issue_owner_repo listlengths 2 2 | issue_body stringlengths 0 261k ⌀ | issue_title stringlengths 1 925 | issue_comments_url stringlengths 56 81 | issue_comments_count int64 0 2.5k | issue_created_at stringlengths 20 20 | issue_updated_at stringlengths 20 20 | issue_html_url stringlengths 37 62 | issue_github_id int64 387k 2.46B | issue_number int64 1 127k |
|---|---|---|---|---|---|---|---|---|---|
[
"MonetDB",
"MonetDB"
] | Date: 2005-10-04 05:44:51 +0200
From: Wouter Alink <<alink>>
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2005-11-25 01:13:41 +0100
## Comment 4032
Date: 2005-10-04 17:44:51 +0200
From: Wouter Alink <<alink>>
I thought we should at least have a decent amount of
bugs for coming thursday.... ;)
<b>2</b>,
<b>3</b>,
5.000000,
23.000000,
23.000000,
23.000000
The above is the answer, what is the query?
<a><b>2</b><b>3</b></a>/(child::b|child::c),
sum(<a><b>2</b><b>3</b></a>/b),
sum(<a><b>2</b><b>3</b></a>/(b|c)),
sum(<a><b>2</b><b>3</b></a>/(child::b|child::c)),
sum(<a><b>2</b><b>3</b></a>/(self::b|child::b))
which i guess is incorrect...
Note that the casting bug (the results should be
denoted as integers instead of decimals) has already
been reported before, there is at least a test for it
on the Testweb.
## Comment 4033
Date: 2005-10-04 17:46:23 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
by the way,
sum(<a><b>2</b><b>3</b></a>/(b))
also produces 23.000000
## Comment 4034
Date: 2005-11-10 00:13:15 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay 2005-11-09, vzzzbx: ADDED TEST / SUCCESS
p.s. the correct result should be
<b>2</b><b>3</b>, 5, 5, 5, 5
## Comment 4035
Date: 2005-11-10 14:08:31 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
We tracked your problem down to a bug (actually a misunderstanding) in
the implementation of fn:data(). fn:data() is implicitly called due to the
application of fn:sum().
Pathfinder tries to rewrite fn:data() calls into possibly more efficient
implementations, which succeeded for your first example. The other
examples lead to slightly different Core trees. Rewriting failed and you hit
the buggy implementation of fn:data().
I've just committed a patch that makes rewriting slightly more intelligent,
which "accidentally" fixes your example. Jan is currently working on the
fix for the actual problem, fn:data().
## Comment 4036
Date: 2005-11-25 01:39:47 +0100
From: @peterboncz
Logged In: YES
user_id=591107
Jan, was this fixed by your fn:data checkin of today?
## Comment 4037
Date: 2005-11-25 13:13:41 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
yes -- at least it works with my version now :)
## Comment 4038
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1313008 at http://sourceforge.net/support/tracker.php?aid=1313008
| XQuery: sum + (a|b) | https://api.github.com/repos/MonetDB/MonetDB/issues/710/comments | 0 | 2020-11-30T08:50:10Z | 2024-06-27T11:06:30Z | https://github.com/MonetDB/MonetDB/issues/710 | 753,282,715 | 710 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-10-04 05:06:55 +0200
From: Wouter Alink <<alink>>
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2007-11-30 04:20:05 +0100
## Comment 4025
Date: 2005-10-04 17:06:55 +0200
From: Wouter Alink <<alink>>
The attached query (i could not produce a smaller
example, it might be related to an earlier reported bug
(if-for-let-bug)) returns the following error message
(CVS head of pathfinder).
ERROR: interpret_assignment: unknown variable 'ipik001'
In the generated MIL of the XQuery the ipik001 is
referenced only for destruction (ipik001 := nil), but
never initialized. If this line is removed from the
MIL-script, MonetDB complains about some other
non-initialized variables.
I guess the scopes are not handled correctly.
The source-XML-doc is in my home-folder
/home/alink/test-data/bug_ipik.xml
I would appreciate it not being spread around too much.
(we'll have a happy bug-day ;)
## Comment 4026
Date: 2005-11-10 00:05:06 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: NO TEST ADDED
note; i'm (at the moment) not able to reproduce the bug with
a smaller document/query. when i come across a smaller
example, i'll add it. If anyone knows what the problem,
he/she is welcome to provide a test-script.
(p.s. the query+doc provided do generate the bug but cannot
be used in the testweb)
## Comment 4027
Date: 2005-11-25 00:56:04 +0100
From: @peterboncz
Logged In: YES
user_id=591107
I tested the query on the music.xml document as I lack your
result.xml. On my document, the error did not occur anymore.
could you re-check this with the latest version?
## Comment 4028
Date: 2006-11-25 04:20:02 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 4029
Date: 2006-11-29 10:09:23 +0100
From: @drstmane
Logged In: YES
user_id=572415
Originator: NO
over-ruling "sf-robot":
re-set to "Pending" since adding a proper test has neither been done nor finally been discarded, yet.
User (submitter) & developer (assignee),
please check, again.
## Comment 4030
Date: 2007-11-30 04:20:05 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 4031
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1312972 at http://sourceforge.net/support/tracker.php?aid=1312972
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| XQuery: unknown variable 'ipik001' | https://api.github.com/repos/MonetDB/MonetDB/issues/709/comments | 0 | 2020-11-30T08:50:07Z | 2024-06-28T13:34:11Z | https://github.com/MonetDB/MonetDB/issues/709 | 753,282,685 | 709 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-10-04 11:28:43 +0200
From: Wouter Alink <<alink>>
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2005-11-14 08:10:20 +0100
## Comment 4017
Date: 2005-10-04 11:28:43 +0200
From: Wouter Alink <<alink>>
from what i could find on the internet, i assume that
distinct-values should accept integer values, and also
a sequence of mixed types.
let $i := (1,1,2)
return distinct-values($i)
type error: no variant of function fn:distinct-values
accepts the given argument type(s): integer+
----------------------
let $i := (1,"aap",1)
return distinct-values($i)
type error: no variant of function fn:distinct-values
accepts the given argument type(s): (string | integer)+
## Comment 4018
Date: 2005-11-09 23:57:59 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: TEST ADDED / SUCCESS
## Comment 4019
Date: 2005-11-10 11:44:51 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
this was an old one -- during playing around in august 2004
the type of distinct-values got mixed up.
It was string* and is with my next checkin atomic* again.
Please close the bug report after my commit.
## Comment 4020
Date: 2005-11-14 18:29:33 +0100
From: @drstmane
Logged In: YES
user_id=572415
cannot yet close this bug:
query:
let $i := (1,1,2)
return distinct-values($i)
IMHO correctly returns:
1 2
But query:
$i := (1,"aap",2.5)
return distinct-values($i)
results in:
1 2.500000 "aap"
while I'd expect:
1 "aap" 2.500000
... or am I wrong ... ?
(cf.,
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora4/tests_BugDay_2005-11-09_0.9.3/distinct_integers.SF-1312716.out.00.html)
## Comment 4021
Date: 2005-11-14 19:59:30 +0100
From: @Teggy
Logged In: YES
user_id=642632
Hi Stefan,
you are "wrong". The result of distinct-values() is
considered a set and the order of items in the resulting
sequence is implementation-dependent:
http://www.w3.org/TR/xpath-functions/func-distinct-values
--Teggy
## Comment 4022
Date: 2005-11-14 20:10:20 +0100
From: @drstmane
Logged In: YES
user_id=572415
Thanks, Teggy!
I approved the stable output and close this bug report.
Stefan.
## Comment 4023
Date: 2005-11-15 09:43:37 +0100
From: @drstmane
Logged In: YES
user_id=572415
Thanks, Teggy!
I approved the stable output and close this bug report.
Stefan.
## Comment 4024
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1312716 at http://sourceforge.net/support/tracker.php?aid=1312716
| XQuery: distinct-values doesn't accept integers | https://api.github.com/repos/MonetDB/MonetDB/issues/708/comments | 0 | 2020-11-30T08:50:04Z | 2024-06-27T11:06:28Z | https://github.com/MonetDB/MonetDB/issues/708 | 753,282,650 | 708 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-10-03 08:02:57 +0200
From: @mlkersten
To: @njnes
Version: -- development
Last updated: 2007-02-02 09:31:53 +0100
## Comment 4008
Date: 2005-10-03 20:02:57 +0200
From: @mlkersten
Problem reported below indicates we need better protection
against overflowing log directories and at least ensure
that
database directory and log directory are in sink.
The recovery process should be safe against removal of
any accidental removal of directories.
The long awaited for Mguardian should include a scheme to
deal with overflow in this area, e.g. sent email to the dba
when the space for logs become tight.
---------------------------
It seems MonetDB has problem with handling of huge
tables (larger than
amount of memory available).
I tried to load large amount of data into a table using
COPY command.
I loaded same data (6 millions records, 700+MBytes) few
times. Loading
was successful two times, on third time server crushed.
After crush I
could not start the server again, it output messages
!WARNING: GDKlockHome: ignoring empty or invalid .gdk_lock.
!FATAL: BBPinit: cannot properly process bat/BACKUP/.
and exited.
Removing of the database data directory did not help
much - server
output messages
!ERROR:mvc_init: unable to create system tables
!WARNING: BBPincref: range error 410
Couple minutes later it output
!ERROR: BATSIGcrash: Mserver internal error
(Segmentation fault),
please restart .
!ERROR: (One potential cause could be that your disk
might be full...)
Well, insufficient disk space could be a cause, before
server crashed
I had approximately 2GB free. But I have cleaned up 10
GB and problem
persists.
How can I troubleshoot and fix the server ? What are
memory / disk
space requirements when bulk loading data ?
I tested with 32-bit system, 512 MB RAM + 1GB swap,
Ubuntu Linux,
MonetDB-4.9.3, nightly build dated September 30.
--
Best regards,
Andrei Martsinchyk
## Comment 4009
Date: 2005-10-03 20:29:05 +0200
From: @grobian
Logged In: YES
user_id=963970
the same problem was reported today by a few students who
more or less did the same: loading a lot of data into MonetDB.
## Comment 4010
Date: 2005-10-04 13:31:01 +0200
From: @grobian
Logged In: YES
user_id=963970
from Andrei:
It was due to insufficient disk space.
After freeing up 10 GB I successfully loaded more than 2GB
of data.
Server crush if there is lack of disk space is an issue, but
not critical one.
I guess we need some better/more indicative error messaging.
## Comment 4011
Date: 2005-10-05 22:08:20 +0200
From: @njnes
Logged In: YES
user_id=43556
Protecting against users removing important data is impossible.
The bugs we should fix are 1) don't crash on loading (to
much) data, but report the problem and gracefully drop the
inserts.
When we startup without enough diskspace we should give more
informative error messages.
## Comment 4012
Date: 2005-12-21 11:28:32 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay: stmane: NO TEST / POSTPONED:
I'm not sure, whether it is feasible to provide a proper
test script ...
## Comment 4013
Date: 2006-06-06 21:03:16 +0200
From: @grobian
Logged In: YES
user_id=963970
how about adding support for some "informative" error
messages to start with?
## Comment 4014
Date: 2007-01-17 19:44:00 +0100
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
Why this bug is not closed?
Regards,
Romulo
## Comment 4015
Date: 2007-02-02 09:31:53 +0100
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
Testing against out of space may cause other parts of the system to become instable (OS etc). So the output of such a test won't be stable. So we close this bug without a test!
## Comment 4016
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1312130 at http://sourceforge.net/support/tracker.php?aid=1312130
| SQL:segfault upon recovery/restart | https://api.github.com/repos/MonetDB/MonetDB/issues/707/comments | 0 | 2020-11-30T08:50:00Z | 2024-06-27T11:06:27Z | https://github.com/MonetDB/MonetDB/issues/707 | 753,282,607 | 707 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-28 08:01:59 +0200
From: Christian Gruen <<christiangruen>>
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2005-10-04 01:36:24 +0200
## Comment 4002
Date: 2005-09-28 20:01:59 +0200
From: Christian Gruen <<christiangruen>>
I successfully installed MonetDB on a Darwin 8.0
64bit-machine, using the --enable-bi1970-01-01 01:01:04 +0100 flag,
whereas Pathfinder stops because of warnings,
which are treated as errors. As soon as I remove
the -Werror flags, I can't properly load the
pathfinder module:
_________________________________
-> FINK FILES ADDED (for Bison)
export PATH=/sw/bin/:$PATH
-> BUILD OPTION ADDED (otherwise wrong
architecture for libxml2 is reported)
> ./configure --prefix=/Users/test/monet --enable-
bi1970-01-01 01:01:04 +0100 --build=powerpc64
> make
warning being treated as errors:
pf_support.glue.c: In function
'CMDpflock_unpack19420754':
pf_support.glue.c:1217: warning: cast to pointer
from integer of different size
...
-> "-Werror" flags removed from runtime/
Makefile, line 294/295:
> make
> make install
> Mserver --dbinit="module(pathfinder); pfstart();"
Monet Database Server V4.8.2
Copyright (c) 1993-2005, CWI. All rights
reserved.
Compiled for powerpc-apple-darwin8.2.0/64bit
with 64bit OIDs;
dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
!ERROR: TBL_openmod: DL_open: library not
found.
!ERROR: TBL_loadmod: module load _pf_support
cancelled.
!ERROR: moduleClient: module(_pf_support) load
error.
!ERROR: TBL_openmod: DL_open: library not
found.
!ERROR: TBL_loadmod: module load _pathfinder
cancelled.
!ERROR: moduleClient: module(_pathfinder) load
error.
!ERROR: interpret: no matching MIL operator to
'pflock()'.
MonetDB> loaded();
-> Pathfinder module is listed
MonetDB> shred_doc("/Users/test/test/books.xml",
"books.xml");
!ERROR: interpret: no matching MIL operator to
'lock_set(void)'.
!MAYBE YOU MEAN:
! lock_set(lock) : void
_________________________________
Regards,
Christian G.
## Comment 4003
Date: 2005-09-29 09:12:27 +0200
From: @grobian
Logged In: YES
user_id=963970
About the build option:
OSX Tiger provides dev-libs/libxml2-2.6.16 which appears to
be pretty up-to-date to me. Fink builds 32-bits
binaries/libraries, because they usually compile with gcc3
and also want to maintain compatibility with pre-G5
machines. Apple on the other hand usually builds 'fat'
binaries/libraries that are 64-bits capable. It might be
worth trying to just use the system provided libxml2 library.
Because a compilation error on pf_support, I expect it
normal that the pf_support library (dylib) cannot be loaded.
This might be related to the powerpc64 build env.
## Comment 4004
Date: 2005-09-29 12:45:02 +0200
From: @grobian
Logged In: YES
user_id=963970
I tried configuring xquery on 64-bits using Apple provided
tools and a bunch of extra files to compile MonetDB/XQuery
on top of a stock install of OSX Tiger used by Gentoo on
OSX. The extra file can be downloaded from:
http://distfiles.gentoo.org/monetdb-pathfinder-bison-sources-0.8.2.tar.bz2
Actually I altered the ebuild to use 64-bits. It seems that
the libxml2 provided by Apple doesn't do 64-bits, as
configure bails out in config.log:
configure:31139: checking for xmlParseDocument in -lxml2
configure:31169: gcc -m64 -o conftest -O2 -std=c99 -Wall -W
-O6 -fomit-frame- pointer -finline-functions
-I/usr/include/libxml2 -L/usr/lib -lxml2 -lz -
lpthread -liconv -lm conftest.c -lxml2 >&5
ld64 warning: in /usr/lib/libxml2.dylib, file does not
contain requested architecture
ld64 warning: in /usr/lib/libz.dylib, file does not contain
requested architecture
ld64 warning: in /usr/lib/libiconv.dylib, file does not
contain requested architecture
ld64 warning: in /usr/lib/libxml2.dylib, file does not
contain requested architecture
ld64 warning: in /usr/lib/libgcc.a, missing required
architecture ppc64 in fat file
can't resolve symbols:
_xmlParseDocument, referenced from:
_main in ccU3iGQA.o
ld64 failed: symbol(s) not found
collect2: ld returned 1 exit status
So the question now is what fink's library supplies (maybe
`file` helps?)
I compiled libxml2 as ppc64 library to see if it would work
out. (env CFLAGS='-arch ppc64 -m64' ./configure
--disable-dependency-tracking --with-python=no) Building a
fat library somehow didn't work :/
However, I ended up with a 64-bit libxml2:
% file /var/tmp/libxml2/lib/libxml2.2.6.20.dylib
/var/tmp/libxml2/lib/libxml2.2.6.20.dylib: Mach-O 64-bit
dynamically linked shared library ppc64
Then configure happily went to the end when doing appending
"--enable-bi1970-01-01 01:01:04 +0100 --with-libxml2=/var/tmp/libxml2/" to my
configure call.
Make and make install also succeeded.
Starting however revealed that it will take a lot more
compiling to eventually get it going, because:
% bin/Mserver -c etc/MonetDB.conf
!WARNING: GDKlockHome: created directory
/var/tmp/monetdb/var/lib/MonetDB/dbfarm/default/
!WARNING: GDKlockHome: ignoring empty or invalid .gdk_lock.
!WARNING: BBPdir: initializing BBP.
Monet Database Server V4.8.2
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for powerpc-apple-darwin8/64bit with 64bit OIDs;
dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module(pathfinder);
!ERROR: DL_open:
dlopen(/var/tmp/monetdb/lib/MonetDB/lib_pf_support.dylib,10)
-> Error:
dlopen(/var/tmp/monetdb/lib/MonetDB/lib_pf_support.dylib,
10): Symbol not found: _gzread
!ERROR: Referenced from: /var/tmp/libxml2/lib/libxml2.2.dylib
!ERROR: Expected in: flat namespace
!ERROR: .
!ERROR: TBL_openmod: DL_open:
dlopen(/var/tmp/monetdb/lib/MonetDB/lib_pf_support.dylib,
10): Symbol not found: _gzread
!ERROR: Referenced from: /var/tmp/libxml2/lib/libxml2.2.dylib
!ERROR: Expected in: flat namespace
!ERROR: .
!ERROR: TBL_loadmod: module load _pf_support cancelled.
!ERROR: moduleClient: module(_pf_support) load error.
!ERROR: DL_open:
dlopen(/var/tmp/monetdb/lib/MonetDB/lib_pathfinder.dylib,10)
-> Error:
dlopen(/var/tmp/monetdb/lib/MonetDB/lib_pathfinder.dylib,
10): Symbol not found: _gzread
!ERROR: Referenced from: /var/tmp/libxml2/lib/libxml2.2.dylib
!ERROR: Expected in: flat namespace
!ERROR: .
!ERROR: TBL_openmod: DL_open:
dlopen(/var/tmp/monetdb/lib/MonetDB/lib_pathfinder.dylib,
10): Symbol not found: _gzread
!ERROR: Referenced from: /var/tmp/libxml2/lib/libxml2.2.dylib
!ERROR: Expected in: flat namespace
!ERROR: .
!ERROR: TBL_loadmod: module load _pathfinder cancelled.
!ERROR: moduleClient: module(_pathfinder) load error.
!ERROR: interpret: no matching MIL operator to 'pflock()'.
MonetDB>pf_start();
!ERROR: interpret: no matching MIL operator to 'pf_start()'.
MonetDB>
zlib isn't 64-bits either, and I guess a few other libraries
which are used (libiconv) as well, so it would basically
require a full system install using 64-bits.
In the long run this should eventually be possible using
Portage on OSX. For now I think it's too much of a hassle
to actually get it going.
## Comment 4005
Date: 2005-10-04 13:36:23 +0200
From: @grobian
Logged In: YES
user_id=963970
Marking CLOSED/CANTFIX.
(there is no CANTFIX, won't fix sounded the closest :/ )
## Comment 4006
Date: 2005-11-09 23:50:46 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: NO TEST ADDED / COMPILATION ISSUE
## Comment 4007
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1307181 at http://sourceforge.net/support/tracker.php?aid=1307181
| Compilation problems on G5 / Darwin 8.0 | https://api.github.com/repos/MonetDB/MonetDB/issues/706/comments | 0 | 2020-11-30T08:49:57Z | 2024-06-28T13:34:10Z | https://github.com/MonetDB/MonetDB/issues/706 | 753,282,572 | 706 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-27 09:34:16 +0200
From: roeland <<rweve>>
To: @njnes
Version: -- development
Last updated: 2005-10-09 01:06:49 +0200
## Comment 3992
Date: 2005-09-27 21:34:16 +0200
From: roeland <<rweve>>
When I was working on a benchmarking program and
testing it, I did get the following exception:
Exception in thread "main" java.sql.SQLException:
Server sent unparsable update count: [ 4294966896 ]
To re-create the error, I developed a simple program,
called bug.java (see attachment). This is not nicely
coded, but it does the trick.
The program first initializes by creating a simple
table, called tweety. If it already exists, it deletes
all records in the table. Next, it inserts 25 records
in this table.
When the initialization is done, the program goes into
a loop. Every time in the loop, it executes a count,
insert and delete.
This loop will be repeated 200 times.
The program repeats itself 3 times. The first 2 times,
the program does what it supposed to do, but the 3rd
time it prints the exception (hopefully on your
machines as well).
I tested the program on Windows XP, with Monet 4.8.2
(msi installer).
I'm a bit puzzled about this problem; I hope someone
can help me out.
The last part of the logfile of the last loop (I can
only attach one file?):
TX 1127848808984: write block: 67 bytes
TX 1127848808984: sCREATE TABLE tweety (someid INT NOT
NULL, somechar VARCHAR(255));;
TX 1127848808984: zero block (flush)
RX 1127848809015: zero block (flush)
RX 1127848809015: new block: 45 bytes
RX 1127848809015: read chunk: 45 bytes, left: 0 bytes
RX 1127848809015: !CREATE TABLE: name tweety already in use
TX 1127848809015: write block: 21 bytes
TX 1127848809015: sDELETE FROM tweety;;
TX 1127848809015: zero block (flush)
RX 1127848809015: zero block (flush)
RX 1127848809015: new block: 37 bytes
RX 1127848809015: read chunk: 37 bytes, left: 0 bytes
RX 1127848809015: -
4 querytype
[ 4294966896 ]
If more information is needed, please let me know.
Roeland
## Comment 3993
Date: 2005-09-28 09:21:15 +0200
From: @grobian
Logged In: YES
user_id=963970
This problem seems to be twofold:
1) JDBC makes the incorrect assumption update counts are
integers, and
2) it appears that the SQL backend goes crazy somehow, and
it looks as if the transaction management fails to properly
bring BATs in a consistent state.
consider 1) as unfixeable: the specs define an integer
update count, then 2) is left for Niels.
I could not reproduce your bug exactly, but after running
the example a few times and manually deleting, I got this,
which indicates a bug to me:
monetdb-> select * from tweety;
+------------+----------+
| someid | somechar |
+============+==========+
| 1 | test1 |
| 2 | test2 |
| 3 | test2 |
| 4 | test2 |
| 5 | test2 |
| 1 | test1 |
| 2 | test2 |
| 3 | test2 |
| 4 | test2 |
| 5 | test2 |
| 1 | test1 |
| 2 | test2 |
| 3 | test2 |
| 4 | test2 |
| 5 | test2 |
| 1 | test1 |
| 2 | test2 |
| 3 | test2 |
| 4 | test2 |
| 5 | test2 |
| 1 | test1 |
| 2 | test2 |
| 3 | test2 |
| 4 | test2 |
| 5 | test2 |
+------------+----------+
25 rows
monetdb-> delete from tweety;
-3375 affected rows
## Comment 3994
Date: 2005-09-28 09:55:54 +0200
From: @grobian
Logged In: YES
user_id=963970
I altered the bug.java slightly to make the testing phase
somewhat longer. Altered the subject to reflect the bug
problem somewhat better.
## Comment 3995
Date: 2005-09-28 13:04:43 +0200
From: @grobian
Logged In: YES
user_id=963970
the example code also reveals another (probably related)
issue: a select count(*) gives 0 as answer for a table
where just about 25 items have been inserted.
## Comment 3996
Date: 2005-10-09 01:06:49 +0200
From: @njnes
Logged In: YES
user_id=43556
When deleting all values of a table we clear the tables, a
few cases were missed however. This is now fixed.
## Comment 3997
Date: 2005-11-09 23:49:40 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: RE-ASSIGNED TO Fabian
Fabian, do you still have the test-script you ran to
reproduce the bug?
## Comment 3998
Date: 2005-11-10 09:28:06 +0100
From: @grobian
Logged In: YES
user_id=963970
vzzzbx: see attached files.
## Comment 3999
Date: 2005-11-10 11:26:16 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
that's java...
do you perhaps have a non-java example? otherwise i would
propose to leave this bug for now without a test-script.
## Comment 4000
Date: 2005-11-10 11:30:39 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-11-09, meltdown: NO TEST / TOO LARGE / POSTPONED
## Comment 4001
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1306171 at http://sourceforge.net/support/tracker.php?aid=1306171
| SQL: update count makes no sense when deleting | https://api.github.com/repos/MonetDB/MonetDB/issues/705/comments | 0 | 2020-11-30T08:49:54Z | 2024-06-27T11:06:25Z | https://github.com/MonetDB/MonetDB/issues/705 | 753,282,539 | 705 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-27 02:23:43 +0200
From: roeland <<rweve>>
To: MonetDB4 devs <<bugs-monetdb4>>
Version: -- development
Last updated: 2005-09-27 07:27:42 +0200
## Comment 3982
Date: 2005-09-27 14:23:43 +0200
From: roeland <<rweve>>
We have created a Java program that opens a connection
to the Mserver using:
[code]
Class.forName("nl.cwi.monetdb.jdbc.MonetDriver");
Connection con =
DriverManager.getConnection("jdbc:monetdb://localhost/database?debug=true",
"monetdb", "monetdb");
Statement st = con.createStatement();
...
[/code]
after that, it executes a couple of queries to the
server for, let's say, 10 minutes. After those 10
minutes, the connection is closed with:
[code]
con.close();
[/code]
If we start this program 2 times, everything runs fine.
But if we want to start this program 3 times, the 3rd
program is probably queued in Mserver, because it waits
to connect to the server.
If you kill one of the 2 programs running, the 3rd
program immediately starts a connection to the server
and executes the queries.
Operating System: Windows XP
(MonetDB-Installer-4.8.2.msi) with JDBC connection
Java version: 1.5.0_04
Roeland and Michiel
## Comment 3983
Date: 2005-09-27 17:40:08 +0200
From: @grobian
Logged In: YES
user_id=963970
I can confirm this on linux, using MapiClient as well. A
strace shows that the client blocks on receiving the
challenge string. I cannot really figure out why the server
doesn't send it.
## Comment 3984
Date: 2005-09-27 17:47:52 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
maybe too obvious, but did you try to change:
mapi_clien1970-01-01 01:00:02 +0100
to a higher number in /etc/MonetDB.conf?
## Comment 3985
Date: 2005-09-27 18:12:02 +0200
From: roeland <<rweve>>
Logged In: YES
user_id=1352508
Well, apparently not many people are aware of this
configuration option (we were told it was a bug).
I deleted the comment before the variable and changed 2 in a
higher number, restarted the server and it works fine now.
## Comment 3986
Date: 2005-09-27 18:17:33 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
I completely agree with you Roeland. I already proposed to
add a link to the MonetDB.conf in the start-menu in the
Windows-version (like Apache does), but the argument against
this link at the time was that nobody complaint about it
before.... i guess the time has come... ;)
## Comment 3987
Date: 2005-09-27 19:27:42 +0200
From: @grobian
Logged In: YES
user_id=963970
since when is that option there?!?!?
Seems to me this rediculously low default should be
increased and at least the server should actively reject the
connections instead of keeping them in the TCP queue.
## Comment 3988
Date: 2005-09-27 20:12:50 +0200
From: @mlkersten
Logged In: YES
user_id=490798
I am not sure increasing the default is safe. Concurrency
conflicts were the reason (as far as I can remember) to
reduce it. For the time
being is may be invalid, until a repeatable test is obtained
to isolate
any concurrency conflict.
The reject/wait should indeed be an option of the Mapi library
## Comment 3989
Date: 2005-09-27 20:31:07 +0200
From: roeland <<rweve>>
Logged In: YES
user_id=1352508
The default setting (2) is a little bit annoying when
working with Aqua Data Server (ADS) as well. Starting ADS
starts 1 connection and a Query Analyzer (ctrl-q) window
starts another one. If you want a 2nd Query Analyzer, ADS
doesn't respond any more (actually it is waiting for
Mserver, but that isn't visible).
## Comment 3990
Date: 2005-11-09 23:40:45 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: NO TEST ADDED
NEEDS FOLLOW UP: a test-script needs to be written that
opens n client-sessions, checks if Mserver serves all of
them, and then closes the sessions. As far as I know, such a
test is not yet present. Please correct me if i'm wrong.
## Comment 3991
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1305797 at http://sourceforge.net/support/tracker.php?aid=1305797
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-monetdb4@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Only max. of 2 open connections to Mserver possible | https://api.github.com/repos/MonetDB/MonetDB/issues/704/comments | 0 | 2020-11-30T08:49:51Z | 2024-06-28T13:34:09Z | https://github.com/MonetDB/MonetDB/issues/704 | 753,282,491 | 704 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-27 02:00:44 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-11-15 11:54:36 +0100
## Comment 3975
Date: 2005-09-27 14:00:44 +0200
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__mapi_start_talking referenced in
function _embedded_mil
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__pthread_create referenced in function
_embedded_mil
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__mapi_embedded_init referenced in
function _embedded_mil
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__monet_exec referenced in function
_start_mil_server
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__THRnew referenced in function
_start_mil_server
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__MT_getpid referenced in function
_start_mil_server
embeddedclient.obj : error LNK2019: unresolved external
symbol _static_module_init referenced in function
_start_mil_server
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__mo_free_options referenced in function
_start_mil_server
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__mo_system_config referenced in
function _start_mil_server
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__mo_builtin_settings referenced in
function _start_mil_server
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__monet_init referenced in function
_Monet_init
embeddedclient.obj : error LNK2001: unresolved external
symbol __imp__monet_daemon
embeddedclient.obj : error LNK2001: unresolved external
symbol __imp__GDKembedded
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__GDKgetenv_isyes referenced in function
_Monet_init
embeddedclient.obj : error LNK2001: unresolved external
symbol __imp__GDK_vm_minsize
embeddedclient.obj : error LNK2001: unresolved external
symbol __imp__GDK_mem_bigsize
embeddedclient.obj : error LNK2001: unresolved external
symbol __imp__GDKdebug
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__GDKgetenv referenced in function
_Monet_init
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__GDKsetenv referenced in function
_Monet_init
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__GDKinit referenced in function
_Monet_init
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__mo_substitute referenced in function
_Monet_init
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__GDKfree referenced in function
_Monet_init
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__mo_find_option referenced in function
_Monet_init
embeddedclient.obj : error LNK2019: unresolved external
symbol __imp__GDKmalloc referenced in function
_Monet_init
NMAKE : fatal error U1077: 'cl' : return code '0x2'
Stop.
## Comment 3976
Date: 2005-09-29 12:36:25 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
By trial and error, I just managed it to compile.
Here are the changes I made:
1) Added missing LIBS
2) Changed embeddedclient.mx such that embedded_mil()
is exported and embeddedclient.lib is created
## Comment 3977
Date: 2005-11-09 23:05:38 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: NO TEST ADDED / COMPILATION ISSUE
## Comment 3978
Date: 2005-11-15 10:49:21 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Do you still have difficulty compiling or can this report be
closed?
I can compile MonetDB and SQL without problems using the
Microsoft tools (using up-to-date buildtools obviously).
## Comment 3979
Date: 2005-11-15 11:44:40 +0100
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
MonetDB isok, but building SQL fails:
cd "winsetup"
...
LINK : fatal error LNK1181: cannot open input
file 'libodbc32.lib'
All libs have this 'lib' prefix in Makefile.msc!?
## Comment 3980
Date: 2005-11-15 11:54:35 +0100
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
Aah, it works with up-to-date buildtools!
I close the bug.
## Comment 3981
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1305779 at http://sourceforge.net/support/tracker.php?aid=1305779
| libembeddedmil.dll: missing libs | https://api.github.com/repos/MonetDB/MonetDB/issues/703/comments | 0 | 2020-11-30T08:49:47Z | 2024-06-28T13:34:09Z | https://github.com/MonetDB/MonetDB/issues/703 | 753,282,435 | 703 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-25 07:03:12 +0200
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2006-01-15 02:48:29 +0100
## Comment 3968
Date: 2005-09-25 19:03:12 +0200
From: @grobian
/bin/bash ../../libtool --mode=compile gcc -m64
-DHAVE_CONFIG_H -I. -I../../../src/gdk -I../..
-I../../../src/gdk -I../common -DLIBGDK -g -O2
-std=c99 -D__EXTENSIONS__ -Wall -W -fno-strict-aliasing
-DHWCOUNTERS -DHW_SunOS -DHW_sun4u
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label
-Wno-uninitialized -D_REENTRANT -c -o
libbat_la-gdk_atoms.lo `test -f 'gdk_atoms.c' || echo
'../../../src/gdk/'`gdk_atoms.c
gcc -m64 -DHAVE_CONFIG_H -I. -I../../../src/gdk
-I../.. -I../../../src/gdk -I../common -DLIBGDK -g -O2
-std=c99 -D__EXTENSIONS__ -Wall -W -fno-strict-aliasing
-DHWCOUNTERS -DHW_SunOS -DHW_sun4u
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label
-Wno-uninitialized -D_REENTRANT -c gdk_atoms.c -fPIC
-DPIC -o .libs/libbat_la-gdk_atoms.o
../../../src/gdk/gdk_atoms.mx: In function `fltFromStr':
../../../src/gdk/gdk_atoms.mx:1498: error:
`__builtin_infinity' undeclared (first use in this
function)
../../../src/gdk/gdk_atoms.mx:1498: error: (Each
undeclared identifier is reported only once
../../../src/gdk/gdk_atoms.mx:1498: error: for each
function it appears in.)
../../../src/gdk/gdk_atoms.mx:1501: error: invalid
operands to binary ==
gmake[4]: *** [libbat_la-gdk_atoms.lo] Error 1
gmake[4]: Leaving directory
`/nfs/poseidon/export/scratch2/Temp/monet/MonetDB/build-solaris/src/gdk'
gmake[3]: *** [all] Error 2
gmake[3]: Leaving directory
`/nfs/poseidon/export/scratch2/Temp/monet/MonetDB/build-solaris/src/gdk'
gmake[2]: *** [all-recursive] Error 1
gmake[2]: Leaving directory
`/nfs/poseidon/export/scratch2/Temp/monet/MonetDB/build-solaris/src'
gmake[1]: *** [all-recursive] Error 1
gmake[1]: Leaving directory
`/nfs/poseidon/export/scratch2/Temp/monet/MonetDB/build-solaris'
gmake: *** [all] Error 2
% gcc --version
gcc (GCC) 3.4.2
Copyright (C) 2004 Free Software Foundation, Inc.
This is free software; see the source for copying
conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
math.h doesn't seem to define any NAN values for me.
## Comment 3969
Date: 2005-09-25 23:37:06 +0200
From: @drstmane
Logged In: YES
user_id=572415
Without access to a Solaris 10 machine, I cannot help much,
here --- except:
- search through all your system- & compiler-include files,
where "NAN" is defined as "__builtin_infinity" (or something
else) and under which circumstance/conditions
- search through all your system- & compiler-include files,
where "__builtin_infinity" is declared, and under which
circumstance/conditions
- check, whether the respective include files are
(transitively) included by src/gdk/gdk_atoms.mx, and whether
the respetive circumstance/conditions du hold at time of
inclusion
## Comment 3970
Date: 2005-09-26 14:01:06 +0200
From: @grobian
Logged In: YES
user_id=963970
% find /usr/include -name "*.h" | xargs grep __builtin_infinity
/usr/include/iso/math_c99.h:define INFINITY
__builtin_infinity
The only thing I can find on this is this:
http://gcc.gnu.org/ml/gcc-patches/2005-05/msg02123.html
(which doesn't sound too much as a quick solution)
## Comment 3971
Date: 2005-09-26 14:04:18 +0200
From: @grobian
Logged In: YES
user_id=963970
how do I easily tell GCC to look for it's includes here?
/usr/sfw/lib/gcc/sparc-sun-solaris2.10/3.4.3/include/
instead of
/usr/include/
## Comment 3972
Date: 2005-11-09 23:04:39 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: NO TEST ADDED / COMPILATION ISSUE
## Comment 3973
Date: 2006-01-15 14:48:29 +0100
From: @grobian
Logged In: YES
user_id=963970
Solaris support has been dropped till there's actually a
demand for it.
## Comment 3974
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1304078 at http://sourceforge.net/support/tracker.php?aid=1304078
| MonetDB: __builtin_infitiny on Solaris 10 | https://api.github.com/repos/MonetDB/MonetDB/issues/702/comments | 0 | 2020-11-30T08:49:44Z | 2024-06-28T13:34:52Z | https://github.com/MonetDB/MonetDB/issues/702 | 753,282,404 | 702 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-25 04:57:06 +0200
From: @grobian
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-09-26 01:06:55 +0200
## Comment 3959
Date: 2005-09-25 16:57:06 +0200
From: @grobian
[helios:monet/MonetDB/build-solaris] % gmake
make all-recursive
/bin/bash: make: command not found
gmake: *** [all] Error 127
[helios:monet/MonetDB/build-solaris] % which make
make: Command not found.
[helios:monet/MonetDB/build-solaris] % which gmake
/opt/sfw/bin/gmake
autoconf/automake/configure doesn't like gmake, or am I
missing an obvious flag here?
## Comment 3960
Date: 2005-09-25 17:12:01 +0200
From: @grobian
Logged In: YES
user_id=963970
using some script-fu I replaced make with gmake and am so
far still compiling
## Comment 3961
Date: 2005-09-25 17:31:04 +0200
From: @drstmane
Logged In: YES
user_id=572415
I never expeirenced similar problem on our Solaris 8 box.
Couls you please check with the following simple Makefile,
whether your gmake correctly sets $(MAKE) to "gmake"?
17:25:45 manegold@corona:/tmp $ cat Makefile
all:
echo $(MAKE)
17:25:48 manegold@corona:/tmp $ make
echo make
make
17:25:53 manegold@corona:/tmp $ gmake
echo gmake
gmake
17:25:56 manegold@corona:/tmp $
## Comment 3962
Date: 2005-09-25 19:02:27 +0200
From: @grobian
Logged In: YES
user_id=963970
[helios:/var/tmp] % cat Makefile
all:
echo $(MAKE)
[helios:/var/tmp] % gmake
echo gmake
gmake
## Comment 3963
Date: 2005-09-25 19:06:00 +0200
From: @grobian
Logged In: YES
user_id=963970
maybe I should add that the bootstrap script had some
problems but appeared to continue. It complained about some
syntax options for grep iirc that were too GNU-ish to be
accepted
## Comment 3964
Date: 2005-09-25 23:39:30 +0200
From: @drstmane
Logged In: YES
user_id=572415
I don't think that the bootstrap problems are related to the
make problems,
but I cannot be sure unless I see the detailed bootstrap
output...
Further, maybe "gmake MAKE=gmake ..." helps...
## Comment 3965
Date: 2005-09-26 13:06:55 +0200
From: @grobian
Logged In: YES
user_id=963970
adding Solaris "make" to the path seemed to have cured
whatever tool generated MAKE=make, probably because it
realised it would never work.
gmake now actually starts compiling without any script-fu.
## Comment 3966
Date: 2005-11-09 23:03:52 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: NO TEST ADDED / COMPILATION ISSUE
## Comment 3967
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1303997 at http://sourceforge.net/support/tracker.php?aid=1303997
| Build: unable to use gmake on Solaris 10 | https://api.github.com/repos/MonetDB/MonetDB/issues/701/comments | 0 | 2020-11-30T08:49:41Z | 2024-06-28T13:34:52Z | https://github.com/MonetDB/MonetDB/issues/701 | 753,282,371 | 701 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-23 10:46:32 +0200
From: Jan Rittinger <<jan.rittinger>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-10-19 01:00:20 +0200
## Comment 3956
Date: 2005-09-23 10:46:32 +0200
From: Jan Rittinger <<jan.rittinger>>
just try the query '1 to 10'
## Comment 3957
Date: 2005-11-09 23:01:39 +0100
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-11-09, vzzzbx: TEST ADDED / SUCCESS
## Comment 3958
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1300366 at http://sourceforge.net/support/tracker.php?aid=1300366
| constant stuff does not work in op:to | https://api.github.com/repos/MonetDB/MonetDB/issues/700/comments | 0 | 2020-11-30T08:49:38Z | 2024-06-28T13:34:51Z | https://github.com/MonetDB/MonetDB/issues/700 | 753,282,343 | 700 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-22 02:57:45 +0200
From: Jens Teubner <<teubner>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2006-01-24 08:32:42 +0100
## Comment 3948
Date: 2005-09-22 14:57:45 +0200
From: Jens Teubner <<teubner>>
I just got this reported from Konstanz. I cannot confirm myself, as I
don't have bison 2.0 here. System is Darwin.
Compilation fails with:
gcc -m32 -DHAVE_CONFIG_H -I. -I. -I.. -I. -g -O2 -std=c99 -Wall -
W -Werror-implicit-function-declaration -Werror -Wno-unused-
function -Wno-unused-label -Wno-uninitialized -c gram.tab.c
/usr/share/bison.simple: In function 'yyparse':
/usr/share/bison.simple:361: error: implicit declaration of function
'free'
make[3]: *** [gram.tab.o] Error 1
yyparse() is entirely under control of bison. To me this looks like
bison generates a call to free() without including stdlib.h first, which
would make this a bison bug.
Adding an include <stdlib.h> into gram.y apparently fixed the
problem.
As I said, I can neither confirm nor invalidate the problem, for the
lack of bison 2.0.
## Comment 3949
Date: 2005-09-22 18:04:44 +0200
From: @grobian
Logged In: YES
user_id=963970
/usr/share/bison.simple
this is a system installed bison.
So question is, is this really bison 2, or just bison 1.31
or something which is shipped by default on OSX. (Assuming
the OSX version of Darwin is meant here)
emerge monetdb
should also work on OSX machines
(though I doubt whether that results in the desired
situation here)
## Comment 3950
Date: 2005-09-22 18:15:37 +0200
From: @Teggy
Logged In: YES
user_id=642632
The bison shipped with OS X (GNU bison v1.28) is incapable
of dealing with token position information and thus not a
viable choice in any case.
I am using GNU bison v1.35 on my OS X box and cannot
reproduce the problem reported by Jens (or by the Konstanz
folks, for that matter). Why not ask Konstanz to use that
version of bison instead? (Available via fink.)
## Comment 3951
Date: 2005-09-23 11:44:10 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Teggy,
the Konstanz people got their stuff compiled, don't worry. I mainly
submitted that bug report to have the problem "on file".
If somebody really feels like it, he/she can look into this. Otherwise we are
at least aware of a possible issue if somebody runs into that problem.
Jens
## Comment 3952
Date: 2005-09-25 23:47:13 +0200
From: @drstmane
Logged In: YES
user_id=572415
On Fedora Core 4 (64-bit), MonetDB, Pathfinder & Burg
compile without problems with GNU bison 2.0 . Hence, it
doesn't seem to be a GNU bison 2.0 problem/bug in general,
but (only?) on Darwin, or at least some (which?) version of
Darwin...
## Comment 3953
Date: 2005-12-19 16:59:25 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
## Comment 3954
Date: 2006-01-24 20:32:41 +0100
From: @grobian
Logged In: YES
user_id=963970
Since bison isn't necessary any more to build
MonetDB/XQuery, I think this bug is obsolete now. Please
reopen if you think otherwise. For me the problem doesn't
occur anymore on Darwin8.
## Comment 3955
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1298707 at http://sourceforge.net/support/tracker.php?aid=1298707
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| possible problem with burg and bison 2.0 | https://api.github.com/repos/MonetDB/MonetDB/issues/699/comments | 0 | 2020-11-30T08:49:35Z | 2024-06-27T11:06:18Z | https://github.com/MonetDB/MonetDB/issues/699 | 753,282,309 | 699 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-21 05:15:33 +0200
From: Wouter Alink <<alink>>
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2005-11-25 01:03:20 +0100
## Comment 3943
Date: 2005-09-21 17:15:33 +0200
From: Wouter Alink <<alink>>
Given the query:
let $d := doc("bug.xml")
return
if (1 = 0) then
()
else
let $f := $d//file
return element { "result" } {
element { "matches" } { count($f) },
element { "results" } {
for $i in 1 to 10
return element { "result" } { count($f[$i]/*) }
}
}
And the document:
<file><file>a</file><file>b</file></file>
Results in:
ERROR: CTrefine: both BATs must have the same
cardinality and their heads must form a 1-1 match.
The query cannot be reduced very much. Leaving out the
'if' or the 'element { "matches" }', or the '/*' will
make the query behave correctly.
## Comment 3944
Date: 2005-09-23 10:13:27 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
In addition to the mil optimizer which prunes necessary code
(see example below), the join recognition ignored that an
if-then-else expression also introduces a new scope...
The last checkin fixes the latter problem (and this bug).
example:
var v_item001 := v_item000;
var v_kind001; prunes the kind code
v_item001 := v_item001.insert(nil,12@0);
v_item000.fetch(0); returns 1 tuple
v_kind000.fetch(0); returns 0 tuples
## Comment 3945
Date: 2005-11-25 01:03:20 +0100
From: @peterboncz
Logged In: YES
user_id=591107
this works for me now. closing..
## Comment 3946
Date: 2005-12-19 16:57:42 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: ALREADY IN TESTWEB
pathfinder/tests/BugsViaSourgeforce/Tests/ID.1297836-1.xq
## Comment 3947
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1297836 at http://sourceforge.net/support/tracker.php?aid=1297836
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| XQuery: if+let+for+element | https://api.github.com/repos/MonetDB/MonetDB/issues/698/comments | 0 | 2020-11-30T08:49:32Z | 2024-06-27T11:06:17Z | https://github.com/MonetDB/MonetDB/issues/698 | 753,282,251 | 698 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-21 11:29:12 +0200
From: Jens Teubner <<teubner>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: unspecified
Last updated: 2005-09-21 12:47:01 +0200
## Comment 3939
Date: 2005-09-21 11:29:12 +0200
From: Jens Teubner <<teubner>>
"[an error occurred while processing this directive]"
http://monetdb.cwi.nl/Research/Articles/index.html
## Comment 3940
Date: 2005-09-21 12:47:01 +0200
From: @drstmane
Logged In: YES
user_id=572415
thanks for lettin us know!
It should be fixed, again.
## Comment 3941
Date: 2005-12-19 16:56:11 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / DOCUMENTATION
## Comment 3942
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1297454 at http://sourceforge.net/support/tracker.php?aid=1297454
| website: research articles page broken | https://api.github.com/repos/MonetDB/MonetDB/issues/697/comments | 0 | 2020-11-30T08:49:29Z | 2024-06-27T11:06:16Z | https://github.com/MonetDB/MonetDB/issues/697 | 753,282,222 | 697 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-20 03:31:57 +0200
From: @grobian
To: Romulo Goncalves <<goncalve>>
Version: -- development
Last updated: 2007-03-28 07:34:32 +0200
## Comment 3927
Date: 2005-09-20 15:31:57 +0200
From: @grobian
monetdb-> create table a (id int primary key);
Operation successful
monetdb-> create table b (id int, foreign key (id)
references a(id) on delete cascade);
Operation successful
monetdb-> insert into a values (1);
1 affected row
monetdb-> insert into b values(1);
1 affected row
monetdb-> delete from a;
1 affected row
monetdb-> select * from b;
+------------+
| id |
+============+
| 1 |
+------------+
1 row
monetdb->
although the ON DELETE CASCADE clause is accepted, it
is not implemented. We better not accept this, or best
just implement it.
## Comment 3928
Date: 2005-12-19 16:55:28 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / FAILURE
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/delete_cascade.SF-1296395.sql
## Comment 3929
Date: 2006-01-15 14:46:50 +0100
From: @grobian
Logged In: YES
user_id=963970
this needs major work, postponing for now.
## Comment 3930
Date: 2006-01-22 21:26:44 +0100
From: @drstmane
Logged In: YES
user_id=572415
postponed until after the release, I suppose?
## Comment 3931
Date: 2006-01-22 21:58:36 +0100
From: @grobian
Logged In: YES
user_id=963970
yes
## Comment 3932
Date: 2006-06-06 21:00:30 +0200
From: @grobian
Logged In: YES
user_id=963970
Niels, shall we remove the syntactical support for this from
the parser? (Like is done for WITH CHECK OPTION?)
## Comment 3933
Date: 2007-01-18 11:03:25 +0100
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
In my opinion it should be implemented.
It will take some time, but it is a nice feature for our system.
What do you think?
Regards,
Romulo
## Comment 3934
Date: 2007-01-24 11:35:54 +0100
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
well, if you now try to execute a DELETE over a primary key with foreign keys you get an error message.
See this example.
The tables and constraints:
create table t1(id int, name varchar(1024));
alter table t1 add constraint id_p primary key(id);
create table t2(id_f int, age int, foreign key(id_f) references t1(id));
alter table t2 add constraint age_p primary key(age);
The test:
sql>delete from t1 where id = 0;
MAPI = monetdb@localhost:50000
QUERY = delete from t1 where id = 0;
ERROR = !DELETE FROM: cannot delete from table 't1' there are foreign keys depending on it, please use DELETE CASCADED!
sql>drop table t2;
sql>delete from t1 where id = 0;
[ 1 ]
sql>
The same it will implemented for updates over primary keys.
My question is:
Is this enough to close the bug or the implementation of DELETE/UPDATE CASCADED is needed?
Regards,
Romulo
## Comment 3935
Date: 2007-01-24 11:41:01 +0100
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
> My question is:
> Is this enough to close the bug or the implementation of DELETE/UPDATE
> CASCADED is needed?
No, this is not enough, because as far as I understand, "ON DELETE CASCADE" is still accepted, but not acted upon.
## Comment 3936
Date: 2007-03-28 17:09:15 +0200
From: @romulogoncalves
Logged In: YES
user_id=1498628
Originator: NO
This bug, can be closed.
The example given by this bug is covered by the test:
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/delete_cascade.SF-1296395.sql
Which now is green. :)
Regards,
Romulo
## Comment 3937
Date: 2007-03-28 19:34:32 +0200
From: @grobian
Logged In: YES
user_id=963970
Originator: YES
Closing, kudos to Romulo.
## Comment 3938
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1296395 at http://sourceforge.net/support/tracker.php?aid=1296395
| SQL: ON DELETE/UPDATE CASCADE accepted but not executed | https://api.github.com/repos/MonetDB/MonetDB/issues/696/comments | 0 | 2020-11-30T08:49:26Z | 2024-06-27T11:06:15Z | https://github.com/MonetDB/MonetDB/issues/696 | 753,282,188 | 696 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-20 03:27:55 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2005-10-23 12:09:26 +0200
## Comment 3924
Date: 2005-09-20 15:27:55 +0200
From: @grobian
monetdb-> create table t (id int);
Operation successful
monetdb-> insert into t values (1);
1 affected row
monetdb-> insert into t values (select max(id) + 1 from t);
Error: syntax error, unexpected SELECT in: "insert into
t values (select"
monetdb-> insert into t values ((select max(id) + 1
from t));
1 affected row
I think the extra ( ) here should be superfluous and it
should be allowed to go without them.
## Comment 3925
Date: 2005-12-19 16:49:19 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/subquery_syntax.SF-1296390.sql
## Comment 3926
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1296390 at http://sourceforge.net/support/tracker.php?aid=1296390
| SQL: relaxing subqueries | https://api.github.com/repos/MonetDB/MonetDB/issues/695/comments | 0 | 2020-11-30T08:49:23Z | 2024-06-27T11:06:14Z | https://github.com/MonetDB/MonetDB/issues/695 | 753,282,154 | 695 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-19 11:08:09 +0200
From: Wouter Alink <<alink>>
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2005-10-06 11:00:16 +0200
## Comment 3912
Date: 2005-09-19 23:08:09 +0200
From: Wouter Alink <<alink>>
from what i understand of;
http://www.w3.org/TR/xmlschema-2/int
it seems that the xs:integer type should support
integers of at least 18 digits to meet the XQuery-specs
as it is derived from xs:decimal.
however, it seems that currently integers above the 2G-
border return an error, even if monetDB is compiled in
everything-64-bits-mode (at least MonetDB says:
Compiled for x86_64-redhat-linux-gnu/64bit with
64bit OIDs; dynamically linked.)
## Comment 3913
Date: 2005-09-19 23:21:52 +0200
From: @drstmane
Logged In: YES
user_id=572415
In MonetDB, "int" are 32-bit (signed, i.e.,
[-((2^31)-1),(2^31)-1], or at most 9 digits) and "lng" are
64-bit (signed, i.e., [-((2^63)-1),(2^63)-1], or at most 18
digits. This is independed of whether MonetDB is compiled
with 32 or 64 bit.
Currently, xs:integer seems to be implemented as "int", and
is hence limited to 9 digits. To allow for 18 digits,
xs:integer would have to be implemented as "lng".
## Comment 3914
Date: 2005-09-19 23:27:30 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
that's already what i guessed, but i wanted to make sure i
didn't overlook a special feature of monetdb... ;)
i guess this means the bug-report is valid as the
implementation doesn't met the specs.
## Comment 3915
Date: 2005-09-20 10:21:57 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
i think i have my mserver working with xs:integers as lng's
now... have to wait for mtest to finish, if all is ok, i
will check it in
## Comment 3916
Date: 2005-09-20 17:58:02 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
attached is the diff-script for the fix: the xs:integer type
is now represented as 'lng' in MIL (used to be 'int')
according to the testweb everything is fine, but i doubt if
the tests are sufficient
from what i understand of the milprint_summer code is that
sometimes the 'num'-field of a 'PFcnode_t' is (ab)used for
storing 'fid's ('c->sem.num = fid;'). these fid's are
actually oid's, disguised as int's. this will not be very
safe, but could somebody (preferably an expert,
Jan/Stefan??) have a look at the attached code?
## Comment 3917
Date: 2005-09-20 19:15:30 +0200
From: @grobian
Logged In: YES
user_id=963970
wouter,
nice diff! Too bad sf doesn't allow inline viewing, but ok.
I was just wondering whether long long int would be
portable, and whether %lld is also portable. I'm not sure
on this, so ignore if invalid, but aren't there macros
defined in MonetDB somewhere to deal with the right "long"
on each and every exotic platform we test on?
## Comment 3918
Date: 2005-09-20 19:27:56 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Stefan suggested something like "LNGFMT", but don't know in
which library it is. Any clues?
The diff at least fixes some casting from signed to unsigned.
## Comment 3919
Date: 2005-09-20 19:29:56 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fabian,
you're right, %lld is for sure not portable to (among
others) native Windows compilers and the same holds most
probably for long long int.
However, pathfinder does not compile with native Windows
compilers (at least not with the Microsoft compiler); on
Windows, we're bound to using MinGW, and that is GNU.
Nevertheless, we should indeed consider using the macros
"lng" and "LNGFMT" as in MonetDB; probabaly, they'll have to
be re-defined for pathfinder since the compiler is supposed
to compile stand-alone without (too many) MonetDB
dependencies...
## Comment 3920
Date: 2005-10-06 22:33:42 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
BugDay_2005-10-06: Done.
CLAIMED BY vzzzbx
ALREADY IN TESTWEB
test-script "tests/XQuery/Tests/type_integer.xq" was already
added
## Comment 3921
Date: 2005-10-06 23:00:15 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
fixed in CVS current-branch.
due to the rather large amount of changes to the code, i
checked it in as a "feature" in the current- branche ;)
## Comment 3922
Date: 2005-12-19 16:46:15 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: ALREADY IN TESTWEB
pathfinder/tests/XQuery/Tests/type_integer.xq
## Comment 3923
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1295786 at http://sourceforge.net/support/tracker.php?aid=1295786
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| XQuery: ceiling xs:integer | https://api.github.com/repos/MonetDB/MonetDB/issues/694/comments | 0 | 2020-11-30T08:49:20Z | 2024-06-27T11:06:13Z | https://github.com/MonetDB/MonetDB/issues/694 | 753,282,122 | 694 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-19 04:47:51 +0200
From: @mlkersten
To: @njnes
Version: -- development
Last updated: 2005-09-19 06:53:38 +0200
## Comment 3908
Date: 2005-09-19 16:47:51 +0200
From: @mlkersten
It is possible to generate serious crashes (Segfault)
using your ability to drop tables from the catalog. It
recovers, as far as I can detect upon restart.
## Comment 3909
Date: 2005-09-19 18:53:38 +0200
From: @grobian
Logged In: YES
user_id=963970
I assume you mean system tables here. But I bet you login
as monetdb:monetdb, which makes you the equivalent of root
on your linux box, hence I deem your bug to be invalid.
Please reopen if I did not understand your bug correctly.
## Comment 3910
Date: 2005-12-19 16:42:46 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST
## Comment 3911
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1295482 at http://sourceforge.net/support/tracker.php?aid=1295482
| SQL: self protection too minimalistic | https://api.github.com/repos/MonetDB/MonetDB/issues/693/comments | 0 | 2020-11-30T08:49:17Z | 2024-06-28T13:34:50Z | https://github.com/MonetDB/MonetDB/issues/693 | 753,282,087 | 693 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-19 10:41:45 +0200
From: Jan Rittinger <<jan.rittinger>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2006-01-25 09:13:57 +0100
## Comment 3900
Date: 2005-09-19 10:41:45 +0200
From: Jan Rittinger <<jan.rittinger>>
The help of MapiClient says about the option '-s':
-s stmt | --statement=stmt /* run single statement */
It never came into my mind that this could be used to
change the serialization type of my result if I use it
in combination with '-l xquery'. As I started with '-s
foo' I got some nice text which can be interpreted as
sax events. However other options (dm, xml-seq, xml,
sax) are not mentioned anywhere I looked (website &
Mapi help). Can this please be mentioned at a prominent
place and are there perhaps even more options? -- e.g.
'bar' seems to be the same as sax again. Therefore I
propose to disable non-existing serialization modes.
## Comment 3901
Date: 2005-09-19 10:54:35 +0200
From: @drstmane
Logged In: YES
user_id=572415
In fact, this "-sxml" is kind of a (quick...) hack (sorry,
Peter ;-)).
No being a MAPI expert myself, I resently discussed it with
Niels, and he said that the "challange" feature (and an
extra option to MapiClient) would actually be the correct
way to specify the serialization mode and other
configurations for MonetDB/XQuery.
I (temporarly?) assign this to Peter, as he intorduced the
"-sxml" hack;
of course, any MAPI expert it more than welcome to enlighten
me or Peter how to solve this problem properly.
## Comment 3902
Date: 2005-09-19 16:29:12 +0200
From: @grobian
Logged In: YES
user_id=963970
I think niels meant that there is a 'language' field in the
string sent by the client. This field can be exploited/used
to tell the server in a nice way what kind out output it
should produce. This would then involve a -l (--language)
switch instead.
## Comment 3903
Date: 2005-09-19 16:44:52 +0200
From: @drstmane
Logged In: YES
user_id=572415
"-l xquery" / "--language=xquery" is already used (and
documented) for XQuery.
The point here is, that in addition to setting the language
to XQuery, we also can set the "output style" (xml, dm, ...)
and switch on/off other options like debugging,
optimization, timing for the XQuery compiler, mis-using
MapiClient's "-s"/"--statement=" option.
This hack should be replaced by a proper implementation
using the "challange-string" feature of the MAPI protocol,
i.e., instead of prefixing the XQuery by a line that is
derived from the mis-used "-s" option, we should send a
challange string that contains not only things like
"mode=blocked", but also things like "debug=<dbg-level>",
"optimization=<opt-level>", "timing={on/off}" ...
(Provided my humble knowledge of the MAPI protocol is not
completely bogus...)
## Comment 3904
Date: 2005-09-19 19:04:27 +0200
From: @grobian
Logged In: YES
user_id=963970
The clean solution you propose is already forseen in MCL.
I don't think the current solution is that bad (because it's
under the hood, and the user doesn't see this happening, you
can also change the output per query, although I don't think
it is used now), the real problem is the interface through
MapiClient I guess. I could imagine either one of these two
hacks:
1% MapiClient --language=xquery:xml
2% MapiCliient -lxquery -o xml
at 1 the language can be followed by a colon with the output
mode, if not specified, the default is used, where at 2 a
separate flag is used which can only be used with language
xquery.
I don't really have a preference, so I think it will be
what's easier to implement in MapiClient.
## Comment 3905
Date: 2005-12-19 16:41:36 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / DOCUMENTATION
Either the (mis)use of the -s option for XQuery should be
better document, or an alternative solution should be
implemented. In either case, no test.
## Comment 3906
Date: 2006-01-25 21:13:57 +0100
From: @peterboncz
Logged In: YES
user_id=591107
it is documented as of now..
## Comment 3907
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1295145 at http://sourceforge.net/support/tracker.php?aid=1295145
| documentation of option '-s' missing | https://api.github.com/repos/MonetDB/MonetDB/issues/692/comments | 0 | 2020-11-30T08:49:14Z | 2024-06-27T11:06:11Z | https://github.com/MonetDB/MonetDB/issues/692 | 753,282,047 | 692 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-18 08:28:04 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-09-19 04:05:53 +0200
## Comment 3895
Date: 2005-09-18 20:28:04 +0200
From: @drstmane
After changes made between Thursday, Sep 15 2005, 0:02
CEST, and Friday, Sep 16 2005, 0:02 CEST, tests
tests/XQuery/Tests/let_for[23].xq do not generate/show
"warning: variable reuse: at (0,0-0,0): $a"
any more.
I'm not sure, whether this was intended, or is a bug...
## Comment 3896
Date: 2005-09-19 10:49:15 +0200
From: @drstmane
Logged In: YES
user_id=572415
For the bookkeeping (might be helpful to fix the bug):
The warning is still generated by the (stand-alone)
compiler, but seems to ge swallowed in the Mserver-embedded
compiler after Peter's following changes:
compiler/compile.c: 1.41 -> 1.43
compiler/mil/milprint_summer.c: 1.103 -> 1.104
compiler/include/milprint_summer.h: 1.7 -> 1.8
"
- fixed a bug introduced yesterday in mil_opt
- we now report out-of-memory errors in milprint_summer code
generation
- hm. let's make it compile.
"
Sorry, but I have no time, now, to analyse the details and
fix the problem
...
## Comment 3897
Date: 2005-09-19 16:05:53 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed by Peter.
## Comment 3898
Date: 2005-12-19 16:38:28 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / REDUNDANT
## Comment 3899
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1294712 at http://sourceforge.net/support/tracker.php?aid=1294712
| XQuery: tests tests/XQuery/Tests/let_for[23].xq fail | https://api.github.com/repos/MonetDB/MonetDB/issues/691/comments | 0 | 2020-11-30T08:49:11Z | 2024-06-27T11:06:10Z | https://github.com/MonetDB/MonetDB/issues/691 | 753,282,011 | 691 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-16 09:34:47 +0200
From: @grobian
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2006-06-06 08:58:48 +0200
## Comment 3883
Date: 2005-09-16 21:34:47 +0200
From: @grobian
This applies to the MonetDB 4.8.2 en SQL 2.8.2 and
XQuery 0.8.2 tar distributions. (no group for 4.8.2 is
available yet)
the following files pollute the user's filesystem when
installing the full family suite of MonetDB:
testing stuff that the user didn't ask for IMHO:
/usr/bin/Mapprove.py -> Mtest.py
/usr/bin/Mdiff
/usr/bin/Mfilter.py
/usr/bin/MkillUsers
/usr/bin/Mlog
/usr/bin/Mprofile.py -> Mtest.py
/usr/bin/Mtest.py
/usr/bin/Mtimeout
/usr/bin/prof.py
/usr/lib/MonetDB/Tests
/usr/lib/MonetDB/Tests/sample0
/usr/lib/MonetDB/Tests/sample1
/usr/lib/MonetDB/Tests/sample2
/usr/lib/MonetDB/Tests/sample3
/usr/lib/MonetDB/Tests/sample4
/usr/lib/MonetDB/Tests/smack00
/usr/lib/MonetDB/Tests/smack01
/usr/lib/sql/Tests
/usr/lib/sql/Tests/odbcsample1
/usr/share/MonetDB/Mprofile-commands.lst
/usr/share/MonetDB/Tests
/usr/share/MonetDB/Tests/sample.class
/usr/share/sql/Tests
/usr/share/sql/Tests/Test_Cautocommit.class
/usr/share/sql/Tests/Test_Csavepoints.class
/usr/share/sql/Tests/Test_PStimedate.class
/usr/share/sql/Tests/Test_PStimezone.class
/usr/share/sql/Tests/Test_Rbooleans.class
/usr/share/sql/Tests/Test_Rmetadata.class
/usr/share/sql/Tests/Test_Rpositioning.class
/usr/share/sql/Tests/Test_Rtimedate.class
/usr/share/sql/Tests/Test_Sbatching.class
compilation tools that should not/never be installed:
/usr/bin/burg
/usr/bin/epsffit (note: this file conflicts with
app-text/psutils !!!!)
/usr/bin/prefixMxFile
/usr/lib/autogen
/usr/lib/autogen/am.py
/usr/lib/autogen/autogen.py
/usr/lib/autogen/codegen.py
/usr/lib/autogen/msc.py
/usr/lib/autogen/var.py
an application which I am not sure on what it is doing
in the distribution:
/usr/bin/calibrator
/usr/include/MonetDB/calibrator
/usr/include/MonetDB/calibrator/calib.m
/usr/lib/MonetDB/lib_calib.la
/usr/lib/MonetDB/lib_calib.so -> lib_calib.so.0.0.0
/usr/lib/MonetDB/lib_calib.so.0 -> lib_calib.so.0.0.0
/usr/lib/MonetDB/lib_calib.so.0.0.0
is this a useful .h file to have?:
/usr/include/MonetDB/monet/yytree.h
is this an installer file, or a configuration file?
/usr/include/sql/odbcinst.h
misc:
/usr/share/MonetDB/conf
/usr/share/MonetDB/conf/monet.m4
/usr/share/MonetDB/general.mil ?
/usr/share/MonetDB/lady.gif
/usr/share/MonetDB/quit.mil
## Comment 3884
Date: 2005-09-16 23:08:12 +0200
From: @grobian
Logged In: YES
user_id=963970
I forgot these files which were eliminated in my install:
remove Mx and mel, they are only needed for compilation
rm -f ${D}/usr/bin/Mx ${D}/usr/bin/mel ${D}/usr/bin/idxmx
## Comment 3885
Date: 2005-12-19 16:29:34 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: NO TEST / PACKAGING
## Comment 3886
Date: 2006-01-24 20:26:29 +0100
From: @njnes
Logged In: YES
user_id=43556
The buildtools aren't distributed ie can we close this bug.
## Comment 3887
Date: 2006-01-24 20:29:36 +0100
From: @grobian
Logged In: YES
user_id=963970
No, it still installs:
/usr/bin/epsffit
and that one is the major reason this bug was created,
because that file collides with psutils.
## Comment 3888
Date: 2006-01-24 21:06:04 +0100
From: @drstmane
Logged In: YES
user_id=572415
epsffit is in buildtools, hence not in any distribution.
nevertheless, epsffit and other tools like idxmx are IMHO
legacy and could be removed.
## Comment 3889
Date: 2006-01-24 21:14:02 +0100
From: @grobian
Logged In: YES
user_id=963970
That is quite interesting.
% grep epsffit $INSTALL_DIR/filelist.*
/ufs/fabian/scratch/monetdb/stable/program-x86_64/filelist.buildtools.20060122180501:ufs/fabian/scratch/monetdb/stable/program-x86_64/bin/epsffit
You seem to be right. Now I'm wondering where the file
comes from...
## Comment 3890
Date: 2006-01-25 09:59:52 +0100
From: @grobian
Logged In: YES
user_id=963970
ok, rm -f doesn't return a non-zero exit code if files are
missing... that explains a lot.
The update list of questions/problems then is for 4.10 and
beyond:
testing stuff that the user didn't ask for IMHO:
/usr/bin/Mapprove.py -> Mtest.py
/usr/bin/Mdiff
/usr/bin/Mfilter.py
/usr/bin/MkillUsers
/usr/bin/Mlog
/usr/bin/Mprofile.py -> Mtest.py
/usr/bin/Mtest.py
/usr/bin/Mtimeout
/usr/bin/prof.py
/usr/lib/MonetDB/Tests
/usr/lib/MonetDB/Tests/sample0
/usr/lib/MonetDB/Tests/sample1
/usr/lib/MonetDB/Tests/sample2
/usr/lib/MonetDB/Tests/sample3
/usr/lib/MonetDB/Tests/sample4
/usr/lib/MonetDB/Tests/smack00
/usr/lib/MonetDB/Tests/smack01
/usr/lib/sql/Tests
/usr/lib/sql/Tests/odbcsample1
/usr/share/MonetDB/Mprofile-commands.lst
/usr/share/MonetDB/Tests
/usr/share/MonetDB/Tests/sample.class
/usr/share/sql/Tests
/usr/share/sql/Tests/Test_Cautocommit.class
/usr/share/sql/Tests/Test_Csavepoints.class
/usr/share/sql/Tests/Test_PStimedate.class
/usr/share/sql/Tests/Test_PStimezone.class
/usr/share/sql/Tests/Test_Rbooleans.class
/usr/share/sql/Tests/Test_Rmetadata.class
/usr/share/sql/Tests/Test_Rpositioning.class
/usr/share/sql/Tests/Test_Rtimedate.class
/usr/share/sql/Tests/Test_Sbatching.class
n application which I am not sure on what it is doing
in the distribution:
/usr/bin/calibrator
/usr/include/MonetDB/calibrator
/usr/include/MonetDB/calibrator/calib.m
/usr/lib/MonetDB/lib_calib.la
/usr/lib/MonetDB/lib_calib.so -> lib_calib.so.0.0.0
/usr/lib/MonetDB/lib_calib.so.0 -> lib_calib.so.0.0.0
/usr/lib/MonetDB/lib_calib.so.0.0.0
is this a useful .h file to have?:
/usr/include/MonetDB/monet/yytree.h
is this an installer file, or a configuration file?
/usr/include/sql/odbcinst.h
misc:
/usr/share/MonetDB/conf
/usr/share/MonetDB/conf/monet.m4
/usr/share/MonetDB/general.mil ?
/usr/share/MonetDB/lady.gif
/usr/share/MonetDB/quit.mil
## Comment 3891
Date: 2006-03-09 12:37:57 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
I removed epsffit from the buildtools. It's available in
other packages so there is no need to include it in our
source tree.
Does this close the bug?
## Comment 3892
Date: 2006-03-09 14:30:44 +0100
From: @grobian
Logged In: YES
user_id=963970
I don't think so. There are a lot other files that I wonder
of whether they should be in or not (like the yytree.h file)
## Comment 3893
Date: 2006-06-06 20:58:48 +0200
From: @grobian
Logged In: YES
user_id=963970
Ok. Most of the stuff is cleaned up now, I'll punt whatever
I don't like in the ebuild myself.
## Comment 3894
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1293196 at http://sourceforge.net/support/tracker.php?aid=1293196
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| distribution: clashing and foreign files are installed | https://api.github.com/repos/MonetDB/MonetDB/issues/690/comments | 0 | 2020-11-30T08:49:08Z | 2024-06-28T13:34:50Z | https://github.com/MonetDB/MonetDB/issues/690 | 753,281,965 | 690 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-16 11:39:22 +0200
From: Jan Rittinger <<jan.rittinger>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2006-01-15 02:16:48 +0100
## Comment 3869
Date: 2005-09-16 11:39:22 +0200
From: Jan Rittinger <<jan.rittinger>>
This was my first generated algebra query...
First I thought some errors are not handled in the
staircasejoin, but after removing a 'col_name' call (in
line 74) it worked...
That somehow seems *very* strange
## Comment 3870
Date: 2005-09-16 14:08:33 +0200
From: @drstmane
Logged In: YES
user_id=572415
Jan,
I just ran your query against the latest Mserver (v4.9.3,
CVS head) and it works fine for me (see attached output).
Hence, could you please provide some more information about
the crash, e.g., run Mserver in a debugger to locate the
code where the crash occurs, and maybe why (deref. of
null-pointer, stackoverflow, etc.)?
## Comment 3871
Date: 2005-09-17 23:30:59 +0200
From: @drstmane
Logged In: YES
user_id=572415
The crash occurs in "mallinfo", but apparently only on Jan's
SuSE 9.3 (32bit) ... ???
Further investigation pending...
## Comment 3872
Date: 2005-09-19 13:03:47 +0200
From: @drstmane
Logged In: YES
user_id=572415
The same crash seems to occur with our nightly testing on
Fedora Core 4 (32-bit) and Gentoo 1.6.13; cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_BugsViaSourgeforce/ID.1292727.col_name-segfault.run.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Gentoo1.6.13/tests_BugsViaSourgeforce/ID.1292727.col_name-segfault.run.out.00.html
It might hence be related to 32-bit and/or
(GNU-)compiler-version and/or [g]libc-version ...
## Comment 3873
Date: 2005-11-06 20:58:25 +0100
From: @drstmane
Logged In: YES
user_id=572415
Jan,
your MIL script does not seem to work any more with the
latest version of pathfinder (CVS head), at least it now
produces an empty result.
Could you generate a new MIL(?) test script that reproduces
the bug for you?
Stefan
## Comment 3874
Date: 2005-11-07 09:30:38 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
completed your attempt to apply the recent changes. It now
crashes again and produces a result without col_name.
## Comment 3875
Date: 2005-11-20 18:25:15 +0100
From: @drstmane
Logged In: YES
user_id=572415
On none of out test machines, I can reproduce the problem;
see
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests03/GNU.64.64.d-Fedora4/tests_BugsViaSourgeforce/ID.1292727.col_name-segfault.shred.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora4/tests_BugsViaSourgeforce/ID.1292727.col_name-segfault.shred.out.00.html
Jan,
does the crash still occur for you, and if so, on which
platform (OS +version, compiler +version, MonetDB version,
pathfinder version) and under which circumstances (configure
flags, within/outside Mtest, ...)?
## Comment 3876
Date: 2005-11-24 15:37:30 +0100
From: @drstmane
Logged In: YES
user_id=572415
As Jan just pointed out, the problems also occurs with
nightly testing (again) on Fedora Core 4 (32-bit), Gentoo
1.6.13 (32-bit), and SuSE 9.3 (32-bit), cf.:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_BugsViaSourgeforce/ID.1292727.col_name-segfault.run.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Gentoo1.6.13/tests_BugsViaSourgeforce/ID.1292727.col_name-segfault.run.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-SuSE9.3/tests_BugsViaSourgeforce/ID.1292727.col_name-segfault.run.out.00.html
## Comment 3877
Date: 2005-12-19 16:29:03 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
pathfinder/tests/BugsViaSourgeforce/Tests/ID.1292727.col_name-segfault.*
## Comment 3878
Date: 2005-12-30 16:15:58 +0100
From: @drstmane
Logged In: YES
user_id=572415
The same bug occurs (only) on the same platforms with all
XMark tests when using the algebra translation; cf. e.g.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_XMark_alg/q01.out.00.html
## Comment 3879
Date: 2006-01-14 13:17:04 +0100
From: @drstmane
Logged In: YES
user_id=572415
The below mentioned XMark/alg tests have been "fixed" by
avoising the use of "col_name()".
However, the test script for http://sourceforge.net/support/tracker.php?aid=1314982 "kill(0) causes SEGFAULT"
https://sourceforge.net/tracker/index.php?func=detail&aid=1314982&group_id=56967&atid=482468
now segvault on (only) the same platforms as this one;
hence, they might be related...
## Comment 3880
Date: 2006-01-14 13:35:18 +0100
From: @drstmane
Logged In: YES
user_id=572415
The same platforms segfault also with the test of http://sourceforge.net/support/tracker.php?aid=921173 "extremely long and complex query causes SEGFAULT"
https://sourceforge.net/tracker/index.php?func=detail&aid=921173&group_id=56967&atid=482468
## Comment 3881
Date: 2006-01-15 14:16:48 +0100
From: @drstmane
Logged In: YES
user_id=572415
Closed as the bug is covered by this new report
1406591 "several tests cause segfault on the same subset
of
platforms"
https://sourceforge.net/tracker/index.php?func=detail&aid=1406591&group_id=56967&atid=482468
## Comment 3882
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1292727 at http://sourceforge.net/support/tracker.php?aid=1292727
This bug was marked DUPLICATE in the database it was moved from.
Changing resolution to "MOVED"
| Mserver segfault because of 'col_name' | https://api.github.com/repos/MonetDB/MonetDB/issues/689/comments | 0 | 2020-11-30T08:49:05Z | 2024-06-28T13:34:49Z | https://github.com/MonetDB/MonetDB/issues/689 | 753,281,928 | 689 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-16 11:20:34 +0200
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-09-18 11:20:43 +0200
## Comment 3859
Date: 2005-09-16 11:20:34 +0200
From: Wouter Alink <<alink>>
Shredding the document
/ufs/alink/test-data/trecvid_bug.xml
and answering the query
for $d in $doc("trecvid_bug.xml")
return element { "result" } {
$d//VideoSegment
}
returns the following error:
ERROR: XML Generation: idxTV, BAT[tmp]: 12228 not
within [0,12227]
what i discovered was that the transient WS was
corrupted: PRE_SIZE didn't match PRE_LEVEL at some point.
The xml-doc itself is valid xml and also firing queries
NOT asking for VideoSegment-elements return correct
results.
Unfortunately the content of the xml-doc cannot be
freely distributed, but those interested can contact me
for a copy.
## Comment 3860
Date: 2005-09-16 13:51:00 +0200
From: @drstmane
Logged In: YES
user_id=572415
Wouter,
three questions:
1)
Do you refer to the CVS Head version of pathfinder (as
indecated in this bug report) or to the stable 0.8.2 version
resp. the XQuery_0-8 CVS branch (as you told me yesterday)?
2)
Did you make sure, that it is not your code (-changes) that
do corrupt the
transient WS?
3)
Did you try to run your Mserver with property checking
enabled (i.e., Mserver --debug=10 ...)?
## Comment 3861
Date: 2005-09-16 15:23:51 +0200
From: @drstmane
Logged In: YES
user_id=572415
Here's a simple example:
15:22:15
manegold@corona:~/_/scratch0/Monet/Development/Current/MonetDB
$ echo '<x><a><a/><a/></a></x>//a' | pf | Mserver --set
monet_prompt='' Monet Database Server V4.9.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult>
<a><a/><a/></a><a/><a/>
</XQueryResult>
15:22:39
manegold@corona:~/_/scratch0/Monet/Development/Current/MonetDB
$ echo 'element{"y"}{<x><a><a/><a/></a></x>//a}' | pf |
Mserver --set monet_prompt=''
Monet Database Server V4.9.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult>
<y><a><a/><a/></a><a/>!ERROR: XML Generation: idxTV,
BAT[tmp]: 9 not within [0,8]
!ERROR: xquery_print_result_main: operation failed.
15:22:41
manegold@corona:~/_/scratch0/Monet/Development/Current/MonetDB $
## Comment 3862
Date: 2005-09-16 18:28:12 +0200
From: @drstmane
Logged In: YES
user_id=572415
.. as Jan (R) already discovered, this is indeed caused a
bug in the loop_lifted descendant (of self) step
implementation --- it ook me a while to understand why, and
what's actually going wrong --- by now, I have a vague idea,
what's wrong, but I won't have time to fix it today ...
## Comment 3863
Date: 2005-09-18 19:53:45 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS:
loop-lifted descendant step did not work properly
in some cases of overlapping subtrees in different iterations.
Wouter,
could you please check, whether it also works for you, now,
and if so close this bug?
Thanks!
## Comment 3864
Date: 2005-09-18 21:14:53 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
i'm sorry i have to disappoint you, but the fix didn't solve the
bug (at least not on my system). I updated
ll_staircasejoin.mx and recompiled, and the small test-
script 'element { "y" } { ... }' now works ok, but the larger
example (actually all four of the trecvid docs) still fail with the
same error-message. Stefan, could you check this
behaviour by trying the initial query of the bug report
for $d in $doc("trecvid_bug.xml")
return element { "result" } {
$d//VideoSegment
}
on the trecvid data (/ufs/alink/test-data/trecvid_bug.xml)?
## Comment 3865
Date: 2005-09-18 23:08:02 +0200
From: @drstmane
Logged In: YES
user_id=572415
Second attempt:
Now, Wouter's query doesn't produce any error anymore.
However, I did not verify the (quite large) result by hand ...
## Comment 3866
Date: 2005-09-18 23:20:43 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
none of the 4 trecvid-documents produce errors anymore,
but the content is not validated (a little difficult for the large
documents). Stefan, do you think your second small
example suffices assuming correct behaviour? i close the
bug.
## Comment 3867
Date: 2005-12-19 16:28:29 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
pathfinder/tests/BugsViaSourgeforce/Tests/ID.1292711-1.*
## Comment 3868
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1292711 at http://sourceforge.net/support/tracker.php?aid=1292711
| XQuery: incorrect WS | https://api.github.com/repos/MonetDB/MonetDB/issues/688/comments | 0 | 2020-11-30T08:49:02Z | 2024-06-27T11:06:07Z | https://github.com/MonetDB/MonetDB/issues/688 | 753,281,900 | 688 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-16 11:11:06 +0200
From: @drstmane
To: @mlkersten
Version: -- development
Last updated: 2005-09-17 11:26:43 +0200
## Comment 3855
Date: 2005-09-16 11:11:06 +0200
From: @drstmane
!ERROR: interpret: no matching MIL operator to
'antijoin(BAT[int,int], BAT[int,int])'.
cf. e.g.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/MonetDB/.mTests103/GNU.64.64.d-Fedora4/tests_suite_00/test_36.out.00.html
## Comment 3856
Date: 2005-09-17 23:26:43 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed: Martin added the missing MIL antijoin command.
## Comment 3857
Date: 2005-12-19 16:27:18 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB
## Comment 3858
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1292701 at http://sourceforge.net/support/tracker.php?aid=1292701
| MonetDB: new test tests/suite_00/Tests/test_36.milS fails | https://api.github.com/repos/MonetDB/MonetDB/issues/687/comments | 0 | 2020-11-30T08:48:58Z | 2024-06-27T11:06:06Z | https://github.com/MonetDB/MonetDB/issues/687 | 753,281,862 | 687 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-13 01:20:03 +0200
From: @drstmane
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2005-09-13 04:40:00 +0200
## Comment 3851
Date: 2005-09-13 13:20:03 +0200
From: @drstmane
In case an error occurs while parsing an XML document,
the error handler ("print_error()") that is registered
via "xmlSetGenericErrorFunc()" in "untime/shredder.mx"
seems to be called incorrectly, i.e., not with a
variable argument list in one call, but with an extra
call per argument!??
(see trace below for details)
In case the argument (i.e., the error message) contains
a format string, the respective arguments hence seem to
be missing in the same call, causing at least "strange"
error messages (arguments are filled in by random data
from the stack), or even a crash.
My knowledge of libxml2 seems to be too limited to
analyse and potentially fix this bug in reasonable
time. I'm actually not sure, whether the problem is
libxml2 inherent, caused by a wrong definition
or implementation of the error handler
("print_error()"), caused by an incorrect/insufficient
registration of this error handler with libxml2, or
caused by an incorrect/insufficient
initialization/usage of libxml2's XML parser.
The problem occurs at least with libxml2 2.6.20 on
Fedora Core 4 (64-bit) and libxml2 2.6.17 on SuSE 9.3
(32-bit).
Jan (F.),
assuming that your libxml2 experience is more than mine
(which is none), would you have time to have a look at
this problem?
Of course, any other libxml2 expert's opinion is also
more than welcome!
========
12:48:20
manegold@corona:~/_/scratch0/Monet/Development/Current/pathfinder
$ l /tmp/foo.xml ls: /tmp/foo.xml: No such file or
directory
12:48:21
manegold@corona:~/_/scratch0/Monet/Development/Current/pathfinder
$ echo 'shred_doc("/tmp/foo.xml","foo.xml");' | Mserver
--dbinit='module(pathfinder);' --set monet_prompt=''
Monet Database Server V4.9.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
!ERROR: I/O
!ERROR: warning :
!ERROR:
!ERROR: shred:mainSHREDDER:
xmlCreateFileParserCtxt("/tmp/foo.xml") failed
!ERROR: CMDshred2bats: operation failed.
12:48:30
manegold@corona:~/_/scratch0/Monet/Development/Current/pathfinder
$ gdb --args Mserver --dbinit='module(pathfinder);' GNU
gdb Red Hat Linux (6.3.0.0-1.21rh)
Copyright 2004 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public
License, and you are
welcome to change it and/or distribute copies of it
under certain conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB. Type "show
warranty" for details.
This GDB was configured as
"x86_64-redhat-linux-gnu"...Using host libthread_db
library "/lib64/libthread_db.so.1".
(gdb) break print_error
Function "print_error" not defined.
Make breakpoint pending on future shared library load?
(y or [n]) y
Breakpoint 1 (print_error) pending.
(gdb) r
Starting program:
/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Development/Current/.GNU.64.64.d..PREFIX./bin/Mserver
--dbinit=module\(pathfinder\)\;[Thread debugging using
libthread_db enabled]
[New Thread 46912515620032 (LWP 6809)]
[New Thread 1082132832 (LWP 6845)]
Monet Database Server V4.9.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
Breakpoint 2 at 0x2aaab5076ae0: file
/ufs/manegold/_/scratch0/Monet/Development/Current/pathfinder/runtime/shredder.mx,
line 1380.
Pending breakpoint "print_error" resolved
MonetDB>shred_doc("/tmp/foo.xml","foo.xml");
[Switching to Thread 46912515620032 (LWP 6809)]
Breakpoint 2, print_error (ctx=0x0, msg=0x3ff07deb4d
"I/O ") at
/ufs/manegold/_/scratch0/Monet/Development/Current/pathfinder/runtime/shredder.mx:1380
1380 {
(gdb) cont
Continuing.
!ERROR: I/O
Breakpoint 2, print_error (ctx=0x0, msg=0x3ff07debce
"warning : ") at
/ufs/manegold/_/scratch0/Monet/Development/Current/pathfinder/runtime/shredder.mx:1380
1380 {
(gdb) cont
Continuing.
!ERROR: warning :
Breakpoint 2, print_error (ctx=0x0, msg=0x3ff07e8bee
"%s") at
/ufs/manegold/_/scratch0/Monet/Development/Current/pathfinder/runtime/shredder.mx:1380
1380 {
(gdb) cont
Continuing.
!ERROR:
!ERROR: shred:mainSHREDDER:
xmlCreateFileParserCtxt("/tmp/foo.xml") failed
!ERROR: CMDshred2bats: operation failed.
MonetDB>
Program exited normally.
(gdb) 12:49:20
manegold@corona:~/_/scratch0/Monet/Development/Current/pathfinder
$
12:49:21
manegold@corona:~/_/scratch0/Monet/Development/Current/pathfinder
$
## Comment 3852
Date: 2005-09-13 16:40:00 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Retrieved original libxml2-error.c from net for correct
example.
Problems were:
- we should use vfprintf()/vsnprintf() for correct
argument handling.
- function may be called more than once for 1 error
message so we cannot
use GDKerror() here.
## Comment 3853
Date: 2005-12-19 16:35:08 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: NO TEST / REDUNDANT:
shred_doc error messages are covered by test
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/shred_doc_with_space.SF-1211372.*
## Comment 3854
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1289651 at http://sourceforge.net/support/tracker.php?aid=1289651
| XQuery: error handler incorrectly called from libxml2?? | https://api.github.com/repos/MonetDB/MonetDB/issues/686/comments | 0 | 2020-11-30T08:48:55Z | 2024-06-27T11:06:05Z | https://github.com/MonetDB/MonetDB/issues/686 | 753,281,826 | 686 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-09 03:10:39 +0200
From: @drstmane
To: Fabian Groffen <<fabian>>
Version: -- development
Last updated: 2005-09-13 01:05:00 +0200
## Comment 3842
Date: 2005-09-09 15:10:39 +0200
From: @drstmane
After Peter's "noroot" changes in runtime/serialize.mx
(revision 1.32),
test runtime/Tests/text1.milS
========
module(pathfinder);
xquery("dm", "<a>aap</a>/text()").printf();
xquery("dm", "<a>2</a>/text()").printf();
xquery("dm-mapi", "<a>aap</a>/text()").printf();
xquery("dm-mapi", "<a>2</a>/text()").printf();
xquery("xml-seq-mapi", "<a>aap</a>/text()").printf();
xquery("xml-seq-mapi", "<a>2</a>/text()").printf();
xquery("xml-mapi", "<a>aap</a>/text()").printf();
xquery("xml-mapi", "<a>2</a>/text()").printf();
========
shows the following differences compared with the
"stable" output:
========
---
../.GNU.64.64.d..PREFIX./mTests.pathfinder.43/runtime/text1.stable.out
2005-08-21 13:17:37.000000000 +0200
+++
../.GNU.64.64.d..PREFIX./mTests.pathfinder.43/runtime/text1.test.out
2005-09-09 14:38:47.000000000 +0200
@@ -30,7 +30,8 @@
60 length
1 tuplecount
0 id
-[ "text {\"aap\"}" ]
+>
+=[ "aap" ]
xquery("xml-seq-mapi", "<a>2</a>/text()").printf();
-
3 querytype
@@ -40,7 +41,8 @@
60 length
1 tuplecount
1 id
-[ "text {\"2\"}" ]
+>
+=[ "2" ]
xquery("xml-mapi", "<a>aap</a>/text()").printf();
=<?xml version="1.0" encoding="utf-8"?>
========
I have to admit, I'm not sure, whether this is a bug,
whether the original "stable" output is not correct, or
whether the output changes was intended.
Peter, could you (or someone else) please have a look
at this?
Thanks in advance!
## Comment 3843
Date: 2005-09-09 15:21:54 +0200
From: @grobian
Logged In: YES
user_id=963970
> @@ -30,7 +30,8 @@
> 60 length
> 1 tuplecount
> 0 id
> -[ "text {\"aap\"}" ]
> +>
> +=[ "aap" ]
This is wrong in any case (the '>' and '=' certainly don't
belong there, hence XML:DB is broken. I remember from a
discussion with Wouter that the output as it was (text
{"aap"}) was the desired one.
> xquery("xml-seq-mapi", "<a>2</a>/text()").printf();
> -
> 3 querytype
> @@ -40,7 +41,8 @@
> 60 length
> 1 tuplecount
> 1 id
> -[ "text {\"2\"}" ]
> +>
> +=[ "2" ]
again the '>' and '=' don't belong here and break XML:DB.
This bug might be fixable in serialize.mx. I guess they are
caused by Peter's change/decision to dismiss the modes
somehow. I don't know since when this bug appeared. It is
obvious to me that there is a mode missing somehow because
the '=' indicates it is in another mode when printing the
results.
## Comment 3844
Date: 2005-09-09 15:52:13 +0200
From: @drstmane
Logged In: YES
user_id=572415
The bug occurs since Peter's "noroot" changes in
runtime/serialize.mx
(revision 1.32) on Sunday Sep 04 2005, and might be releated
to his "noroot" changes in compiler/compile.c (revision
1.40), compiler/include/pathfinder.h (revision 1.34), &
compiler/mil/milprint_summer.c (revision 1.91) on Friday Sep
02 2005.
## Comment 3845
Date: 2005-09-13 09:45:15 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
the following fails as well:
xquery: "aap", 1+1
using: MapiClient -lx -sxml
gives:
<XQueryResult>
"aap" 2
</XQueryResult>
In this example I would think that the double quotes should
not be there (the answer is not even valid XML). I guess it
has something to do with the same checkin.
## Comment 3846
Date: 2005-09-13 10:28:06 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Wouter,
I cannot see any problem in your example. The output is of course valid
XML. There is no need to escape quotes in text content. You only need to
escape them if you want to have quotes within quotes (within attribute
values).
The quotes appear in the serialization to denote that the output is actually
a string. Note that the query
"foo"
is NOT the same as
text { "foo" } !
The former returns a string, while the latter returns a text node. This is why
the two are serialized differently.
Of course, there is the ever-lasting discussion on how serialization should
be performed, and you can argue for the one or the other solution. The
output for your example appears to be consistent and correct for me,
though.
## Comment 3847
Date: 2005-09-13 11:02:20 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
I apologize for claiming it to be non-xml, but on the XQuery
website the following example is given:
----
<b>{1, "2", "3"}</b>
The constructed element node has one child, a text node
containing the value "1 2 3".
----
And because we do add a <XQueryResult/>-element around the
answer, i would expect, correct me if i'm wrong, that the
string-node would become a text-node. Of course this
behaviour would only happen in case the 'xml'-serialization
method is used.
## Comment 3848
Date: 2005-09-13 13:05:00 +0200
From: @grobian
Logged In: YES
user_id=963970
Fixed in CVS.
Please continue (restart) discussion on the representation
of multiple resource sets in a separate bug.
## Comment 3849
Date: 2005-12-19 16:22:56 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB
## Comment 3850
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1286005 at http://sourceforge.net/support/tracker.php?aid=1286005
| XQuery: test runtime/Tests/text1.milS failes | https://api.github.com/repos/MonetDB/MonetDB/issues/685/comments | 0 | 2020-11-30T08:48:52Z | 2024-06-27T11:06:04Z | https://github.com/MonetDB/MonetDB/issues/685 | 753,281,797 | 685 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-09 02:30:26 +0200
From: Jens Teubner <<teubner>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-11-20 07:01:56 +0100
## Comment 3837
Date: 2005-09-09 14:30:26 +0200
From: Jens Teubner <<teubner>>
Since Peter's parser modifications, one of the parser test cases
fails.
I actually have a slight preference to use the XQuery module syntax
to declare functions for the SOAP stuff. (No full implementation of
the module functionality, but just use it to declare your SOAP
functions.) That may, by the way, would also allow you to omit the
query part in the MIL code altogether and just print the PROCs that
you need, no need to post-process anything.
As (fortunately) your modifications to the parser are really small and
should not have any negative effect on the remaining parsing rules, I
could also live with the current solution/hack. Least thing to do then,
however, is to fix up the test cases.
## Comment 3838
Date: 2005-11-20 18:30:27 +0100
From: @drstmane
Logged In: YES
user_id=572415
According to the TestWeb, this test is working fine, again.
Peter, Jens,
could you please check/confirm this, and in case you agree,
mark this report "fixed" and close it.
Thanks,
Stefan
## Comment 3839
Date: 2005-11-20 19:01:56 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
SOAP stuff has been implemented the "proper" way in the
meantime, and Pathfinder now supports modules to some
degree. Peter's parser hack is no longer neccessary since
then, and the problem is fixed again.
## Comment 3840
Date: 2005-12-19 16:22:38 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB
## Comment 3841
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1285957 at http://sourceforge.net/support/tracker.php?aid=1285957
| XQuery: test compiler/parser/bad-006.xq fails | https://api.github.com/repos/MonetDB/MonetDB/issues/684/comments | 0 | 2020-11-30T08:48:50Z | 2024-06-27T11:06:03Z | https://github.com/MonetDB/MonetDB/issues/684 | 753,281,756 | 684 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-08 01:35:36 +0200
From: mihalych <<mihalych_>>
To: @njnes
Version: -- development
Last updated: 2005-11-26 09:00:38 +0100
## Comment 3833
Date: 2005-09-08 13:35:36 +0200
From: mihalych <<mihalych_>>
monetdb-> create table t1 (id1 int, val1 varchar(255));
Operation successful
monetdb-> create table t2 (id2 int, val2 varchar(255));
Operation successful
monetdb-> insert into t1 values (1,'1');
1 affected row
monetdb-> insert into t2 values (1,'2');
1 affected row
monetdb-> update t1 set val1 = (select val2 from t2
where id1 = id2)
where id1 in (select id2 from t2);
Error: ERROR: interpret: unknown variable 's23'.
ERROR: interpret_params: join(param 2): evaluation error.
## Comment 3834
Date: 2005-11-26 09:00:38 +0100
From: @njnes
Logged In: YES
user_id=43556
The update code now handles this case, ie the update
statement is now a lot more powerfull (not allways set =
'value').
Test script is added to
src/test/BugTracker/Tests/update_with_correlated_subselect.SF-1284791.sql
## Comment 3835
Date: 2005-12-19 16:22:15 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
sql/src/test/BugTracker/Tests/update_with_correlated_subselect.SF-1284791.*
## Comment 3836
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1284791 at http://sourceforge.net/support/tracker.php?aid=1284791
| SQL: Update with correlated subselect | https://api.github.com/repos/MonetDB/MonetDB/issues/683/comments | 0 | 2020-11-30T08:48:47Z | 2024-06-27T11:06:02Z | https://github.com/MonetDB/MonetDB/issues/683 | 753,281,733 | 683 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-07 05:21:11 +0200
From: Wouter Alink <<alink>>
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2005-11-25 12:51:58 +0100
## Comment 3826
Date: 2005-09-07 17:21:11 +0200
From: Wouter Alink <<alink>>
(100 div 7) mod 5
gives the following error:
ERROR: interpret: no matching MIL operator to '[%](BAT
[void,dbl], BAT[void,dbl])'.
My interpretation of the XQuery specs would be that we
need a new MIL operator:
[%](BAT[void,dbl], BAT[void,dbl])
because the following is being said on the w3c website:
(by the way, the definition doesn't mention if it is allowed
for A and B to have different types, but i guess not, as
the last sentence identifies 'four' different operators, am i
right?)
-----------------
A div B
A of type numeric
B of type numeric
op:numeric-divide(A, B)
result of type:
numeric; but xs:decimal if both operands are xs:integer
A mod B
A of type numeric
B of type numeric
op:numeric-mod(A, B)
result of type:
numeric
[Definition: When referring to a type, the term numeric
denotes the types xs:integer, xs:decimal, xs:float, and
xs:double.] An operator whose operands and result are
designated as numeric might be thought of as
representing four operators, one for each of the numeric
types.
## Comment 3827
Date: 2005-09-16 14:41:49 +0200
From: @drstmane
Logged In: YES
user_id=572415
Wouter,
to be honest, I don't know how '%' (mod) is/should be
defined on double (i.e., floating point numbers, or even
non-integers in general).
What's the result of 5.1 mod 4.2? 0.9? (Then 5.1 div 4.2 =
1.0 while 5.1/4.2 = 1.214285714?)
What do the XQuery specs say about this?
## Comment 3828
Date: 2005-09-16 15:53:27 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
http://www.w3.org/TR/xquery-operators/
Section 6.2.6 gives a detailed description of the mod-operator.
## Comment 3829
Date: 2005-09-19 09:49:26 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
the current version of 'mod' just uses the '%' operator in
MIL. Because it is probably rather hard to find a sensitive
solution in MIL, I would call it an un unsupported
feature... (& it should be mentioned in the KnownIssues file
then)
## Comment 3830
Date: 2005-11-25 00:51:58 +0100
From: @peterboncz
Logged In: YES
user_id=591107
should work now
## Comment 3831
Date: 2005-12-19 16:21:40 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: TEST ADDED / SUCCESS:
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/div_mod.SF-1284049.*
## Comment 3832
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1284049 at http://sourceforge.net/support/tracker.php?aid=1284049
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| XQuery: (100 div 7) mod 5 | https://api.github.com/repos/MonetDB/MonetDB/issues/682/comments | 0 | 2020-11-30T08:48:43Z | 2024-06-27T11:06:01Z | https://github.com/MonetDB/MonetDB/issues/682 | 753,281,691 | 682 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-07 01:33:27 +0200
From: @MarcinZukowski
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-09-17 11:29:22 +0200
## Comment 3821
Date: 2005-09-07 13:33:27 +0200
From: @MarcinZukowski
This script behaves incorrectly:
var b:=new(oid,oid);
b.insert(0@0,0@0);
b.insert(1@0,1@0);
b.insert(10@0,10@0);
b.info().find("hsorted").print();
[ "65" ] ok
b.info().find("tsorted").print();
[ "65" ] ok
var m:=b.merge();
m.info().find("hsorted").print();
[ "65" ] ok
m.info().find("tsorted").print();
[ "65" ] ok
b.insert(5@0,0@0);
b.info().find("hsorted").print();
[ "0" ] ok
b.info().find("tsorted").print();
[ "0" ] ok
var m:=b.merge();
m.info().find("hsorted").print();
[ "65" ] NOT OK!
m.info().find("tsorted").print();
[ "65" ] NOT OK!
var s:=m.split();
m.info().find("hsorted").print();
[ "65" ] NOT OK!
m.info().find("tsorted").print();
[ "65" ] NOT OK!
BATs returned by merge() are incorrectly marked sorted,
causing the same data after split() to be also marked
sorted.
With -d10 Monet naturally complains, and that breaks
amdb testweb :)
## Comment 3822
Date: 2005-09-07 13:57:18 +0200
From: @MarcinZukowski
Logged In: YES
user_id=607094
BTW, merge() in general doesn't have to work with 64-bit oids...
## Comment 3823
Date: 2005-09-17 23:29:22 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS by Marcin (and me).
## Comment 3824
Date: 2005-12-19 16:11:55 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: TEST ADDED / SUCCESS:
MonetDB/tests/BugDay_2005-12-19_4.9.3/Tests/merge_properties.SF-1283842.*
## Comment 3825
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1283842 at http://sourceforge.net/support/tracker.php?aid=1283842
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| merge() incorrectly marks BATs as sorted | https://api.github.com/repos/MonetDB/MonetDB/issues/681/comments | 0 | 2020-11-30T08:48:40Z | 2024-06-27T11:06:00Z | https://github.com/MonetDB/MonetDB/issues/681 | 753,281,657 | 681 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-05 11:25:39 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-10-05 10:42:12 +0200
## Comment 3813
Date: 2005-09-05 11:25:39 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select 2 * interval '1' second;
[ 2 ]
works, but
sql>select interval '1' second * 2;
!ERROR: second_interval_sht: operation failed.
doesn't work.
## Comment 3814
Date: 2005-09-05 11:42:41 +0200
From: @grobian
Logged In: YES
user_id=963970
according to the big SQL book this should be:
SELECT INTERVAL SECOND(1) * 2;
(which should yield in an INTERVAL)
## Comment 3815
Date: 2005-09-05 12:31:50 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
Well, I can only use unofficial sources, e.g. drafts or:
<http://savage.net.au/SQL/sql-2003-2.bnf.htmlinterval%
20value%20expression>
This would indicate, that there's nothing special about
<interval term 2>, i.e. my <interval literal> would be
allowed.
Out of curiosity, what's the 'big SQL book'?
## Comment 3816
Date: 2005-09-05 14:22:52 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
If you search at http://print.google.com for "you can also
multiply or divide an interval", you'll find Melton/Simon:
SQL: 1999, p. 158:
INTERVAL ‘2' DAY * 3 -- results in an interval of 6 days
## Comment 3817
Date: 2005-09-05 23:24:25 +0200
From: @njnes
Logged In: YES
user_id=43556
added missing *,+,/,- functions on intervals
## Comment 3818
Date: 2005-09-06 09:46:43 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
Currently, interval <op> interval no longer works:
sql>select interval '1' second + interval '1' second;
!ERROR: interpret: no matching MIL operator to '+(lng, oid)'.
## Comment 3819
Date: 2005-12-19 16:28:52 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/interval_times_int.SF-1281996.sql
## Comment 3820
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1281996 at http://sourceforge.net/support/tracker.php?aid=1281996
| SQL: select interval '1' second * 2 | https://api.github.com/repos/MonetDB/MonetDB/issues/680/comments | 0 | 2020-11-30T08:48:37Z | 2024-06-27T11:05:59Z | https://github.com/MonetDB/MonetDB/issues/680 | 753,281,621 | 680 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-03 11:44:11 +0200
From: Jens Teubner <<teubner>>
To: Jens Teubner <<teubner>>
Version: -- development
Last updated: 2005-09-12 01:23:09 +0200
## Comment 3809
Date: 2005-09-03 11:44:11 +0200
From: Jens Teubner <<teubner>>
The
let $a := doc("foo.xml") cast as xs:anyNode
return $a/bar
returns a type error within Jan's "simplifier" in milprint_summer.c:
type error: can't cast type 'untypedAtomic?' to type 'node'
The original Core tree is apparently type correct, while Jan's
simplifier introduces some type error.
## Comment 3810
Date: 2005-09-12 13:23:09 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
As Jan noticed correctly, this query is actually illegal. `cast as' only allows
casting to atomic types (no node types). I've just committed a fix for the
typechecker, which now ensures that a cast's target type is always a
subtype of atomic?.
## Comment 3811
Date: 2005-12-19 16:26:24 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/type_error_in_cast.SF-1281176.xq
## Comment 3812
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1281176 at http://sourceforge.net/support/tracker.php?aid=1281176
| Core simplification in milprint_summer.c leads to type error | https://api.github.com/repos/MonetDB/MonetDB/issues/679/comments | 0 | 2020-11-30T08:48:34Z | 2024-06-27T11:05:58Z | https://github.com/MonetDB/MonetDB/issues/679 | 753,281,587 | 679 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-02 06:25:14 +0200
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-09-05 12:49:40 +0200
## Comment 3805
Date: 2005-09-02 18:25:14 +0200
From: @drstmane
When using a multiplex operation (e.g., "[+]") with
void-headed synchronized/aligned BATs that have
var-sized tail values, the dense (non-nil) void-heads
are (IMHO unnecessarily) materialized to oid.
The reason seems to be, that "array_opt" in struct
"monet_sig_t" is set to FALSE for var-sized atoms
(cf. function "TBL_multiplex_resolve" in
"src/monet/monet_tbl.mx":
line 2307: "sig->array_opt = sig->merge_opt = 0;",
line 2436" "sig->array_opt = !ATOMvarsized(tpe);"), and
hence, the "array" implementation in
"src/monet/monet_multiplex.mx" (functions
"interpret_{special,multiplex}_array") is NOT used.
Could someone please check, whether this is indeed
necessary?
There seems to be code in
"interpret_{special,multiplex}_array" that can handle
var-sized atoms (cf. marco "FIRST_PARAM_FIXEDSIZE" in
"src/monet/monet_multiplex.mx").
Here's a small experiment that depicts the problem:
========
$ cat /tmp/m.mil
--------
var i :=
new(void,int).seqbase(0@0).insert(nil,1).access(BAT_READ);
var s := [str](i);
print(i,s);
print(i.[+](1));
print(i.[+](i));
print(s.[+]("1"));
print(s.[+](s));
quit();
========
$ cat /tmp/m.mil | Mserver --trace --set
monet_prompt="`echo -e '\n-> '`"
--------
Monet Database Server V4.8.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
-> var i :=
new(void,int).seqbase(0@0).insert(nil,1).access(BAT_READ);
-> var s := [str](i);
-> print(i,s);
-------------------------
h t t name
void int str type
-------------------------
[ 0@0, 1, "1" ]
-> print(i.[+](1));
interpret_multiplex
interpret_multiplex: min_size == REALLY_REALLY_BIG =>
tpe=BAThtype(b)=6 (b->htype=0)
interpret_multiplex: sig->pack_fcn=0x2aaaaada76c7,
sig->fast_fcn=0x2aaaaadb404b, sig->special=0,
sig->array_opt=1, all_synced=0x730fd8,
sig->merge_opt=0, all_sorted=0
interpret_multiplex: sig->fast_fcn != NULL &&
sig->array_opt && all_synced => interpret_multiplex_array
interpret_multiplex_array
-----------------
h t name
void int type
-----------------
[ 0@0, 2 ]
-> print(i.[+](i));
interpret_multiplex
interpret_multiplex: min_size == REALLY_REALLY_BIG =>
tpe=BAThtype(b)=6 (b->htype=0)
interpret_multiplex: sig->pack_fcn=0x2aaaaada76c7,
sig->fast_fcn=0x2aaaaadb404b, sig->special=0,
sig->array_opt=1, all_synced=0x730fd8,
sig->merge_opt=1, all_sorted=0
interpret_multiplex: sig->fast_fcn != NULL &&
sig->array_opt && all_synced => interpret_multiplex_array
interpret_multiplex_array
-----------------
h t name
void int type
-----------------
[ 0@0, 2 ]
-> print(s.[+]("1"));
interpret_multiplex
interpret_multiplex: min_size == REALLY_REALLY_BIG =>
tpe=BAThtype(b)=6 (b->htype=0)
interpret_multiplex: sig->pack_fcn=0x2aaaab10e443,
sig->fast_fcn=0x2aaaab10ff79, sig->special=0,
sig->array_opt=0, all_synced=0x7ae2a8,
sig->merge_opt=0, all_sorted=0
interpret_multiplex: fall-through => 'durchfall'
implementation => interpret_multiplex_fcn
interpret_multiplex_fcn
interpret_multiplex_fcn: argv[i].vtype == TYPE_bat &&
argv[i].len != int_nil => real_type = b->htype = 0
interpret_multiplex_fcn: desc.nbats == 1 && real_type
== TYPE_void => desc.id.vtype = real_type = 0
-----------------
h t name
void str type
-----------------
[ 0@0, "11" ]
-> print(s.[+](s));
interpret_multiplex
interpret_multiplex: min_size == REALLY_REALLY_BIG =>
tpe=BAThtype(b)=6 (b->htype=0)
interpret_multiplex: sig->pack_fcn=0x2aaaab10e443,
sig->fast_fcn=0x2aaaab10ff79, sig->special=0,
sig->array_opt=0, all_synced=0x7ae2a8,
sig->merge_opt=0, all_sorted=0
interpret_multiplex: fall-through => 'durchfall'
implementation => interpret_multiplex_fcn
interpret_multiplex_fcn
interpret_multiplex_fcn: argv[i].vtype == TYPE_bat &&
argv[i].len != int_nil => real_type = b->htype = 0
interpret_multiplex_fcn: argv[i].vtype == TYPE_bat &&
argv[i].len != int_nil => real_type = b->htype = 0
interpret_multiplex_fcn: desc.nbats == 2, real_type ==
0 => desc.id.vtype = tpe = 6
-----------------
h t name
oid str type
-----------------
[ 0@0, "11" ]
-> quit();
========
$ cvs diff -u src/monet/monet_multiplex.mx
--------
Index: src/monet/monet_multiplex.mx
===================================================================
RCS file:
/cvsroot/monetdb/MonetDB/src/monet/monet_multiplex.mx,v
retrieving revision 1.41.2.2
diff -u -u -r1.41.2.2 monet_multiplex.mx
--- src/monet/monet_multiplex.mx 23 Aug 2005
17:20:47 -0000 1.41.2.2
+++ src/monet/monet_multiplex.mx 2 Sep 2005
16:09:22 -0000
@@ -201,6 +201,8 @@
size_t min_size = REALLY_REALLY_BIG;
char name[128];
+ THRprintf(GDKout, "interpret_multiplex\n");
+
strcpy(name, s);
if (sig->assignop) {
BBPfix(argv[1].val.bval);
@@ -221,6 +223,7 @@
if (min_size ==
REALLY_REALLY_BIG) {
all_synced = b;
tpe = BAThtype(b);
+ THRprintf(GDKout,
"interpret_multiplex: min_size == REALLY_REALLY_BIG =>
tpe=BAThtype(b)=%d (b->htype=%d)\n", tpe, b->htype);
if (sig->assignop &&
b->batRestricted) {
GDKerror("interpret_multiplex: %s: param 1 must be
writeable.\n", name);
return -2;
@@ -276,25 +279,34 @@
@- execution
@c
+ THRprintf(GDKout, "interpret_multiplex: "
+ "sig->pack_fcn="PTRFMT",
sig->fast_fcn="PTRFMT", sig->special=%d,
sig->array_opt=%d, all_synced="PTRFMT",
sig->merge_opt=%d, all_sorted=%d\n",
+ PTRFMTCAST sig->pack_fcn, PTRFMTCAST
sig->fast_fcn, (int)sig->special, (int)sig->array_opt,
PTRFMTCAST all_synced, (int)sig->merge_opt,
(int)all_sorted);
if ((GDKfcn) sig->pack_fcn == sig->fast_fcn) {
/* TOK_PROC */
+ THRprintf(GDKout, "interpret_multiplex:
(GDKfcn) sig->pack_fcn == sig->fast_fcn =>
interpret_multiplex_fcn\n");
return interpret_multiplex_fcn(name,
argc, argv, res, sig->assignop, stk, min_size, tpe,
sig->ret_tpe, (GDKfcn) t);
} else if (sig->special) {
if (sig->array_opt && all_synced) {
+ THRprintf(GDKout,
"interpret_multiplex: sig->special && sig->array_opt &&
all_synced => interpret_special_array\n");
return
interpret_special_array(name, argc, argv, res,
sig->fast_fcn, min_size, sig->ret_tpe, sig->assignop,
sig->special);
}
if (sig->merge_opt && all_sorted) {
+ THRprintf(GDKout,
"interpret_multiplex: sig->special && sig->merge_opt &&
all_sorted => interpret_special_merge\n");
return
interpret_special_merge(name, argc, argv, res,
sig->fast_fcn, min_size, sig->ret_tpe, sig->assignop,
sig->special);
}
} else if (sig->fast_fcn != NULL) {
if (sig->array_opt && all_synced) {
+ THRprintf(GDKout,
"interpret_multiplex: sig->fast_fcn != NULL &&
sig->array_opt && all_synced =>
interpret_multiplex_array\n");
return
interpret_multiplex_array(name, argc, argv, res,
sig->fast_fcn, min_size, sig->ret_tpe, sig->assignop,
FALSE);
}
if (sig->merge_opt && all_sorted) {
+ THRprintf(GDKout,
"interpret_multiplex: sig->fast_fcn != NULL &&
sig->merge_opt && all_sorted =>
interpret_multiplex_merge\n");
return
interpret_multiplex_merge(name, argc, argv, res,
sig->fast_fcn, min_size, sig->ret_tpe, sig->assignop,
FALSE);
}
}
{
/* fall-through => 'durchfall'
implementation */
+ THRprintf(GDKout, "interpret_multiplex:
fall-through => 'durchfall' implementation =>
interpret_multiplex_fcn\n");
return interpret_multiplex_fcn(name,
argc, argv, res, sig->assignop, -1, min_size, tpe,
sig->ret_tpe, sig->pack_fcn);
}
}
@@ -416,6 +428,8 @@
exec_t desc;
RowFcn wrap;
+ THRprintf(GDKout, "interpret_multiplex_fcn\n");
+
memset(&desc, 0, sizeof(exec_t));
desc.name = s;
desc.fcn = imp;
@@ -434,6 +448,7 @@
if (b == NULL)
return 0;
real_type = b->htype;
+ THRprintf(GDKout,
"interpret_multiplex_fcn: argv[i].vtype == TYPE_bat &&
argv[i].len != int_nil => real_type = b->htype =
%d\n", real_type);
desc.piv[desc.nbats] = b;
desc.c[++desc.nbats] =
desc.argv + i;
} else if (argv[i].vtype == TYPE_bat) {
@@ -450,8 +465,10 @@
desc.c[i + 1]->len = ATOMsize(desc.c[i
+ 1]->vtype);
}
if (desc.nbats == 1 && real_type == TYPE_void) {
+ THRprintf(GDKout,
"interpret_multiplex_fcn: desc.nbats == 1 && real_type
== TYPE_void => desc.id.vtype = real_type = %d\n",
real_type);
desc.id.vtype = real_type;
} else {
+ THRprintf(GDKout,
"interpret_multiplex_fcn: desc.nbats == %d, real_type
== %d => desc.id.vtype = tpe = %d\n", desc.nbats,
real_type, tpe);
desc.id.vtype = tpe;
}
desc.id.len = ATOMsize(tpe);
@@ -785,6 +802,8 @@
bulkerror_t bulkerror;
BUN dst;
+ THRprintf(GDKout, "interpret_@1_array\n");
+
/* keep compilers happy */
ah = 0; (void) ah;
hdff = 0; (void) hdff;
@@ -1047,6 +1066,8 @@
bulkerror_t bulkerror;
BAT* retb;
+ THRprintf(GDKout, "interpret_@1_merge\n");
+
@:mergeinit(1)@
@:mergeinit(2)@
if (argc==4)
========
## Comment 3806
Date: 2005-09-05 12:49:40 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed in a different way.
In monet_multiplex, check whether all input BATs are
synchronized (this was already done), and whether there is a
dense input BAT (any one will do). If that is the case, the
result BAT can be VOID-headed.
## Comment 3807
Date: 2005-12-19 16:21:01 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
MonetDB/tests/BugDay_2005-12-19_4.9.3/Tests/void_multiplex.SF-1280802.milS
## Comment 3808
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1280802 at http://sourceforge.net/support/tracker.php?aid=1280802
| multiplex fails to maintain void-heads with var-sized tails | https://api.github.com/repos/MonetDB/MonetDB/issues/678/comments | 0 | 2020-11-30T08:48:31Z | 2024-06-27T11:05:57Z | https://github.com/MonetDB/MonetDB/issues/678 | 753,281,548 | 678 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-02 04:18:54 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-10-05 10:55:28 +0200
## Comment 3801
Date: 2005-09-02 16:18:54 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select cast( 1 as interval year );
!ERROR: interpret: no matching MIL operator
to 'month_interval(sht, int, int)'.
!MAYBE YOU MEAN:
! month_interval(str, int, int) : int
sql>select cast( 1 as interval day );
!ERROR: interpret: no matching MIL operator
to 'second_interval(sht, int, int)'.
!MAYBE YOU MEAN:
! second_interval(str, int, int) : lng
## Comment 3802
Date: 2005-09-05 23:23:25 +0200
From: @njnes
Logged In: YES
user_id=43556
added missing conversion functions
## Comment 3803
Date: 2005-12-19 16:15:53 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/cast_interval.SF.1280682.sql
## Comment 3804
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1280682 at http://sourceforge.net/support/tracker.php?aid=1280682
| SQL: select cast( 1 as interval ... | https://api.github.com/repos/MonetDB/MonetDB/issues/677/comments | 0 | 2020-11-30T08:48:27Z | 2024-06-27T11:05:56Z | https://github.com/MonetDB/MonetDB/issues/677 | 753,281,508 | 677 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-02 01:01:13 +0200
From: @peterboncz
To: Jens Teubner <<teubner>>
Version: -- development
Last updated: 2005-09-08 10:32:12 +0200
## Comment 3791
Date: 2005-09-02 13:01:13 +0200
From: @peterboncz
I admit that I may be in need of a course on XQuery but I
do not understand why the below query:
(: q4 'SELECT alb.Title, ass.Title FROM Assets as ass,
Albums as alb WHERE ass.AlbumId=alb.AlbumId LIMIT 2' :)
let $assets := fn:subsequence(fn:doc("dataset.xml")//Album,
1, 2) cast as xs:anyNode*
let $albums := $assets/parent::*
for $album in $albums
for $asset in $album/Asset
return <Result><Album> { $album/Title }
</Album><Track> { $asset/Title } </Track></Result>
on dataset:
<Music>
<Artist>
<ArtistId>1</ArtistId>
<Name>artist1</Name>
<Album>
<AlbumId>1</AlbumId>
<Title>album1</Title>
<Asset>
<GlobalId>1</GlobalId>
<Title>asset1</Title>
<TrackNr>1</TrackNr>
</Asset>
<Asset>
<GlobalId>2</GlobalId>
<Title>asset2</Title>
<TrackNr>2</TrackNr>
</Asset>
<Asset>
<GlobalId>3</GlobalId>
<Title>asset3</Title>
<TrackNr>3</TrackNr>
</Asset>
<Asset>
<GlobalId>4</GlobalId>
<Title>asset4</Title>
<TrackNr>4</TrackNr>
</Asset>
</Album>
</Artist>
<Artist>
<ArtistId>2</ArtistId>
<Name>artist2</Name>
<Album>
<AlbumId>2</AlbumId>
<Title>album2</Title>
<Asset>
<GlobalId>5</GlobalId>
<Title>asset5</Title>
<TrackNr>1</TrackNr>
</Asset>
<Asset>
<GlobalId>6</GlobalId>
<Title>asset6</Title>
<TrackNr>2</TrackNr>
</Asset>
</Album>
</Artist>
</Music>
produces the error:
$ pf -M 04.xq
syntax error, unexpected "let $" on line 3 (next token is `let
$')
parse error: XQuery parsing failed
halted in ../../../compiler/parser/parser.y (PFparse), line
2239
## Comment 3792
Date: 2005-09-02 13:09:26 +0200
From: @drstmane
Logged In: YES
user_id=572415
Hm, having even less XQuery knowledge, I added a "return"
after the first "let ..." statement, and get:
========
illegal reference to context node: at (4,27-4,32): ``.'' is
unbound
halted in
/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/core/fs.brg
(reduce), line 2796
========
... no idea, whether this helps ...
## Comment 3793
Date: 2005-09-02 13:23:34 +0200
From: @drstmane
Logged In: YES
user_id=572415
Apparently, the assignement to Jens got lost due to
concurrent changes by Peter & me --- SF != ACID |-(
## Comment 3794
Date: 2005-09-02 14:05:10 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
from what i know casting can only be applied to atomic
types. you should use 'treat as' for non-atomic types.
so it should read:
fn:subsequence(...) treat as anyNode*
## Comment 3795
Date: 2005-09-02 14:20:26 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
of course the question is whether anyNode is an atomic type
or not... (i'm starting to having doubts)
## Comment 3796
Date: 2005-09-03 11:42:34 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Pathfinder behaves correctly in Peter's query.
XQuery's `cast as' only allows `SingleType's. That is, only QNames
with optionally a question mark, no star or plus occurrence
operator. Pathfinder will parse the * as a multiplication sign
and the following keyword as a location step. The location step,
however, is relative, without a legally bound context set, thus the
error message.
However, while playing around with your example, I noticed another
("real") bug, probably located in Jan's "simplifier" in
milprint_summer.c. I will file a new bug report for that one.
## Comment 3797
Date: 2005-09-03 16:16:51 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Thanks Jens,
ok, I see. in the below, I iterate over the items and cast
individually, however this fails. It does work with 'treat
as', I noted.
I re-opened this bug to raise another issue. The reason I
was casting was that fn:subsequence returns item* when
invoked on node*. This is contrary to what is mandated by
the formal semantics.
http://www.w3.org/TR/xquery-semantics/sec_fn_subsequence
Maybe I did not correctly specify the signature in
semantics/xquery_fo.c Is there soimething like
type-variables (a la any::1 in MIL) that can specify that
the input type is equal to the oputput type?
this is not only an issue with subsequence, but with many
more operators, as described in the formal semantics doc.
Peter
let $items := fn:subsequence(fn:doc("dataset.xml")//Asset,
1, 100)
let $assets := for $item in $items return $item cast as
xs:anyNode
let $albums := $assets/parent::*
for $album in $albums
for $asset in $album/Asset
where $asset/TrackNr = 1
return <Result><Album> { $album/Title } </Album><Track> {
$asset/Title } </Track></Result>
## Comment 3798
Date: 2005-09-08 10:32:12 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Specific typing rule is now implemented in CVS HEAD.
## Comment 3799
Date: 2005-12-19 16:09:47 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/cast_as_anyNode.SF-1280519.xq
I'm assuming the result (a syntax error) is correct.
## Comment 3800
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1280519 at http://sourceforge.net/support/tracker.php?aid=1280519
| cast as anyNode* | https://api.github.com/repos/MonetDB/MonetDB/issues/676/comments | 0 | 2020-11-30T08:48:24Z | 2024-06-28T13:34:48Z | https://github.com/MonetDB/MonetDB/issues/676 | 753,281,476 | 676 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-02 11:22:01 +0200
From: @yzchang
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-09-05 08:57:59 +0200
## Comment 3787
Date: 2005-09-02 11:22:01 +0200
From: @yzchang
This link doesn't work
(http://monetdb.cwi.nl/TechDocs/Core/gdk/index.html):
MonetDB Kernel API and data structures.
[an error occurred while processing this directive]
## Comment 3788
Date: 2005-09-02 13:00:19 +0200
From: @drstmane
Logged In: YES
user_id=572415
Indeed. Thanks for reporting it.
The problem is, that texi2html was nolonger available on the
new machines,
and hence the respective documents were not created during
the nightly testing's "make docs".
I installed texi2html on my machine, and will update the
website as soon as next night's testing has finished.
## Comment 3789
Date: 2005-12-19 16:00:18 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / DOCUMENTATION
## Comment 3790
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1280452 at http://sourceforge.net/support/tracker.php?aid=1280452
| Kernel guide not avialable | https://api.github.com/repos/MonetDB/MonetDB/issues/675/comments | 0 | 2020-11-30T08:48:22Z | 2024-06-27T11:05:54Z | https://github.com/MonetDB/MonetDB/issues/675 | 753,281,442 | 675 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-09-01 03:26:36 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-09-02 11:18:03 +0200
## Comment 3783
Date: 2005-09-01 15:26:36 +0200
From: @drstmane
[Unless compiled with debugging explicitely enabled
(`configure --enable-debug`) and hence optimization
disabled (i.e., `gcc -g [-O0]` iso. the default `gcc -g
-O2`),]
the latest version of compiler/mil/milprint_summer.c
(revision 1.89) fails to compile with:
========
make[3]: Entering directory
`/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..BUILD./PATHFINDER/compiler/mil'
/bin/sh ../../libtool --tag=CC --mode=compile gcc -m64
-DHAVE_CONFIG_H -I.
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil
-I../..
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil
-I../include
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil/../include
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/common
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/gdk
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/monet
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/plain
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/contrib
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/mapi
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/C
-DLIBMIL -g -O2 -std=c99 -Wall -W
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label -c -o
libmil_la-milprint_summer.lo `test -f
'milprint_summer.c' || echo
'/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil/'`milprint_summer.c
gcc -m64 -DHAVE_CONFIG_H -I.
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil
-I../..
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil
-I../include
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil/../include
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/common
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/gdk
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/monet
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/plain
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/contrib
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/mapi
-I/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..PREFIX./include/MonetDB/C
-DLIBMIL -g -O2 -std=c99 -Wall -W
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label -c
/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil/milprint_summer.c
-fPIC -DPIC -o .libs/libmil_la-milprint_summer.o
cc1: warnings being treated as errors
/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil/milprint_summer.c:
In function 'translate2MIL':
/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/pathfinder/compiler/mil/milprint_summer.c:7750:
warning: 'rc' is used uninitialized in this function
make[3]: *** [libmil_la-milprint_summer.lo] Error 1
make[3]: Leaving directory
`/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..BUILD./PATHFINDER/compiler/mil'
make[2]: *** [install] Error 2
make[2]: Leaving directory
`/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..BUILD./PATHFINDER/compiler/mil'
make[1]: *** [install-recursive] Error 1
make[1]: Leaving directory
`/net/corona.ins.cwi.nl/export/scratch0/manegold/Monet/Testing/Current/.GNU.64.64.d..BUILD./PATHFINDER/compiler'
make: *** [install-recursive] Error 1
========
## Comment 3784
Date: 2005-09-02 11:17:50 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed by Peter.
## Comment 3785
Date: 2005-12-19 15:59:45 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
## Comment 3786
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1279145 at http://sourceforge.net/support/tracker.php?aid=1279145
| Pathfinder (milprint_summer.c) fails to compile | https://api.github.com/repos/MonetDB/MonetDB/issues/674/comments | 0 | 2020-11-30T08:48:19Z | 2024-06-27T11:05:53Z | https://github.com/MonetDB/MonetDB/issues/674 | 753,281,411 | 674 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-30 10:16:41 +0200
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-09-05 08:06:37 +0200
## Comment 3778
Date: 2005-08-30 22:16:41 +0200
From: Wouter Alink <<alink>>
The nightly testing results seem not accessible...
clicking on any 'o' or 'X' returns the following error:
Not Found
The requested URL
/projects/monetdb/Stable/pathfinder/.ntv.64.64.d-Fedora3/mTests0/compiler/core/tests/fs/.good-004.out.html
was not found on this server.
Apache/2.0.51 (Fedora) Server at crux.ins.cwi.nl Port 80
## Comment 3779
Date: 2005-08-30 22:34:40 +0200
From: @drstmane
Logged In: YES
user_id=572415
Wouter,
this is actually more a "feature" than a "bug":
As stated on (e.g.)
http://monetdb.cwi.nl/Development/TestWeb/Stable/pathfinder/index.html
, the "Mtest.py -r -d0" test runs (i.e., with GDKdebug=0)
are only executed during weekends (i.e., on Saturday &
Sunday); only the "make check" test runs (i.e., with
GDKdebug=10) are run every night (also during weekends).
The reason is that executing both test runs simply takes too
long (at least on some old & slow machines in our test-farm)...
Hence,
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests03/ntv.64.64.d-Fedora3/compiler_core_tests_fs/good-004.out.00.html
is not available today (Tuesday), while
http://monetdb.cwi.nl/testing/projects/monetdb/Stable/pathfinder/.mTests103/ntv.64.64.d-Fedora3/compiler_core_tests_fs/good-004.out.00.html
is available.
Actually, later this week, I might have a look, whether it
can patch our "TestTool" scripts to keep the last weekend's
results for the "Mtest.py -r -d0" test runs during the whole
week...
## Comment 3780
Date: 2005-09-02 13:01:36 +0200
From: @drstmane
Logged In: YES
user_id=572415
I (hopefully) fixed our TestTools.
Next week will show, whether it works as expected/desired.
## Comment 3781
Date: 2005-12-19 15:58:59 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / DOCUMENTATION
## Comment 3782
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1276925 at http://sourceforge.net/support/tracker.php?aid=1276925
| TestWeb: details not available | https://api.github.com/repos/MonetDB/MonetDB/issues/673/comments | 0 | 2020-11-30T08:48:16Z | 2024-06-27T11:05:52Z | https://github.com/MonetDB/MonetDB/issues/673 | 753,281,370 | 673 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-28 04:04:16 +0200
From: Ethan Aubin <<eaubin>>
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-08-30 01:05:31 +0200
## Comment 3773
Date: 2005-08-28 16:04:16 +0200
From: Ethan Aubin <<eaubin>>
I've a query which fails printing:
GDKmmap(333250560) fail => BBPtrim(enter)
usage[mem=29092948,vm=246087680]
mallinfo.arena = 31477760
BBPTRIM_ENTER: memsize=29092948,vmsize=246087680
BBPTRIM: memtarget=0 vmtarget=1073741824
TRIMSCAN: mem=0 vm=1, start=1, limit=1
TRIMSCAN: 38170 0=doc_sema (0)
TRIMSCAN: 38268 1=attr_own1 (0)
TRIMSCAN: 38269 2=attr_qn1 (0)
TRIMSCAN: 38270 3=attr_prop1 (0)
.... (rest of failure is in attached file).
I've narrowed the problem in the query to:
declare function local:check-function($f as
element(c:function))
as element(c:function)*
{
let $target :=
<c:target id="{ $f/c:name/text() }"
category="Component"/>
return
if (some $performs-target
in $db/c:relation/c:performs/c:target
satisfies $target = $performs-target)
then
()
else
$f
};
<badfunctions>{
for $f in $db/c:element/c:function
return
local:check-function($f)
}</badfunctions>
I've attached my config.log file. Is there any other
information I could provide would be useful to the
developers? - EA
## Comment 3774
Date: 2005-08-30 01:05:13 +0200
From: @peterboncz
Logged In: YES
user_id=591107
Hi,
What happens is that MonetDB simply gets out of resources.
The GDKmmmap() tells that this call for more memory fails on
a request for 333MB of RAM.
It also tells that the memsize is set to about 30MB and it
is at that point already using 246 MB of virtual memory.
This kind of thing is a limitation of the execution model
the makes MonetDB fast (when it works). Other database
systems in these situations justs get awfully slow; MonetDB
throws in the towel.
What I take from this is that you are running a 32-bit
operating system with not too much memory installed. If you
have a 64-bits system and a lot of RAM, MonetDB can take you
much further. But I actually doubt that this is your
problem....
In any case, I suspect that you would not have been content
with a very slow running query either. So, the best solution
is to modify your query in such a way that it gets to
benefit from the strengths of database tehcnology in general.
Basically, you ar submitting a query in a java/C programming
style. The optimizer cannot spot the join pattern and it
resorts to generating a so-called cartesian product. This
leads to O(N*N) complexity, were a O(N) or O(Nlog(N)
solution is possible.
so.. maybe something along these lines works better:
fn:distinct-values(
for $f in $db/c:element/c:function
for $t in $db/c:relation/c:performs/c:target
[@category="Component"]
where $t@id = $f/c:namr/text()
return $f)
this query uses a nested for-loop united by an simple
equality condition in the where clause. MonetDB can detect
this particular pattern as a join.
Maybe it won't work if you split the nested loop in two
parts, the inner being a user defined function. But I
actually think it should work as well.
## Comment 3775
Date: 2005-08-30 11:22:38 +0200
From: @drstmane
Logged In: YES
user_id=572415
Ethan,
did you try Peter's rephrased query?
Does it work for you?
How much main-memory does your system have, and how big is
your XML document (filesize of the .xml file, and in case
you know, number of c:target and c:function nodes).
Thanks in advance,
Stefan
## Comment 3776
Date: 2005-12-19 15:58:18 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / NOT A BUG
## Comment 3777
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1275140 at http://sourceforge.net/support/tracker.php?aid=1275140
| GDKmmap fails executing query | https://api.github.com/repos/MonetDB/MonetDB/issues/672/comments | 0 | 2020-11-30T08:48:13Z | 2024-06-28T13:34:48Z | https://github.com/MonetDB/MonetDB/issues/672 | 753,281,349 | 672 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-28 02:51:52 +0200
From: Ethan Aubin <<eaubin>>
To: Jens Teubner <<teubner>>
Version: -- development
Last updated: 2005-08-31 03:06:21 +0200
## Comment 3768
Date: 2005-08-28 02:51:52 +0200
From: Ethan Aubin <<eaubin>>
I've a query which fails with:
type error: no variant of function
pf:merge-adjacent-text-nodes accepts the given
argument type(s): item*
though there is no mention of
merge-adjacent-text-nodes. E.g.
declare namespace local = "urn:local";
declare variable $db := doc("big.xml")/database;
declare function local:relations($ignore)
{
for $r in $db/relation
return $r
};
<html>{
local:relations($db)
}</html>
Any ideas what going on? - EA
## Comment 3769
Date: 2005-08-28 10:24:49 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Hi Ethan,
functions whose QName start with a `pf' are internal functions of our
compiler (`pf' is for `Pathfinder', and the hash mark keeps you from
explicitly calling these functions in the surface language). These functions
(and their names) should never be visible to you, sorry for that confusion.
I will also look into this as soon as I find the time.
What might help you as a workaround (though I haven't tried), is to assign
types to your function and its parameter.
## Comment 3770
Date: 2005-08-31 15:06:21 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Fixed in CVS (stable branch).
The specific typing rule for item-sequence-to-node-sequence() did not
correctly deal with the type `item'. The function in Ethan's example lacks a
type declaration and is thus treated as `item*' during static typing.
Ethan, you may either checkout the fix from CVS (stable branch, tag
XQuery_0-8), or add a type declaration to your function declaration as a
workaround. Thanks again for the bug report.
## Comment 3771
Date: 2005-12-19 15:55:29 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/type_error_in_function_resolution.SF-1274903.xq
## Comment 3772
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1274903 at http://sourceforge.net/support/tracker.php?aid=1274903
| no variant of function #pf:merge-adjacent-text-nodes accepts | https://api.github.com/repos/MonetDB/MonetDB/issues/671/comments | 0 | 2020-11-30T08:48:10Z | 2024-06-27T11:05:50Z | https://github.com/MonetDB/MonetDB/issues/671 | 753,281,310 | 671 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-27 11:19:37 +0200
From: Ethan Aubin <<eaubin>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-08-29 05:05:39 +0200
## Comment 3759
Date: 2005-08-27 23:19:37 +0200
From: Ethan Aubin <<eaubin>>
Hi, Type declarations on functions don't seem to be
working the way I expect: E.g.
Say I'm using schema and data from the usecase STRONG
1.9.4.2 Q2 at
http://www.w3.org/TR/xquery-use-cases/strong.
I.e. the schema is:
<?xml version="1.0" encoding="UTF-8"?>
<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"
elementFormDefault="qualified"
targetNamespace="http://www.example.com/zips"
xmlns:zips="http://www.example.com/zips">
<xs:element name="zips">
<xs:complexType>
<xs:sequence>
<xs:element maxOccurs="unbounded" ref="zips:row"/>
</xs:sequence>
</xs:complexType>
</xs:element>
<xs:element name="row">
<xs:complexType>
<xs:sequence>
<xs:element ref="zips:city"/>
<xs:element ref="zips:state"/>
<xs:element ref="zips:zip"/>
</xs:sequence>
</xs:complexType>
</xs:element>
<xs:element name="city" type="xs:string"/>
<xs:element name="state" type="xs:string"/>
<xs:element name="zip" type="xs:integer"/>
</xs:schema>
and the data is:
<zips xmlns="http://www.example.com/zips">
<row>
<city>Old Town</city>
<state>PA</state>
<zip>95819</zip>
</row>
<row>
<city>Durham</city>
<state>NC</state>
<zip>27701</zip>
</row>
<row>
<city>Durham</city>
<state>NC</state>
<zip>27703</zip>
</row>
</zips>
And I want to get a list of the city nodes. galax lets
me use the following xquery:
declare namespace local = "urn:local";
import schema namespace z =
"http://www.example.com/zips" at "zips.xsd";
declare function local:cities($r as element(z:zips)) as
element(z:city)*
{
$r/z:row/z:city
};
<cities>
{
local:cities(doc("zips.xml")/z:zips)
}
</cities>
to get:
<cities xmlns:z="http://www.example.com/zips"
xmlns:local="urn:local">
<city xmlns="http://www.example.com/zips">Old Town</city>
<city xmlns="http://www.example.com/zips">Durham</city>
<city xmlns="http://www.example.com/zips">Durham</city>
</cities>
but monetdb fails on this same query printing:
type error: no variant of function local:cities accepts
the given argument type(s): element z:zips { item* }*
Is a problem with the location hint, or the schema
import? Thanks - EA
## Comment 3760
Date: 2005-08-28 10:19:05 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Thanks again for your report, Ethan.
I have to admit that our schema import feature is not really well maintained
(nor thoroughly tested).
I cannot make any promises for the upcoming week (We're in Trondheim
on VLDB to present our engine). But I will try to look into this as soon as
possible.
## Comment 3761
Date: 2005-08-28 13:44:00 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Sorry, Ethan, but the error message is expected behavior.
Your problem has nothing to do with schema import. Your import schema
statment mainly delared the namespace z, the query itself did not
reference any types from the loaded XML schema file.
The problem is that the return type of the path expression `doc("zips.xml")/
z:zips' is element z:zips { item* }*. Note the star at the end, the expression
returns zero or more elements named z:zips. Your function, however, only
accepts arguments of type element z:zips { item* } (that is, exactly one
element with name z:zips). Your query is thus not type-safe, and the
Pathfinder compiler complains.
The reason why you can run the query in Galax is probably that Galax (at
least by default) does not implement XQuery's static type feature.
Pathfinder implements this feature, which makes it appear very picky
about types. I think Galax provides static typing via some additional
command line switch (though as far as I remember that was known to be
broken).
There are two possible fixes for your query:
(a) Tell Pathfinder that the path expression really returns exaclty one
element. The `exactly-one()' function does that:
local:cities (exactly-one (doc ("zips.xml")/z:zips))
(b) Let your function accept sequences of elements:
declare function local:cities($r as element(z:zips)*) as
element(z:city)*
(note the star after the function argument)
I will re-read the specs in Trondheim. I'm pretty sure though, that
Pathfinder behaves correctly here.
(Can somebody close that bug? For some reason I cannot.)
## Comment 3762
Date: 2005-08-28 15:10:35 +0200
From: Ethan Aubin <<eaubin>>
Logged In: YES
user_id=522388
Yikes, how did I miss that? Your explaination makes perfect
sense. By default Galax does catch _some_ type errors but
this gives me less confidence in it. Is there another
implementation you use to sanity check your results? (I'd be
happy to run my scripts on a different implementation or
two, so I only send actual bug reports :-)
Also maybe the error message could print the expected type
along with the used type? The "variant of" part also
confused me, my knee jerk reaction was that there was some
ad-hoc function overloading going on somewhere. I'm guessing
its a reminder of the implicit conversions on atomic types...
Is there any talk of distributing pathfinder with a
testsuite or the xquery usecases?
## Comment 3763
Date: 2005-08-28 15:36:46 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Hi Ethan,
don't hesitate to bother us with bug reports. If it turns out to be a "feature",
don't worry. As for now, your ratio is still not too bad: two out of three are
actual bugs :-).
Your question for other implementations is a good one. I don't investigate
on other implementations on a regular basis, and typically I don't have the
time to really look into the details. Particularly the typing issues in XQuery
are rather tricky, and it is not surprising that implementations have their
problems with it. (Pathfinder does have its problems as well, let's see if
you find them... :-).) One implementation that I like pretty much on the first
view with respect to typing was the online demo at http://www.
cogneticsystems.com/xquery/xquery.html. But I haven't really tested that
one either.
You're right that XQuery does not allow function overloading on the surface
language (well, you may redefine a function with same name, but different
number of arguments). Several built-in functions, however, are
overloaded. Our typechecker does implements some overloading
functionality to deal with those. (You are still not allowed to overload
functions on the surface level, though.)
So that is where the "variant of" comes from. But I admit that our error
messages have some room for improvement...
I'm not really sure what you mean with your question on a test suite. For
development purposes we actually have a HUGE set of test queries that
we run every night through automated scripts. The XPathMark test suite,
for instance, is part of that. If somebody really finds a lot of time, we might
as well add the use cases to our test setup.
## Comment 3764
Date: 2005-08-28 16:26:43 +0200
From: Ethan Aubin <<eaubin>>
Logged In: YES
user_id=522388
Re: The testsuite. I like a small test suite on my machine,
so that I can check that tools on my box did nothing
unexpected during the
build. Its also the best way to see what behavior can be
expected to work right out of the box. Are the nightly test
scripts available
online?
Re: The usecases. They might be nice just to check
cross-implementation consistency. E.g. have the nightly
test queries been ran by a different xquery processor?
> (well, you may redefine a function with same name, but
> different number of arguments).
Wow, I didn't know that. I learn something new everyday.
Cheers - EA
## Comment 3765
Date: 2005-08-29 17:05:39 +0200
From: @mlkersten
Logged In: YES
user_id=490798
On behalf of Jens it is closed
## Comment 3766
Date: 2005-12-19 16:00:50 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: NO TEST / INVALID
## Comment 3767
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1274819 at http://sourceforge.net/support/tracker.php?aid=1274819
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| "type error: no variant of function ... accepts the the giv" | https://api.github.com/repos/MonetDB/MonetDB/issues/670/comments | 0 | 2020-11-30T08:48:07Z | 2024-06-28T13:34:47Z | https://github.com/MonetDB/MonetDB/issues/670 | 753,281,287 | 670 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-27 06:01:30 +0200
From: Ethan Aubin <<eaubin>>
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-08-27 09:19:57 +0200
## Comment 3748
Date: 2005-08-27 18:01:30 +0200
From: Ethan Aubin <<eaubin>>
Hi, The books1.xq example at
http://monetdb.cwi.nl/XQuery/Demo/Books/index.html
fails when run printing
bad usage of XML namespaces: at (1,1-8,1): unknown
namespace in qualified function name local:toc
FWIW, galax seems parse and type the file correctly.
Also the URL for books.xml is different in books2.xq
and the html displayed for books2.xq (missing the
files/ directory).
## Comment 3749
Date: 2005-08-27 18:45:28 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
Hello Ethan,
Thank you for your bug-report. I tried the books1.xq, but
could not reproduce the error you describe. Could you
please provide us with some more information (Operating
system/Version of MonetDB/XQuery?) If you use
MonetDB/XQuery 0.8.0, would you please upgrade to
MonetDB/XQuery 0.8.2 as many bugs have been fixed
since 0.8.0.
Thanks in advance,
Wouter
## Comment 3750
Date: 2005-08-27 19:18:48 +0200
From: Ethan Aubin <<eaubin>>
Logged In: YES
user_id=522388
The README, NEWS, ChangeLog file are empty, but the
directory and config.status file seems to indicate
pathfinder 0.8.2 is being used with MonetDB version 4.8.2.
This is on a debian linux box w/ kernel 2.6.11
If I add
declare namespace local = "urn:local"; to the start of the
file, then is works like a champ.
BTW, Should add a bug report requesting a --version command
line switch?
## Comment 3751
Date: 2005-08-27 21:05:59 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
ah, i see why i couldn't reproduce your bug. you meant
books2.xq (somehow i thought the error was in books1.xq)
I fixed the query books2.xq (and book.xml) on the website
(removed the "local:" several times in the query, and
disconnected the dtd from the xml-file). Note: it will probably
take a few days before the website will show the new files.
The "--version" flag can be translated in "cat VERSION", but
maybe you're right and is it a missing feature. But this is not
my area... (Sjoerd?)
Thanks again for your bug report and for using
MonetDB/XQuery. If you have any other comments or
suggestions, we would be keen to hear from you.
p.s. could somebody please mark this bug fixed (i don't
have those privileges)
## Comment 3752
Date: 2005-08-27 21:19:57 +0200
From: @mlkersten
Logged In: YES
user_id=490798
Closed upon request
## Comment 3753
Date: 2005-08-27 21:43:51 +0200
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Sorry guys, you closed that bug too quickly. Ethan actually found a bug in
the compiler. The webpage was correct (we should have tested those
queries before publishing, though).
Our namespace checker did not predefine the namespaces `local' and
`xdt'. I have committed a fix to the CVS stable branch.
Ethan, thanks for the report (and sorry for the confusion). You will have to
either checkout my fixed version from CVS (stable branch, tag XQuery_0-
8), or wait for the next nightly CVS snaphot.
As a workaround, you can either remove the `local' prefixes in your query,
or define the namespace directly in your query (via `declare namespace').
The --version option is actually a great suggestion. We will think about how
to elegantly incorporate something like that in our release cycle.
## Comment 3754
Date: 2005-08-27 22:56:52 +0200
From: Ethan Aubin <<eaubin>>
Logged In: YES
user_id=522388
I'm very impressed by how responsive you've all been! I'm an
xquery novice and found monet when frustrated with the speed
of galax. It would take 10 minutes for a simple queries on
a small 21meg db, which monetdb runs in a matter of seconds!
Its blindingly fast! The XQuery front end does seem to be a
bit quirky wrt syntax and typing. I'll post a new bug report
about a typing problem in a few minutes.
BTW, I'm also looking forward to when xquery modules are
implemented.
## Comment 3755
Date: 2005-08-27 23:39:28 +0200
From: @grobian
Logged In: YES
user_id=963970
out of curiosity, what 'modules' are you looking for?
## Comment 3756
Date: 2005-08-28 00:11:36 +0200
From: Ethan Aubin <<eaubin>>
Logged In: YES
user_id=522388
MM: Just plain old XQuery modules. E.g.
mymod.xq:
module namespace my = "urn:some-functions";
declare function my:one() { 1 };
main.xq:
import module namespace my = "urn:some-functions" at "mymod.xq";
my:one() + my:one()
## Comment 3757
Date: 2005-12-19 15:56:27 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB
Web site examples are tested in
pathfinder/tests/WebSite/Tests/
## Comment 3758
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1274649 at http://sourceforge.net/support/tracker.php?aid=1274649
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| functions and XML namespaces | https://api.github.com/repos/MonetDB/MonetDB/issues/669/comments | 0 | 2020-11-30T08:48:04Z | 2024-06-27T11:05:48Z | https://github.com/MonetDB/MonetDB/issues/669 | 753,281,255 | 669 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-26 04:49:56 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-09-01 09:58:47 +0200
## Comment 3745
Date: 2005-08-26 16:49:56 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select extract( year from interval '1' year);
[ -1 ]
sql>select extract( day from interval '1' day );
[ 22 ]
sql>select extract( day from interval '1 01' day to hour);
[ 31 ]
sql>select extract( hour from interval '1 01' day to hour);
ERROR = !types sec_interval(64,0) (lng) and time(0,0)
(daytime) are not equal
I have no idea what's going on.
## Comment 3746
Date: 2005-12-19 15:50:40 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
sql/src/test/bugs/Tests/extract_interval_bug-sf-1274088.*
## Comment 3747
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1274088 at http://sourceforge.net/support/tracker.php?aid=1274088
| SQL: select extract( ... from interval ... | https://api.github.com/repos/MonetDB/MonetDB/issues/668/comments | 0 | 2020-11-30T08:48:01Z | 2024-06-27T11:05:47Z | https://github.com/MonetDB/MonetDB/issues/668 | 753,281,224 | 668 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-26 04:44:07 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-09-01 09:59:05 +0200
## Comment 3741
Date: 2005-08-26 16:44:07 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select cast('1-2' as interval year to month );
[ 1 ]
I'd expect 14. Looks like cast() takes only leading
digits (up to the hyphen) into account.
## Comment 3742
Date: 2005-08-30 14:16:04 +0200
From: @njnes
Logged In: YES
user_id=43556
intervals now keep their qualifiers (like year to month)
information in the
types digit and scale (ugh)
this is needed for dynamic type conversions (from string),
which is added to
the convert code generation
## Comment 3743
Date: 2005-12-19 15:50:08 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
sql/src/test/bugs/Tests/interval_convert_bugs-sf-1274077-1274085.*
## Comment 3744
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1274085 at http://sourceforge.net/support/tracker.php?aid=1274085
| SQL: select cast('1-2' as interval year to month ) | https://api.github.com/repos/MonetDB/MonetDB/issues/667/comments | 0 | 2020-11-30T08:47:58Z | 2024-06-27T11:05:46Z | https://github.com/MonetDB/MonetDB/issues/667 | 753,281,191 | 667 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-26 04:35:02 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-09-01 09:59:18 +0200
## Comment 3737
Date: 2005-08-26 16:35:02 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select interval '1' year = interval '12' second;
[ true ]
Sure, the underlying atomic types compare equal, but
the intervals are incomparable.
From the SQL standard:
Year-month intervals are comparable only with other
year-month intervals.
Day-time intervals are comparable only with other
day-time intervals.
## Comment 3738
Date: 2005-08-30 14:14:59 +0200
From: @njnes
Logged In: YES
user_id=43556
conversions between interval are nolonger supported
## Comment 3739
Date: 2005-12-19 15:49:27 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
sql/src/test/bugs/Tests/interval_convert_bugs-sf-1274077-1274085.*
## Comment 3740
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1274077 at http://sourceforge.net/support/tracker.php?aid=1274077
| SQL: select interval '1' year = interval '12' second | https://api.github.com/repos/MonetDB/MonetDB/issues/666/comments | 0 | 2020-11-30T08:47:56Z | 2024-06-27T11:05:46Z | https://github.com/MonetDB/MonetDB/issues/666 | 753,281,170 | 666 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-23 04:41:23 +0200
From: @MarcinZukowski
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-08-23 08:23:45 +0200
## Comment 3732
Date: 2005-08-23 16:41:23 +0200
From: @MarcinZukowski
It seems pump operations do not set the 'hkey' properly
in all the situations, here's the script:
var x:=new(oid,int);
x.insert(20@0,88);
x.insert(21@0,89); hkey=1
var z:=[int](x.slice(0,1)); slice with unique keys,
hkey set
{min}(z).info().find("hkey").print();
[ "1" ]
x.insert(21@0,0); hkey=0
var z:=[int](x.slice(0,1)); slice with unique keys,
hkey not set
{min}(z).info().find("hkey").print();
[ "0" ]
var z:=[int](x.slice(0,2)); slice with duplicates,
hkey not set
{min}(z).info().find("hkey").print();
[ "1" ]
I think pump "by definition" should set hkey to true so
there seems to be a bug.
## Comment 3733
Date: 2005-08-23 19:22:29 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS; see below for details.
Marcin,
could you please verify, whether your bug is indeed fixed,
and then close this bug report?
Thanks,
Stefan
The pump implementation (interpret_setaggr) correctly set
the key property for the head of the result.
However, if the result was the same size as the input
(i.e., the input'\''s head was actually key/unique),
the code is "clever" and calls ALIGNsetH to propagate
the properties of the input'\''s head to the result'\''s head;
in case the input'\''s head was not marked key (though it was),
ALIGNsetH overwrote the previously correctly set key property
of the result'\''s head.
I fixed this, be setting the key property of the result'\''s
head
only after the ALIGNsetH call.
## Comment 3734
Date: 2005-08-23 20:23:45 +0200
From: @MarcinZukowski
Logged In: YES
user_id=607094
After both changes it seems to work again.
Thank you, Stefan!
## Comment 3735
Date: 2005-12-19 15:48:17 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: TEST ADDED / SUCCESS:
MonetDB/tests/BugDay_2005-12-19_4.9.3/Tests/pump_properties.SF-1267155.*
## Comment 3736
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1267155 at http://sourceforge.net/support/tracker.php?aid=1267155
| Pump properties | https://api.github.com/repos/MonetDB/MonetDB/issues/665/comments | 0 | 2020-11-30T08:47:53Z | 2024-06-27T11:05:44Z | https://github.com/MonetDB/MonetDB/issues/665 | 753,281,135 | 665 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-23 03:25:52 +0200
From: @sjoerdmullender
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2007-02-25 06:14:39 +0100
## Comment 3723
Date: 2005-08-23 15:25:52 +0200
From: @sjoerdmullender
In quite a few places in the code (certainly MonetDB,
possibly others) the symbol BUFSIZ is used to allocate
memory. The problem with this is that BUFSIZ may be
too small on certain systems for the intended purposes.
For example: On Windows, BUFSIZ is defined as 512, and
this is too small for a decent-sized monet_mod_path
value (actually encountered when trying to run Mserver
from the build directory using RunMserver.bat).
We should go look at all occurences of BUFSIZ to see if
they are ok if the value is relatively small.
The problem mentioned could perhaps be solved if
mo_substitute were to allocate memory in which it
returns the substituted value instead of substituting
in-place.
## Comment 3724
Date: 2005-12-19 15:39:56 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: NO TEST / CODING TECHNIQUE
## Comment 3725
Date: 2006-01-15 15:00:40 +0100
From: @grobian
Logged In: YES
user_id=963970
any news on this?
## Comment 3726
Date: 2006-01-15 15:53:34 +0100
From: @mlkersten
Logged In: YES
user_id=490798
An assessment without actions yet.
upgrade.mx : OK
str.mx: OK
pcrc.mx/prce_compile_wrap, pcre_select, pcre_uselect,
pcre_replace: possible buffer overflow problems on
pcre_compile calls
logger.mx/logger_open: PROBLEM, BUFSIZ->PATHLENGTH
logger.mx/logger_new: idem
logger.mx/logger_exit: idem
logger.mx/logger_clean: idem
ascii_io.mx: OK
gdk_desc.mx: Dangerous, double check
gdk_posix.mx: OK
gdk_utils.mx:OK
## Comment 3727
Date: 2006-01-20 11:11:05 +0100
From: @mlkersten
Logged In: YES
user_id=490798
gdk_desc.mx: not a real issue as long as atom type names are
less then 512
logger.mx: not serious and protected against overflow
only logger.new might cause a problem at some point.
pcre.mx: doesn't seem a big problem, unless pcre has very
long built in error messages. Most likely not.
## Comment 3728
Date: 2006-01-20 11:29:15 +0100
From: @mlkersten
Logged In: YES
user_id=490798
gdk_desc.mx: not a real issue as long as atom type names are
less then 512
logger.mx: not serious and protected against overflow
only logger.new might cause a problem at some point.
pcre.mx: doesn't seem a big problem, unless pcre has very
long built in error messages. Most likely not.
## Comment 3729
Date: 2006-06-06 20:56:35 +0200
From: @grobian
Logged In: YES
user_id=963970
Does this thing need some more search/replace work, or can
we close it?
## Comment 3730
Date: 2007-02-25 18:14:39 +0100
From: @mlkersten
Logged In: YES
user_id=490798
Originator: NO
Visited once more before closing:
Stable, modules, plain uses it in asii-io, str, pcre, upgrade. No serious threads
Stable, monetdb and gdk all gone.
Stable, modules/mal, no harm
Stable, optimizer, one additional protection added
Stable, clients, seem protected
## Comment 3731
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1267093 at http://sourceforge.net/support/tracker.php?aid=1267093
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| BUFSIZ considered harmful | https://api.github.com/repos/MonetDB/MonetDB/issues/664/comments | 0 | 2020-11-30T08:47:50Z | 2024-06-27T11:05:43Z | https://github.com/MonetDB/MonetDB/issues/664 | 753,281,098 | 664 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-23 01:33:06 +0200
From: @drstmane
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-12-25 11:29:36 +0100
## Comment 3711
Date: 2005-08-23 13:33:06 +0200
From: @drstmane
When executing XQuery queries via the "xquery(str,str)"
MIL command,
error messages that occur during the evaliadtion of the
query get lost/swallowed. When executing XQuery queries
via "MapiClient -lxquery" the respective error messaged
do appeatr as expected.
Most probably, this is caused by the fact that
CMDxquery() in runtime/pathfinder.mx redirects GDKout
(and thus GDKerr) to a buffer to capture the output of
exec_xquery().
However, me knowledge of MonetDB/MIL error message
handling is apparently not sufficient to solve this
problem...
========
$ Mserver --dbinit='module(pathfinder); mapi_start();
pfstart();'
--------
Monet Database Server V4.9.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>xquery("xml","doc(\"abcde.xml\")").printf();
--------
> Missing:
> !ERROR: shredder.mx:fileSize: cannot stat
`abcdef.xml': No such file or directory
> !ERROR: CMDshred2bats: operation failed.
--------
!ERROR: CMDxquery: operation failed.
!ERROR: interpret_params: printf(param 1): evaluation
error.
MonetDB>
========
$ echo 'doc("abcdef.xml")' | MapiClient -lx
--------
ERROR: shredder.mx:fileSize: cannot stat `abcdef.xml':
No such file or directory
$
========
## Comment 3712
Date: 2005-08-23 13:49:35 +0200
From: @yzchang
Logged In: YES
user_id=341633
Hi Stefan,
I thought this was lost by the changes in monet.mx? Didn't
it help?
## Comment 3713
Date: 2005-08-23 13:53:27 +0200
From: @drstmane
Logged In: YES
user_id=572415
Nope. The fixes in monet.mx only solved the problem that the
error state got lost completely. Now, at least the
last/top-level error message "!ERROR: CMDxquery: operation
failed." appears, again; but the actual reason for this
failure (i.e., "!ERROR: shredder.mx:fileSize: cannot stat
`abcdef.xml': No such file or directory") is still lost...
## Comment 3714
Date: 2005-11-18 19:14:20 +0100
From: @drstmane
Logged In: YES
user_id=572415
added test scripts as
tests/BugsViaSourgeforce/Tests/ID.1267018.swallowed_error_messages-[12].*
## Comment 3715
Date: 2005-11-24 23:59:07 +0100
From: @peterboncz
Logged In: YES
user_id=591107
this is fixed now.
## Comment 3716
Date: 2005-12-19 15:39:01 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
pathfinder/tests/BugsViaSourgeforce/Tests/ID.1267018.swallowed_error_messages*
## Comment 3717
Date: 2005-12-21 00:21:34 +0100
From: @drstmane
Logged In: YES
user_id=572415
re-opened as the (proper) error messages seem to get
swallowed, again:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora4/tests_BugsViaSourgeforce/ID.1267018.swallowed_error_messages-2.out.00.html
## Comment 3718
Date: 2005-12-23 18:20:43 +0100
From: @peterboncz
Logged In: YES
user_id=591107
has been fixed
## Comment 3719
Date: 2005-12-23 18:35:10 +0100
From: @drstmane
Logged In: YES
user_id=572415
sorry for being so (too?) "picky", but IMHO, the fix has
"disappeared" again:
in
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_BugsViaSourgeforce/ID.1267018.swallowed_error_messages-2.out.00.html
I would expext the original error message
!ERROR: I/O warning : failed to load external entity "abcde.xml"
but all that appears is
!ERROR: interpret: no matching MIL operator to 'printf(void)'.
checking monet's testing logs, is seems as if the bug was
fixed on Nov 24 or Nov 25, but reappeared due to chenges
made on Dec 02 ... this should help to locate which changes
made the bug re-appear ...
(I'm currently not online enough to check this myself, sorry)
A fix is not "urgend", but I re-open the report, again, to
keep *us* reminded.
## Comment 3720
Date: 2005-12-23 18:36:21 +0100
From: @drstmane
Logged In: YES
user_id=572415
oops, sorry, small typo:
the changes that made the bug re-appear must have been on
Dec 01 ...
## Comment 3721
Date: 2005-12-25 11:29:36 +0100
From: @drstmane
Logged In: YES
user_id=572415
(re-)fixed by Peter.
Thank you very much for the instant action!
## Comment 3722
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1267018 at http://sourceforge.net/support/tracker.php?aid=1267018
| XQuery: xquery() swallows error message | https://api.github.com/repos/MonetDB/MonetDB/issues/663/comments | 0 | 2020-11-30T08:47:46Z | 2024-06-27T11:05:42Z | https://github.com/MonetDB/MonetDB/issues/663 | 753,281,058 | 663 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-22 08:13:30 +0200
From: @drstmane
To: @yzchang
Version: -- development
Last updated: 2005-08-23 09:48:50 +0200
## Comment 3707
Date: 2005-08-22 08:13:30 +0200
From: @drstmane
On our SunOS 5.8 (Solaris 8) machine "apps", pathfinder
fails to compile:
========
[...]
gcc -m64 -DHAVE_CONFIG_H -I.
-I/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime
-I..
-I/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/common
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/gdk
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/monet
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/plain
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/contrib
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/mapi
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/C
-I/ufs/acoi/local/SunOS/64/include/libxml2
-I../compiler/include
-I/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/../compiler/include
-DLIBPATHFINDER -O2 -std=c99 -D__EXTENSIONS__ -Wall -W
-O2 -fomit-frame-pointer -finline-functions
-DHWCOUNTERS -DHW_SunOS -DHW_sun4u
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label
-Wno-uninitialized -DEMBEDDED -DNO_CGI -DNO_AUTH -c
pathfinder.c -fPIC -DPIC -o
.libs/lib_pathfinder_la-pathfinder.o
/bin/bash ../libtool --mode=compile gcc -m64
-DHAVE_CONFIG_H -I.
-I/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime
-I..
-I/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/common
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/gdk
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/monet
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/plain
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/contrib
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/mapi
-I/var/tmp/monet.GNU.64.64.d.24482/MonetDB/.GNU.64.64.d-SunOS5.8/include/MonetDB/C
-I/ufs/acoi/local/SunOS/64/include/libxml2
-I../compiler/include
-I/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/../compiler/include
-DLIBPF_SUPPORT -O2 -std=c99 -D__EXTENSIONS__ -Wall -W
-O2 -fomit-frame-pointer -finline-functions
-DHWCOUNTERS -DHW_SunOS -DHW_sun4u
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label
-Wno-uninitialized -DEMBEDDED -DNO_CGI -DNO_AUTH -c -o
lib_pf_support_la-pf_support.glue.lo `test -f
'pf_support.glue.c' || echo
'/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/'`pf_support.glue.c
/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/pathfinder.mx:
In function `setup_connection':
/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/pathfinder.mx:1133:
syntax error before '.' token
/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/pathfinder.mx:1139:
`S_un' undeclared (first use in this function)
/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/pathfinder.mx:1139:
(Each undeclared identifier is reported only once
/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/pathfinder.mx:1139:
for each function it appears in.)
/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/pathfinder.mx:1142:
`err' undeclared (first use in this function)
/var/tmp/monet.GNU.64.64.d.24482/pathfinder/runtime/pathfinder.mx:1158:
`sock' undeclared (first use in this function)
gmake[3]: *** [lib_pathfinder_la-pathfinder.lo] Error 1
[...]
========
See
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.GNU.64.64.d-SunOS5.8/make.out
for details.
## Comment 3708
Date: 2005-08-23 09:48:50 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Solaris has define s_addr _S_un._S_addr, so s_addr cannot
be used as variable name.
## Comment 3709
Date: 2005-12-19 15:38:21 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: NO TEST / COMPILATION
## Comment 3710
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1265933 at http://sourceforge.net/support/tracker.php?aid=1265933
| pathfinder fails to compile on SunOS/Solaris | https://api.github.com/repos/MonetDB/MonetDB/issues/662/comments | 0 | 2020-11-30T08:47:43Z | 2024-06-27T11:05:41Z | https://github.com/MonetDB/MonetDB/issues/662 | 753,281,031 | 662 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-21 12:08:45 +0200
From: @Teggy
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2005-09-12 03:34:57 +0200
## Comment 3703
Date: 2005-08-21 12:08:45 +0200
From: @Teggy
The creation of element node content seems to exhibit a
problem.
Works OK:
<a> { (1,2,3) } </a> --> <a>1 2 3</a>
<a> { for $x in (1,2,3) return $x } </a> --> <a>1 2 3</a>
Fails:
<a> { for $x in (1,2,3) where $x gt 0 return $x }
</a> --> <a/> (: empty element :)
This seems odd to me (note that the 'where' predicate
lets all bindings of $x pass).
## Comment 3704
Date: 2005-09-12 15:34:57 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
The translation recognized a relational join. Its internal
core representation was of type 'none' which was recognized
as subtype of type 'node*'.
Fixed both return type of relational join and subtype
relationship of type 'none'.
## Comment 3705
Date: 2005-12-19 15:37:58 +0100
From: @drstmane
Logged In: YES
user_id=572415
BugDay_2005-12-19, stmane: ALREADY IN TESTWEB:
pathfinder/tests/BugsViaSourgeforce/Tests/ID.1265413.element_creation_and_content*
## Comment 3706
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1265413 at http://sourceforge.net/support/tracker.php?aid=1265413
| Element creation and element content | https://api.github.com/repos/MonetDB/MonetDB/issues/661/comments | 0 | 2020-11-30T08:47:41Z | 2024-06-27T11:05:40Z | https://github.com/MonetDB/MonetDB/issues/661 | 753,281,003 | 661 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-19 04:18:30 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-08-30 02:03:42 +0200
## Comment 3699
Date: 2005-08-19 16:18:30 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select interval '-1-2' year to month;
[ -10 ]
I'd expect -14, i.e. the first minus sign is a monadic
operator for the whole literal.
## Comment 3700
Date: 2005-08-22 16:29:56 +0200
From: @njnes
Logged In: YES
user_id=43556
Parsing of the minus sign is now done globably instead of
only in the first
part of the interval string.
## Comment 3701
Date: 2005-12-19 15:41:24 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: ALREADY IN TESTWEB
sql/src/test/bugs/interval_parse_bug-sf-1263967.sql
## Comment 3702
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1263967 at http://sourceforge.net/support/tracker.php?aid=1263967
| SQL: select interval '-1-2' year to month | https://api.github.com/repos/MonetDB/MonetDB/issues/660/comments | 0 | 2020-11-30T08:47:38Z | 2024-06-27T11:05:40Z | https://github.com/MonetDB/MonetDB/issues/660 | 753,280,977 | 660 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-19 03:14:33 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-08-30 02:04:07 +0200
## Comment 3695
Date: 2005-08-19 15:14:33 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select date '2005-08-19' + interval '1' day;
[ 2005-08-20 ]
works, but
sql>select interval '1' day + date '2005-08-19';
doesn't work:
ERROR = !binary operator sql_add(date,date) unknown
Fortunately, the standard allowed these elements to be
commutative under addition:
<http://savage.net.au/SQL/sql-2003-
2.bnf.htmldatetime value expression>
## Comment 3696
Date: 2005-08-22 16:31:14 +0200
From: @njnes
Logged In: YES
user_id=43556
Commutative functions are resolved both ways if needed now.
## Comment 3697
Date: 2005-12-19 15:40:48 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: ALREADY IN TESTWEB
sql/src/test/bugs/interval_data_commutative_add_bug-sf-1263928.sql
## Comment 3698
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1263928 at http://sourceforge.net/support/tracker.php?aid=1263928
| SQL: select interval ... + date ... | https://api.github.com/repos/MonetDB/MonetDB/issues/659/comments | 0 | 2020-11-30T08:47:35Z | 2024-06-27T11:05:39Z | https://github.com/MonetDB/MonetDB/issues/659 | 753,280,938 | 659 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-19 03:09:49 +0200
From: Jens Teubner <<teubner>>
To: @yzchang
Version: -- development
Last updated: 2005-08-19 04:41:16 +0200
## Comment 3690
Date: 2005-08-19 15:09:49 +0200
From: Jens Teubner <<teubner>>
Compiling the Pathfinder runtime gives me
pathfinder.mx: In function `CMDxquery':
pathfinder.mx:963: warning: 'dst' might be used
uninitialized in this function
Indeed, the return value of CMDxquery is undefined if
exec_xquery() fails:
/* run the query, but collect all output in a buffer */
THRsetdata(0,s);
ret = exec_xquery(mode, xquery);
THRsetdata(0,o);
stream_close(s);
if (ret == GDK_SUCCEED) {
dst = GDKmalloc(b->pos+1);
memcpy(dst, b->buf, b->pos);
dst[b->pos] = 0;
}
buffer_destroy(b);
stream_destroy(s);
*res = dst;
return ret;
(Btw., indentation seems to be messed up in this file.)
## Comment 3691
Date: 2005-08-19 16:41:16 +0200
From: @yzchang
Logged In: YES
user_id=341633
sorry, my compiler is not strict enough:(
## Comment 3692
Date: 2005-08-22 08:27:11 +0200
From: @yzchang
Logged In: YES
user_id=341633
sorry, my compiler is not strict enough:(
## Comment 3693
Date: 2005-12-19 15:39:54 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
## Comment 3694
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1263926 at http://sourceforge.net/support/tracker.php?aid=1263926
| compilation fails in pathfinder.mx | https://api.github.com/repos/MonetDB/MonetDB/issues/658/comments | 0 | 2020-11-30T08:47:32Z | 2024-06-27T11:05:38Z | https://github.com/MonetDB/MonetDB/issues/658 | 753,280,903 | 658 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-19 03:03:39 +0200
From: @sjoerdmullender
To: Peter Boncz <<boncz>>
Version: -- development
Last updated: 2005-11-25 02:26:47 +0100
## Comment 3686
Date: 2005-08-19 15:03:39 +0200
From: @sjoerdmullender
The third argument of BATfetchjoin (estimate) is
declared as size_t. However inside it compares the
value with oid_nil. An oid is not necessarily the same
as size_t, so this could possibly go wrong. We should
check this.
The same holds for the other functions that take an
estimate.
## Comment 3687
Date: 2005-11-25 02:26:47 +0100
From: @peterboncz
Logged In: YES
user_id=591107
well, I think that size_t is a datatype with a range always
bigger than oid, so I don't think this is a problem.
## Comment 3688
Date: 2005-12-19 15:39:10 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST
## Comment 3689
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1263921 at http://sourceforge.net/support/tracker.php?aid=1263921
| BATfetchjoin takes size_t arg for oid | https://api.github.com/repos/MonetDB/MonetDB/issues/657/comments | 0 | 2020-11-30T08:47:29Z | 2024-06-28T13:34:46Z | https://github.com/MonetDB/MonetDB/issues/657 | 753,280,864 | 657 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-19 10:48:17 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-11-26 06:06:17 +0100
## Comment 3682
Date: 2005-08-19 10:48:17 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
create table t( c time );
insert into t values( time '12:34:56');
select c from t where extract( hour from c ) = 12 and
extract( minute from c ) = 34;
gives an empty result set:
3 querytype
t table_name
c name
time type
8 length
0 tuplecount
12 id
## Comment 3683
Date: 2005-11-26 18:06:17 +0100
From: @njnes
Logged In: YES
user_id=43556
The optimization (reducing selects) found the base columns
equal but incorrectly concluded the functions didn't alter the
column.
Added test script
src/test/BugTracker/Tests/select_extract.SF-1263796.sql
## Comment 3684
Date: 2005-12-19 15:38:03 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: ALREADY IN TESTWEB
sql/src/test/BugTracker/Tests/select_extract.SF-1263796.sql
## Comment 3685
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1263796 at http://sourceforge.net/support/tracker.php?aid=1263796
| SQL: select ... where extract( ...) = ... and extract( ... | https://api.github.com/repos/MonetDB/MonetDB/issues/656/comments | 0 | 2020-11-30T08:47:26Z | 2024-06-27T11:05:35Z | https://github.com/MonetDB/MonetDB/issues/656 | 753,280,827 | 656 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-18 05:31:18 +0200
From: @yzchang
To: @sjoerdmullender
Version: -- development
Last updated: 2006-01-16 02:39:48 +0100
## Comment 3678
Date: 2005-08-18 17:31:18 +0200
From: @yzchang
In stream.mx:
the implementation of socket_read returns a '0' to
indicate no data have been read.
This is not necessarily an error, but the
implementation still sets errnr to READ_ERROR:
s->errnr = READ_ERROR;
I think this line code should be removed.
## Comment 3679
Date: 2005-12-19 15:36:35 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST
Probably the change should be made and we should see what
fails then.
## Comment 3680
Date: 2006-01-16 14:39:48 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
Checked in as revision 1.121 of src/common/stream.mx.
## Comment 3681
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1263146 at http://sourceforge.net/support/tracker.php?aid=1263146
| errnr set without occurence of an error | https://api.github.com/repos/MonetDB/MonetDB/issues/655/comments | 0 | 2020-11-30T08:47:23Z | 2024-06-27T11:05:35Z | https://github.com/MonetDB/MonetDB/issues/655 | 753,280,793 | 655 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-18 03:55:26 +0200
From: @yzchang
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-08-18 10:39:33 +0200
## Comment 3674
Date: 2005-08-18 15:55:26 +0200
From: @yzchang
In stream.mx:
the implementation of socket_read returns a '0' to
indicate no data have been read.
This is not necessarily an error, but the
implementation still sets errnr to READ_ERROR:
s->errnr = READ_ERROR;
I think this line code should be removed.
## Comment 3675
Date: 2005-08-18 22:39:33 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Duplicate of http://sourceforge.net/support/tracker.php?aid=1263146.
## Comment 3676
Date: 2005-12-19 15:34:04 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / REDUNDANT
## Comment 3677
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1263042 at http://sourceforge.net/support/tracker.php?aid=1263042
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
This bug was marked DUPLICATE in the database it was moved from.
Changing resolution to "MOVED"
| errnr set without occurence of an error | https://api.github.com/repos/MonetDB/MonetDB/issues/654/comments | 0 | 2020-11-30T08:47:20Z | 2024-06-28T13:34:46Z | https://github.com/MonetDB/MonetDB/issues/654 | 753,280,755 | 654 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-17 09:53:55 +0200
From: @sjoerdmullender
To: Jan Flokstra <<jflokstra>>
Version: -- development
Last updated: 2005-09-07 02:49:45 +0200
## Comment 3666
Date: 2005-08-17 09:53:55 +0200
From: @sjoerdmullender
Shred a document (e.g. <a><b>c</b><d><e/><f/></d><g
a="42"/></a>) and then look at the produced bats (with
e.g.
print(bat("pre_size0"),bat("pre_level0").[+](0),bat("pre_prop0"),bat("pre_kind0").[+](0));
). Do this on a 64-bit machine. The result show an
interesting value in the prop column:
MonetDB>print(bat("pre_size0"),bat("pre_level0").[+](0),bat("pre_prop0"),bat("pre_kind0").[+](0));
---------------------------------------------------------
h t t t t
name
void int int oid int
type
---------------------------------------------------------
[ 0@0, 10, -1, 18446744071562067968@0, 4 ]
[ 1@0, 8, 0, 0@0, 0 ]
[ 2@0, 2, 1, 1@0, 0 ]
...
Looking more closely at the value:
18446744071562067968 == 0xffffffff80000000.
Some investigation revealed:
The value comes from a (nat) value nil which is passed
as an argument to direct_oid2bat (in
shredder.mx:emitnode). The corresponding argument (the
fourth) of direct_oid2bat is an oid. So here there is
an automatic type conversion from nat to oid. The type
nat is defined as int near the top of shredder.mx.
There is separate code for 32 and 64 bit systems, but
they are currently identical.
Note that the value nat (int) nil is represented as
0x80000000 and the type conversion causes a sign
extension to 64 bits, resulting in the observed value.
I experimented with chaning the type definition of nat to
if (SIZEOF_OID==4)
/* 32 bits systems */
typedef int nat;
define NIL ((nat) INT_MIN)
define NATFMT "u"
else
typedef unsigned long long nat;
define NIL ((nat) LLONG_MIN)
ifdef HAVE___INT64
define NATFMT "I64u"
else
define NATFMT "llu"
endif
However this gave too many errors in the tests.
Note that the if test is also different from the
original. The original has
if (SIZEOF_VOID_P==4)
whereas I have
if (SIZEOF_OID==4)
The comment (elided here) says that the type of a nat
has to correspond to some other type in a bat, but it
is totally unclear to me which type that should be, and
so I wasn't able to check that.
So two bugs really:
- nat -> oid conversion is not correct, especially the
value of nil is not converted to oid_nil.
- the comments near the typedef of nat are incomplete.
## Comment 3667
Date: 2005-08-17 11:09:46 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Repaired the bug by changing the function
direct_oid2bat(..., oid v) to direct_natoid2bat(.... nat_v).
I now check if this nat_v == NIL and if so I insert
'oid_nil' in the BAT. This should solve the problem.
I cannot really test it myself because we have only a
working 64bit monet with 32 bit oid's.
When this all works I think I will remove ALL 32/64 bit
ifdefs because in the current implementation it should work
without these. They were inherited from the old loader by
Torsten.
## Comment 3668
Date: 2005-08-17 11:42:22 +0200
From: @drstmane
Logged In: YES
user_id=572415
Maybe this is also the right time to re-think the usage of
"nat" at all.
Wouldn't is be more natural in a MonetDB context to stick to
the proper MonetDB data types (chr, sht, int, lng, oid)?
... and btw, there's a simple way to get a working 64bit
MonetDB with 64bit oid: just compile it ;-)
(I'm "juggling" with more than 4(!) different MonetDB
versions on a daily basis...)
(if you need access to a 64-bit machine, just let me know!)
## Comment 3669
Date: 2005-08-17 12:26:18 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Last time I tried the 64bit-arch, 64bit oid version it
crashed. But I tried again on our new 'teehuis' and now it
worked.
The fix works, there's a correct [0@0, nil] tuple in
bat("pre_prop0") and printing the doc again also succeeds.
It complains BAT[ATTR_PROP] is not dense but this is another
problem I think. I will have a look at this too.
I will also have a look at the eventual conversion of nat to
oid.
## Comment 3670
Date: 2005-09-05 08:10:07 +0200
From: @drstmane
Logged In: YES
user_id=572415
Any news on this one?
## Comment 3671
Date: 2005-09-07 14:49:44 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
Fixed and closed. Removed 32/64bit arch ifdefs and removed
confusing comments. Furthermore made provisions for the
conversion of nat->oid but did not use this solution by
default because a lot of nat's are counters and not oid's.
Maybe in the future we should retype them into posint's.
## Comment 3672
Date: 2005-12-19 15:32:53 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/nat_oid_conv.SF-1261765.milS
## Comment 3673
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1261765 at http://sourceforge.net/support/tracker.php?aid=1261765
| problematic nat -> oid conversion on 64 bit system | https://api.github.com/repos/MonetDB/MonetDB/issues/653/comments | 0 | 2020-11-30T08:47:17Z | 2024-06-27T11:05:33Z | https://github.com/MonetDB/MonetDB/issues/653 | 753,280,720 | 653 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-16 12:59:06 +0200
From: Jens Teubner <<teubner>>
To: @sjoerdmullender
Version: -- development
Last updated: 2006-01-24 08:52:40 +0100
## Comment 3651
Date: 2005-08-16 12:59:06 +0200
From: Jens Teubner <<teubner>>
Building Pathfinder currently requires a recent version of bison to be
available on the build system (>=1.33, I think). Some platforms (e.g.,
Mac OS X) don't provide this by default.
I think the idea in the autotools is that users are not required to have
flex/bison installed on their system at all. Tarballs should ship with
pre-built .c files, processed with flex/bison before making the tarball.
Flex/bison is only required if the corresponding .l/.y files change, i.e.,
on developers' machines.
I tested simply setting EXTRA_DIST in the respective Makefile.ag
files. This results in the .c files being shipped in the tarball, and
Pathfinder compiles fine even without flex/bison. On `make clean',
however, the .c files are deleted, and a subsequent build without
flex/bison will fail.
My understanding would be that .c files built from flex/bison sources
should be shipped with distribution tarballs, and they should not be
deleted by `make clean'. This way, Pathfinder would easily compile
without flex/bison.
The ./configure script, by the way, tests for the availability of flex/
bison. If a tool is not found, the corresponding variable will be unset.
This is (intended) default behaviour of the autotools (as flex/bison
are not really requirements anyway). However, it leads to misleading
error messages when building Pathfinder (The `d' is the empty
variable for the bison executable, plus the `-d' option for bison.):
Making all in burg
lockfile -r 2 waiting
d ../../dl/pathfinder-0.8.2/burg/gram.y || { rm -f waiting ; exit 1 ; }
/bin/sh: line 1: d: command not found
AFAIK, Fabian stumbled on the same problem, when preparing
ebuilds for OS X.
## Comment 3652
Date: 2005-08-16 14:25:54 +0200
From: @grobian
Logged In: YES
user_id=963970
I "solved" this in the Gentoo ebuild by unpacking an
additional tarball with the generated bison sources. This
works great on OSX, but isn't the nicest solution of course.
The additional tarball is uploaded on the Gentoo mirrors and
only used for OSX users on Gentoo for now.
It would be great if the sources supplied these files by
default, as well as a "make clean-bison-only" or alike
command for users that like to generate the sources
themselves...
## Comment 3653
Date: 2005-11-21 16:25:23 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Seems like it is even worse than I had written it in August.
Apparently, MonetDB's ./configure does *not* check for the
presence of bison/yacc. If nothing is found, $YACC is simply
set to 'yacc', resulting in
yacc -d mel.yy || { rm -f waiting ; exit 1 ; }
/bin/sh: line 1: yacc: command not found
make[3]: *** [mel.tab.h] Error 1
I think we should find a means to compile MonetDB (and
Pathfinder) without lex/flex/yacc/bison installed on the
system. Most software packages I know install fine without
them. If this is not possible, a proper ./configure error
is a must, IMHO.
Btw. I left "Group" at "Pathfinder CVS". However, this
affects *all* MonetDB packages, CVS and stable.
## Comment 3654
Date: 2005-11-21 16:39:32 +0100
From: @drstmane
Logged In: YES
user_id=572415
I completely agree with you, Jens.
I still "dream" of a source distribution (tarball and/or
src.rpm) where all preprocessing has been done, and hence,
neither lex/flex/yacc/bison not the buildtools are required
to compile from that source distribution.
Though this might not be achieved easily, I propose to have
this done (or found evidence that it is not possible) before
the next release (i.e., no release until all bug reports
including this one are properly closed) --- details about
the upcoming release plans will be discussed here at CWI
tomorrow or on Wednesday...
## Comment 3655
Date: 2005-11-21 17:56:34 +0100
From: @grobian
Logged In: YES
user_id=963970
I can redo the trick I did for Gentoo (in fact I will have
to ;) ) and make the sources available on our sourceforge
stuff here, then let Gentoo mirror from there. The
"additional sources" are now solely hosted on the Gentoo
mirrors.
## Comment 3656
Date: 2005-11-21 18:27:06 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Thanks Fabian for your offer.
However, I guess you are realizing yourself that such
"hacks" are not the right way how we should address this
problem, and you'd have to redo that hack on every release/
installation. Users outside the MonetDB developer community
might run into the same problem (in fact, chances that non-
developers don't have the respective tools installed are
even higher).
I will find ways to get MonetDB installed here, don't worry.
But I would still prefer a proper solution in order to make
the installation hassle-free for potential MonetDB users.
Of course, I'd be willing to help. My knowledge on the
MonetDB build process is close to zero, but I could help, e.
g., with testing here in Munich.
Anyway, thanks for caring so much about me :-)
## Comment 3657
Date: 2005-11-21 19:12:02 +0100
From: @grobian
Logged In: YES
user_id=963970
Let me put it like this:
From where I'm sitting what you propose it far from what
MonetDB should support. The problem of bison here (that I
indeed solved with a very dirty hack on OSX) is directly
related to the OS/distro that is being installed on. A
regular Linux system won't have any problems with satisfying
the dependencies for XQuery or any other MonetDB product.
The problem is in the rather crappy stock install Apple
delivers to the user. Solaris might be the same problem.
If bison was not the problem, then the autotools would have
been.
I deal with many many packages, and to be honest, I don't
know of any upstream that includes generated bison/yacc
files. The most they do is simply make their stuff
backwards compatible whenever this is easily possible (often
done with python for True/False I think). There is a good
reason for not including such generated bison files, since a
newer version of bison might be able to do smarter things,
or fix bugs.
While diving into this and actually trying to get the most
ancient archs compiling without any help is a quite generous
thing to do, I believe it draws a far to big burden on the
MonetDB project (probably Stefan...). For architectures
that are problematic, we should just have nightly builds
available. IMHO, MonetDB would be anything but the first
project to just require some decent recent versions of the
used tools.
As practically for OSX:
Fink and DarwinPorts are able to supply more uptodate
versions of the required tools.
Gentoo can build MonetDB using the little cheat I made, but
in the future it will also not need this for OSX, because at
this very moment I have access to a prefixed install that
allows me to install those new tools, like Fink does. For
any other arch such as Solaris: in the end Gentoo for
Solaris will work there too.
Binary installations are for (simple) users, source
installations are for the advanced users (that should be
able to deal with the problem here). Feel free to disagree ;)
## Comment 3658
Date: 2005-11-21 22:54:37 +0100
From: @drstmane
Logged In: YES
user_id=572415
Well, of course there are various solutions.
The minimal one is to properly document that MonetDB
requires (among others) at least bison 1.35 to compile from
the (CVS) sources, and to properly check for this in
configure, and abort compilation in case no proper bison is
found. Then it's the task of who ever tries to compile
MonetDB to ensure that his/her machine meats the requirements.
For the CVS sources, there is no other option.
On our MacOS X testing machine, I simply installed
bison-1.35 in my home --- works even simpler than installing
MonetDB: configure; make; make install ;-)
The ebuild (for MacOS X, only?) should then also depend on a
bison >=1.35 ebuild --- if there is none, yet, we could
consider making one --- cannot be more difficult than making
an ebuild for MonetDB, right?
For the source distributions (tarball, src.rpm), we should
try to make them compile without lex/flex/bison/yacc, and
also without the buildtools (Mx, mel, burg); this would also
simplify documentation, packaging and cross-compiling.
The major hurdel I sees right now is that we cannot have
.mx.in files any more; if/where possible, we should to turn
them into pure .mx files, i.e., eliminate the modifications
done by configure; where/if this doesn't work, we have to
think of another solution (turning .mx.in into .in.mx is
obviously no solution).
Actually, thinking about this, the cleanest solution for all
the (platform independent?) pre-processing (buildtools,
lex/flex/bison/yacc) would be to do it during the bootstrap;
then, the "clean" source distributions are for free ...
## Comment 3659
Date: 2005-11-22 09:35:21 +0100
From: @grobian
Logged In: YES
user_id=963970
> On our MacOS X testing machine, I simply installed
> bison-1.35 in my home --- works even simpler than installing
> MonetDB: configure; make; make install ;-)
> The ebuild (for MacOS X, only?) should then also depend on a
> bison >=1.35 ebuild --- if there is none, yet, we could
> consider making one --- cannot be more difficult than making
> an ebuild for MonetDB, right?
Forget about the ebuild. Dependencies as you describe are
handled like that. I only cheated for OSX because it
otherwise would block on bison, which cannot be installed
since the OS already provides this. This problem is being
solved, so the only relevant part is the source installations.
I argue that for source installations, you could simply let
configure die when it does not find what it wants to have.
Your "clean" source installation idea is pretty much cool
since you don't need the buildtools for it, but I doubt
whether you can get it running. What I remember from last
time we tried to do it together is that we stumbled upon so
many problems that it wasn't fun any more (hence we were not
able to ship updated sources to the students).
## Comment 3660
Date: 2005-11-22 09:49:46 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
> I argue that for source installations, you could simply
> let configure die when it does not find what it wants to
> have. Your "clean" source installation idea is pretty
> much cool since you don't need the buildtools for it, but
> I doubt whether you can get it running.
I don't have any clue about Mx. But for lex/yacc, I read in
the automake manual:
``The intermediate files generated by `yacc' (or `lex') will
be included in any distribution that is made. That way the
user doesn't need to have `yacc' or `lex'.''
So in some way, this functionality appears to be readily
provided by automake. At least for Pathfinder, we don't have
any Mx generated lex/yacc files. So I think I am not
"dreaming" when I ask for a "clean" distribution.
If I understand correctly, lex/yacc need not be called
during bootstrapping. Instead, the `dist' make target will
take care of building the required .c files and put them
into the source tarball.
I will look into this right now...
## Comment 3661
Date: 2005-11-22 11:41:28 +0100
From: Jens Teubner <<teubner>>
Logged In: YES
user_id=731390
Ok, I've now looked into the Pathfinder sources, where it
turned out to be rather simple to remove the dependency on
lex/yacc for users. For now, I have edited Makefile.am files
manually; I'm an absolute novice in Python and autogen,
where a proper fix needed to be made.
The only files affected are scanner.l and parser.y in
compiler/parser. So first I have cleaned up the Makefile.am.
I have replaced the *_SOURCES lines with a simple
libparser_la_SOURCES = scanner.l parser.y abssyn.c
and eliminated *all* explicitly listed dependencies. I've
also removed the BUILT_SOURCES, MOSTLYCLEANFILES, and
EXTRA_DIST variables (though we might still want to fill the
latter with Makefile.ag and Makefile.msc).
Then I stumbled over a minor problem: The lexer depends on
the file parser.tab.h, generated by yacc/bison. The automake
way to handle this is to add the respective header file to
BUILT_SOURCES:
BUILT_SOURCES = parser.h
Apparently, automake prefers `parser.h' over `parser.tab.h',
so I renamed it accordingly in scanner.l.
Pathfinder currently does have a parser.h in its include
directory. To avoid confusion and problems with include
paths, I simply removed include/parser.h and added its
declarations to abssyn.h (and adapted include commands in
source files accordingly).
After running autoheader, automake, and autoconf, `make
dist' will add the generated sources to the distribution
tarball, and everything compiles fine without lex/yacc.
I did not have the `make clean' problem mentioned in the
original bug report. (I did, however, discover a different
problem with .m files in the distribution tarball on the
SourceForge page. If ./configure is called within the source
directory, a `make distclean' apparently deletes a bit too
much. A subsequent ./configure ; make fails with a missing
pathfinder.m.)
So apart from the necessary modification of autogen, does
anything else hinder us from removing the dependency on
lex/yacc?
## Comment 3662
Date: 2005-12-19 15:16:37 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
I'd say this is actually an RFE. not a bug.
## Comment 3663
Date: 2006-01-22 19:39:35 +0100
From: @drstmane
Logged In: YES
user_id=572415
While we managed to get all other buildtools out of the
source distribution, there was/is no time left to do (and
double-check!) the same for yacc/bison/[f]lex before this
release.
Postponed until after the release.
## Comment 3664
Date: 2006-01-24 20:52:40 +0100
From: @grobian
Logged In: YES
user_id=963970
Stefan did a better job than he thought himself. It just
works and looks as if this idea is implemented. Closing bug.
## Comment 3665
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1260796 at http://sourceforge.net/support/tracker.php?aid=1260796
| Building without bison >=1.33 | https://api.github.com/repos/MonetDB/MonetDB/issues/652/comments | 0 | 2020-11-30T08:47:14Z | 2024-06-27T11:05:31Z | https://github.com/MonetDB/MonetDB/issues/652 | 753,280,690 | 652 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-15 10:04:06 +0200
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2006-02-07 11:57:26 +0100
## Comment 3646
Date: 2005-08-15 10:04:06 +0200
From: @grobian
On Gentoo, during the nightly testing, the following
assertion is raised for several tests:
Mserver: dl-open.c:438: dl_open_worker: Assertion
`imap->l_type == lt_loaded' failed.
## Comment 3647
Date: 2005-12-19 15:13:29 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / REDUNDANT
## Comment 3648
Date: 2006-01-15 16:23:47 +0100
From: @drstmane
Logged In: YES
user_id=572415
last night, the bug occured with `make check` (aka `Mtest.py
-d10`) only for one test:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Gentoo1.6.13/tests_XQuery/fun7.err.00.html
with `Mtest.py -d0`, it occured last night for 4 tests:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/sql/.mTests03/GNU.32.32.d-Gentoo1.6.13/src_benchmarks_arno/create.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/sql/.mTests03/GNU.32.32.d-Gentoo1.6.13/src_test_BugDay_2005-10-06_2.9.3/rollback_and_autocommit.SF-926453.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests03/GNU.32.32.d-Gentoo1.6.13/tests_BugsViaSourgeforce/ID.1206985.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests03/GNU.32.32.d-Gentoo1.6.13/tests_XQuery/instance1.err.00.html
I suppose, though, the list of test for which this bug
happens is quite random...
## Comment 3649
Date: 2006-02-07 23:57:26 +0100
From: @njnes
Logged In: YES
user_id=43556
This has to do with the random crashes bug. So closing this
one. Because of the 'exit without cleanup' this is also gone.
## Comment 3650
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1259668 at http://sourceforge.net/support/tracker.php?aid=1259668
This bug was marked DUPLICATE in the database it was moved from.
Changing resolution to "MOVED"
| Testing: Gentoo testing sometimes asserts in Mserver | https://api.github.com/repos/MonetDB/MonetDB/issues/651/comments | 0 | 2020-11-30T08:47:10Z | 2024-06-28T13:34:45Z | https://github.com/MonetDB/MonetDB/issues/651 | 753,280,645 | 651 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-12 12:44:28 +0200
From: @Teggy
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2005-09-12 03:36:56 +0200
## Comment 3642
Date: 2005-08-12 12:44:28 +0200
From: @Teggy
Currently, this query fails at runtime (while it shoud
not):
let $a :=
<a>
<b c="foo"/>
<d/>
</a>
return $a/(d|b/@c)
ERROR: cannot handle attributes in function
pf:distinct-doc-order.
The problem seems to be that pf:distinct-doc-order is
unable to order the 'c' attribute node with respect to
the 'd' element node (the XPath child axis step is
required to return its result in document order).
According to the XQuery spec, attribute nodes
immediately follow their owner elements in document
order (although the order of attributes sharing the
same owner is undefined). Clearly, attribute node 'c'
should precede element node 'd' here.
Can/should we modify pf:distinct-doc-order such that
attributes are ordered according to the document order
rank of their owner elements?
Nice weekend all,
--Teggy
## Comment 3643
Date: 2005-09-12 15:36:56 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Added code which copes with attributes exactly as described
in the bug report (order on iter|frag|pre|attr).
## Comment 3644
Date: 2005-12-19 14:52:15 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: ALREADY IN TESTWEB
pathfinder/tests/BugsViaSourgeforce/Tests/ID.1257463.distinct-doc-order.xq
## Comment 3645
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1257463 at http://sourceforge.net/support/tracker.php?aid=1257463
| Document order and attribute nodes | https://api.github.com/repos/MonetDB/MonetDB/issues/650/comments | 0 | 2020-11-30T08:47:07Z | 2024-06-27T11:05:29Z | https://github.com/MonetDB/MonetDB/issues/650 | 753,280,601 | 650 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-11 05:19:05 +0200
From: @yzchang
To: @sjoerdmullender
Version: -- development
Last updated: 2005-08-12 04:41:47 +0200
## Comment 3638
Date: 2005-08-11 17:19:05 +0200
From: @yzchang
Failed to compile the last version of shredder.mx.
Got the following error during "make":
gcc32 -m64 -DHAVE_CONFIG_H -I. -I../../runtime -I..
-I../../runtime
-I/ufs/zhang/install/monetdb/include/MonetDB
-I/ufs/zhang/install/monetdb/include/MonetDB/common
-I/ufs/zhang/install/monetdb/include/MonetDB/gdk
-I/ufs/zhang/install/monetdb/include/MonetDB/monet
-I/ufs/zhang/install/monetdb/include/MonetDB/plain
-I/ufs/zhang/install/monetdb/include/MonetDB/contrib
-I/ufs/zhang/install/monetdb/include/MonetDB/mapi
-I/ufs/zhang/install/monetdb/include/MonetDB/C
-I/usr/include/libxml2 -I../compiler/include
-I../../runtime/../compiler/include -DLIBPF_SUPPORT
-std=c99 -Wall -W -g
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label -DEMBEDDED
-DNO_CGI -DNO_AUTH -c shredder.c -fPIC -DPIC -o
.libs/lib_pf_support_la-shredder.o
../../runtime/shredder.mx: In function `pf_compress':
../../runtime/shredder.mx:414: implicit declaration of
function `BATins_kunique'
cc1: warnings being treated as errors
../../runtime/shredder.mx:414: warning: assignment
makes pointer from integer without a cast
make[2]: *** [lib_pf_support_la-shredder.lo] Error 1
make[2]: Leaving directory
`/ufs/zhang/src/pathfinder/build/runtime'
make[1]: *** [install] Error 2
make[1]: Leaving directory
`/ufs/zhang/src/pathfinder/build/runtime'
make: *** [install-recursive] Error 1
## Comment 3639
Date: 2005-08-11 17:36:53 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Did you update MonetDB as well?
## Comment 3640
Date: 2005-12-19 14:48:26 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
## Comment 3641
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1256794 at http://sourceforge.net/support/tracker.php?aid=1256794
| Failed to compile shredder.mx | https://api.github.com/repos/MonetDB/MonetDB/issues/649/comments | 0 | 2020-11-30T08:47:05Z | 2024-06-27T11:05:28Z | https://github.com/MonetDB/MonetDB/issues/649 | 753,280,563 | 649 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-11 11:08:10 +0200
From: @drstmane
To: @sjoerdmullender
Version: -- development
Last updated: 2005-08-11 11:49:30 +0200
## Comment 3634
Date: 2005-08-11 11:08:10 +0200
From: @drstmane
/bin/sh ../../../libtool --mode=compile gcc -m32
-DHAVE_CONFIG_H -I.
-I/var/tmp/monet.GNU_32-d.12794/MonetDB/src/modules/plain
-I../../..
-I/var/tmp/monet.GNU_32-d.12794/MonetDB/src/modules/plain
-I../../common -I../../gdk -I../../monet
-I/Users/monet/soft/pcre-5.0/include -DLIBLOGGER
-O2 -std=c99 -Wall -W -O6 -fomit-frame-pointer
-finline-functions
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label
-Wno-uninitialized -D_REENTRANT -c -o
lib_logger_la-logger.lo `test -f 'logger.c' || echo
'/var/tmp/monet.GNU_32-d.12794/MonetDB/src/modules/plain/'`logger.c
gcc -m32 -DHAVE_CONFIG_H -I.
-I/var/tmp/monet.GNU_32-d.12794/MonetDB/src/modules/plain
-I../../..
-I/var/tmp/monet.GNU_32-d.12794/MonetDB/src/modules/plain
-I../../common -I../../gdk -I../../monet
-I/Users/monet/soft/pcre-5.0/include -DLIBLOGGER -O2
-std=c99 -Wall -W -O6 -fomit-frame-pointer
-finline-functions
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label
-Wno-uninitialized -D_REENTRANT -c logger.c -fno-common
-DPIC -o .libs/lib_logger_la-logger.o
cc1: warnings being treated as errors
/var/tmp/monet.GNU_32-d.12794/MonetDB/src/modules/plain/logger.mx:
In function 'log_read_format':
/var/tmp/monet.GNU_32-d.12794/MonetDB/src/modules/plain/logger.mx:215:
warning: passing argument 2 of 'stream_readInt' from
incompatible pointer type
make[5]: *** [lib_logger_la-logger.lo] Error 1
make[4]: *** [all] Error 2
See also
http://monetdb.cwi.nl/testing/projects/monetdb/Current/MonetDB/.GNU_32-d/Darwin8.2.0/make.out
.
## Comment 3635
Date: 2005-08-11 11:49:30 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed in CVS by adding a cast.
I'll close this bug report, even though I didn't actually
test anything. :-)
If the bug is not actually fixed, please reopen.
## Comment 3636
Date: 2005-12-19 14:47:54 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
## Comment 3637
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1256535 at http://sourceforge.net/support/tracker.php?aid=1256535
| module "logger" fails to compiler on Darwin | https://api.github.com/repos/MonetDB/MonetDB/issues/648/comments | 0 | 2020-11-30T08:47:02Z | 2024-06-27T11:05:27Z | https://github.com/MonetDB/MonetDB/issues/648 | 753,280,536 | 648 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-09 04:53:47 +0200
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-08-10 01:22:17 +0200
## Comment 3630
Date: 2005-08-09 16:53:47 +0200
From: Wouter Alink <<alink>>
typing help("hot") gives the result shown below. should
hot() return void or return a bat?
---------------
PROC: hot(BAT[any::1,any::2]) : void
no text available
COMMAND: hot(str) : void
MODULE: bat
COMPILED: by adm on Thu Jul 28 10:15:18 2005
Makes a BAT very hot for the BBP. The chance of being
chosen for swapout is
small, afterwards. Returns the BAT itself.
## Comment 3631
Date: 2005-08-10 13:22:17 +0200
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS:
neither hot() nor cold() return anything;
fixed their help messages accordingly.
## Comment 3632
Date: 2005-12-19 14:47:12 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / DOCUMENTATION
## Comment 3633
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1255017 at http://sourceforge.net/support/tracker.php?aid=1255017
| Documentation: hot() | https://api.github.com/repos/MonetDB/MonetDB/issues/647/comments | 0 | 2020-11-30T08:46:59Z | 2024-06-27T11:05:26Z | https://github.com/MonetDB/MonetDB/issues/647 | 753,280,507 | 647 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-08 11:44:19 +0200
From: @grobian
To: @njnes
Version: -- development
Last updated: 2006-01-19 08:11:30 +0100
## Comment 3622
Date: 2005-08-08 11:44:19 +0200
From: @grobian
When running the server as follows:
% Mserver
Monet Database Server V4.8.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module(sql_server,mapi,monettime);
MonetDB>mapi_start();
Then starting a MapiClient:
% MapiClient -lmil
Monet Database Server V4.8.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
mil>quit();
MAPI = anonymous@localhost:50000
ACTION= read_line
QUERY = quit();
ERROR = Connection terminated
The server appears to have crashed:
MonetDB>Segmentation fault
while, when doing it as follows:
% Mserver
Monet Database Server V4.8.3
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module(sql_server);
MonetDB>module(mapi);
MonetDB>module(monettime);
MonetDB>mapi_start();
the server keeps on running.
Perhaps when starting all modules in one module() call
the refcounting of the modules goes wrong.
## Comment 3623
Date: 2005-08-08 13:33:54 +0200
From: @arjenpdevries
Logged In: YES
user_id=572978
I think it is the order in which modules are loaded that is
the problem;
must be related to the code generated by MEL for unloading
modules.
We recently encountered something similar, and the
workaround was to trace the dependency graph of the modules,
and make sure all are loaded before using MapiClient.
So, maybe this does not lead to segmentation fault:
module(streams,ascii_io);
module(mapi);
module(sqlserver);
(sorry for not having submitted the bug report before - mea
culpa)
## Comment 3624
Date: 2005-08-08 18:16:29 +0200
From: @arjenpdevries
Logged In: YES
user_id=572978
I think it is the order in which modules are loaded that is
the problem;
must be related to the code generated by MEL for unloading
modules.
We recently encountered something similar, and the
workaround was to trace the dependency graph of the modules,
and make sure all are loaded before using MapiClient.
So, maybe this does not lead to segmentation fault:
module(streams,ascii_io);
module(mapi);
module(sqlserver);
(sorry for not having submitted the bug report before - mea
culpa)
## Comment 3625
Date: 2005-12-19 14:45:56 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / FAILURE
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/multi-module_crash.SF-1254041.py
This bug is very likely the same as http://sourceforge.net/support/tracker.php?aid=1385152.
## Comment 3626
Date: 2006-01-15 03:06:24 +0100
From: @drstmane
Logged In: YES
user_id=572415
Most probably also related to
1400200 "Mserver: module-order does matter"
https://sourceforge.net/tracker/index.php?func=detail&aid=1400200&group_id=56967&atid=482468
and/or
1385247: "Mserver assertion failed with
module(sql_server,mapi);"
http://sourceforge.net/tracker/index.php?func=detail&aid=1385247&group_id=56967&atid=482468
## Comment 3627
Date: 2006-01-18 19:38:03 +0100
From: @drstmane
Logged In: YES
user_id=572415
Fixed by Niels'
"
too many bugs related to module(x,y,z) so we simply remove
this feature
ie use module(x);module(y);module(z); instead.
".
To be closed tomorrow, once confirmed by the test web; cf.
http://monetdb.cwi.nl/testing/projects/monetdb/Current/sql/.mTests103/GNU.64.64.d-Fedora4/src_test_BugDay_2005-12-19_2.9.3/multi-module_crash.SF-1254041.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/sql/.mTests103/GNU.64.64.d-Fedora4/src_test_BugDay_2005-12-19_2.9.3/multi-module_crash.SF-1254041.err.00.html
## Comment 3628
Date: 2006-01-19 08:11:30 +0100
From: @drstmane
Logged In: YES
user_id=572415
TestWeb confirmed the fix. Closed.
## Comment 3629
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1254041 at http://sourceforge.net/support/tracker.php?aid=1254041
| server crashes when multiple modules are supplied | https://api.github.com/repos/MonetDB/MonetDB/issues/646/comments | 0 | 2020-11-30T08:46:56Z | 2024-06-27T11:05:25Z | https://github.com/MonetDB/MonetDB/issues/646 | 753,280,472 | 646 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-03 04:34:49 +0200
From: @drstmane
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-08-03 07:28:51 +0200
## Comment 3618
Date: 2005-08-03 16:34:49 +0200
From: @drstmane
Pathfinder fails to compile, as runtime/shttpd.c fails
to compile:
========
/bin/sh ../libtool --tag=CC --mode=compile gcc -m64
-DHAVE_CONFIG_H -I.
-I/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime -I..
-I/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/common
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/gdk
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/monet
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/plain
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/contrib
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/mapi
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/C
-I/usr/include/libxml2 -I../compiler/include
-I/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/../compiler/include
-DLIBPF_HTTPD
-O2 -std=c99 -Wall -W -O6 -fomit-frame-pointer
-finline-functions -falign-loops=4
-falign-jumps=4 -falign-functions=4
-fexpensive-optimizations
-funroll-loops -frerun-cse-after-loop -frerun-loop-opt
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label -DEMBEDDED
-DNO_CGI -DNO_AUTH -c -o lib_pf_httpd_la-shttpd.lo
`test -f 'shttpd.c' || echo
'/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/'`shttpd.c
gcc -m64 -DHAVE_CONFIG_H -I.
-I/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime -I..
-I/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/common
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/gdk
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/monet
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/plain
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/contrib
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/mapi
-I/var/tmp/monet.GNU_64-d.21003/MonetDB/.GNU_64-d/Fedora4/include/MonetDB/C
-I/usr/include/libxml2 -I../compiler/include
-I/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/../compiler/include
-DLIBPF_HTTPD -O2 -std=c99 -Wall -W -O6
-fomit-frame-pointer -finline-functions -falign-loops=4
-falign-jumps=4 -falign-functions=4
-fexpensive-optimizations -funroll-loops
-frerun-cse-after-loop -frerun-loop-opt
-Werror-implicit-function-declaration -Werror
-Wno-unused-function -Wno-unused-label -DEMBEDDED
-DNO_CGI -DNO_AUTH -c
/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/shttpd.c
-fPIC -DPIC -o .libs/lib_pf_httpd_la-shttpd.o
/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/shttpd.c:
In function â<80><98>get_dirâ<80><99>:
/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/shttpd.c:1335:
error: â<80><98>S_IFDIRâ<80><99> undeclared (first use
in this function)
/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/shttpd.c:1335:
error: (Each undeclared identifier is reported only once
/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/shttpd.c:1335:
error: for each function it appears in.)
/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/shttpd.c:
In function â<80><98>decideâ<80><99>:
/var/tmp/monet.GNU_64-d.21003/pathfinder/runtime/shttpd.c:2524:
error: â<80><98>S_IFDIRâ<80><99> undeclared (first use
in this function)
make[3]: *** [lib_pf_httpd_la-shttpd.lo] Error 1
========
See also
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.GNU_64-d/Fedora4/make.out
.
## Comment 3619
Date: 2005-08-03 19:28:51 +0200
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS:
To ensure portability, please make sure that "pf_config.h"
is the first include file in each .c file,
either directly or via "pathfinder.h" !
## Comment 3620
Date: 2005-12-19 13:56:22 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
## Comment 3621
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1251087 at http://sourceforge.net/support/tracker.php?aid=1251087
| Pathfinder fails to compile | https://api.github.com/repos/MonetDB/MonetDB/issues/645/comments | 0 | 2020-11-30T08:46:53Z | 2024-06-27T11:05:24Z | https://github.com/MonetDB/MonetDB/issues/645 | 753,280,434 | 645 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-08-02 10:56:17 +0200
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-08-02 07:29:49 +0200
## Comment 3611
Date: 2005-08-02 10:56:17 +0200
From: Wouter Alink <<alink>>
When shredding a large document, the result was:
mil>shred_doc("/export/scratch0/alink/xmark/xmark_550MB.burk.xml","xmark_550MB.burk");
Shredded XML doc("xmark_550MB.burk"), total time
after commit=-2073.-482s
After some digging, i guess that the time() function is
limited to denote intervals of +-4000 seconds.
## Comment 3612
Date: 2005-08-02 10:59:42 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
by the way, could somebody tell me what the time
"-2073.-482" should have been? So i don't have to re-shred
the document to get the timing results?
## Comment 3613
Date: 2005-08-02 13:21:28 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
(unsigned int) ((-2073*1000 - 482) * 1000) us = 2221.485s
And of course the limits are more like +-2148 seconds.
## Comment 3614
Date: 2005-08-02 14:41:01 +0200
From: @drstmane
Logged In: YES
user_id=572415
the problem is, that MIL "usec()" returns only an int, while
C "GDKusec()" provides a lng...
could be fixed easily (2x "int" -> "lng" in
src/modules/plain/alarm.mx); though I'd like to run (all)
tests first, to see, whether anyone relies on usec()
returning int ...
## Comment 3615
Date: 2005-08-02 19:29:49 +0200
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS:
The problem is actually not "time()" (which returns milliseconds
as int, and hence allows to cover 2147483 secs ~= 24 days),
but "usec()", which returned microseconds as int, and hence
allowed
to cover only 2147 secs ~= 35 minutes.
Since the underlying GDKusec provides the milliseconds as lng,
there is no reason why "usec()" should not return a lng as well.
## Comment 3616
Date: 2005-12-19 13:55:53 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST
Testing is fairly simple: sleep for more than 2150 seconds
and then print the value of usec() in the same way as
shred_doc does it, however it would take prohibitively long
(2150 seconds, i.e. more than 35 minutes).
## Comment 3617
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1250067 at http://sourceforge.net/support/tracker.php?aid=1250067
| MIL: time() limited to < (+-4000s) | https://api.github.com/repos/MonetDB/MonetDB/issues/644/comments | 0 | 2020-11-30T08:46:50Z | 2024-06-27T11:05:23Z | https://github.com/MonetDB/MonetDB/issues/644 | 753,280,393 | 644 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-29 02:50:49 +0200
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-08-03 08:36:52 +0200
## Comment 3606
Date: 2005-07-29 14:50:49 +0200
From: Wouter Alink <<alink>>
Currently the 'key(true)' command is ignored when the
head-column is void. However, sometimes this results in
strange behaviour (see example below). According to
Peter the 'key'-settings should be saved (for the times
when void->oid)
MonetDB>var c:=bat(void,int).seqbase(0@0).insert(0@0,20);
MonetDB>c.insert(2@0,20);
MonetDB>c.key(true);
MonetDB>c.insert(2@0,15);
MonetDB>c.print();
-----------------
h t name
oid int type
-----------------
[ 0@0, 20 ]
[ 2@0, 15 ]
MonetDB>var c:=bat(void,int).seqbase(0@0).insert(0@0,20);
MonetDB>c.key(true);
MonetDB>c.insert(2@0,20);
MonetDB>c.insert(2@0,15);
MonetDB>c.print();
-----------------
h t name
oid int type
-----------------
[ 0@0, 20 ]
[ 2@0, 20 ]
[ 2@0, 15 ]
MonetDB>
## Comment 3607
Date: 2005-08-03 08:36:52 +0200
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS.
## Comment 3608
Date: 2005-08-03 19:01:47 +0200
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS.
## Comment 3609
Date: 2005-12-19 13:48:59 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: ALREADY IN TESTWEB
MonetDB/tests/BugsViaSourgeforce/Tests/ID.1247428.milS
## Comment 3610
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1247428 at http://sourceforge.net/support/tracker.php?aid=1247428
| MIL: key(true) behaviour | https://api.github.com/repos/MonetDB/MonetDB/issues/643/comments | 0 | 2020-11-30T08:46:47Z | 2024-06-27T11:05:22Z | https://github.com/MonetDB/MonetDB/issues/643 | 753,280,356 | 643 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-28 10:55:04 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-07-29 11:34:39 +0200
## Comment 3602
Date: 2005-07-28 10:55:04 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>create table test( i int default next value for
non_ex_seq);
6 querytype
sql>insert into test;
4 querytype
[ 1 ]
sql>select * from test;
3 querytype
test table_name
i name
int type
10 length
1 tuplecount
2 id
[ 0 ]
sql>insert into test;
Dunno if it should be allowed to create the table when
non_ex_seq doesn't exist.
(Similar, if it will be dropped afterwards - would require
to check dependencies between database objects.)
But I suppose that the insert succeeds unintentionally.
(The second insert hangs.)
## Comment 3603
Date: 2005-07-29 11:34:39 +0200
From: @njnes
Logged In: YES
user_id=43556
'next value for' now correctly checks for the existance of
the sequence number. This also makes sure no tables are
created without sequence number. Unfortunatly dropping
sequence number cannot check (at the moment) for existing
tables using it. But the insert after the drop should
correctly fail.
## Comment 3604
Date: 2005-12-19 13:43:45 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/default_next_value_for_non_ex_seq.SF.1246631.sql
## Comment 3605
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1246631 at http://sourceforge.net/support/tracker.php?aid=1246631
| SQL: ... default next value for non_ex_seq | https://api.github.com/repos/MonetDB/MonetDB/issues/642/comments | 0 | 2020-11-30T08:46:43Z | 2024-06-27T11:05:21Z | https://github.com/MonetDB/MonetDB/issues/642 | 753,280,323 | 642 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-25 11:15:52 +0200
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-07-26 07:42:33 +0200
## Comment 3595
Date: 2005-07-25 11:15:52 +0200
From: Wouter Alink <<alink>>
Using PF 0.8 + MDB 4.8 (latest CVS) on FC4, 64-bit
When Mserver is started using:
module(pathfinder); mapi_start(); pfstart();
the attached script will result in a Mserver crash
(mostly after 2 times).
## Comment 3596
Date: 2005-07-25 12:28:02 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
the solution to this bug is rather simple...
what i stated above was incorrect, i used the following
initialization:
module(pathfinder); module(mapi); mapi_start(); pfstart();
when leaving out module(mapi); the server doesn't crash anymore.
## Comment 3597
Date: 2005-07-25 12:51:52 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
just a little bit to early... after retrying it after lunch
the bug reappeared even when module(mapi); was left out.
putting module(pathfinder); module(mapi); mapi_start(); and
pfstart(); in different order also didn't make any difference.
## Comment 3598
Date: 2005-07-25 14:24:52 +0200
From: @drstmane
Logged In: YES
user_id=572415
Some quick tests seem to indicate, that this is also related
to Fedora Core 4's gcc 4.0.0-8 on x86_64 and possibly burg,
since the segfault seem to occur in burg code ...
When compiling pathfinder only (not MonetDB; though I
suppose that it should also if also MonetDB is compiled with
gcc32) with Fedora Core 4's gcc 3.2.3-47.fc4 ('gcc32')
everything seems to work fine...!??
## Comment 3599
Date: 2005-07-26 07:42:33 +0200
From: @drstmane
Logged In: YES
user_id=572415
Testing confirmed that this seems to be a compiler/64bit
problem that only occurs with Fedora Core 4's gcc 4.0.0-8 on
x86_64 and Intel's icc 8.1 on IA64; for Fedora Core 4, I
recommend to compile pathfinder with gcc 3.2.3 (configure
with "--with-gcc=gcc32") instead of with the default gcc 4.0.0.
## Comment 3600
Date: 2005-12-19 13:38:17 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: NO TEST / COMPILATION
Even though there is a test script attached, I didn't add it
to the test suite since the problem is purely a compilation
problem and the resolution was "Invalid".
## Comment 3601
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1244326 at http://sourceforge.net/support/tracker.php?aid=1244326
| XQuery: crash after repeatedly querying | https://api.github.com/repos/MonetDB/MonetDB/issues/641/comments | 0 | 2020-11-30T08:46:41Z | 2024-06-28T13:34:44Z | https://github.com/MonetDB/MonetDB/issues/641 | 753,280,297 | 641 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-21 11:38:00 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-08-17 10:00:10 +0200
## Comment 3591
Date: 2005-07-21 11:38:00 +0200
From: Steffen Goeldner <<sgoeldner>>
Running
select * from env where 1 in ( select 1 from env );
shows
MAPI = monetdb@localhost:45123
QUERY = select * from env where 1 in ( select 1 from
env );
ERROR = !ERROR: interpret: no matching MIL operator
to 'reverse(sht)'.
!MAYBE YOU MEAN:
! reverse(BAT[any::1,any::2]) : BAT
[any::2,any::1]
## Comment 3592
Date: 2005-07-21 22:05:22 +0200
From: @njnes
Logged In: YES
user_id=43556
Added the missing implementation
## Comment 3593
Date: 2005-12-19 13:25:25 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/select_from_env.SF-1240701.1242164.sql
## Comment 3594
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1242164 at http://sourceforge.net/support/tracker.php?aid=1242164
| SQL: ... where 1 in ( ... | https://api.github.com/repos/MonetDB/MonetDB/issues/640/comments | 0 | 2020-11-30T08:46:37Z | 2024-06-27T11:05:19Z | https://github.com/MonetDB/MonetDB/issues/640 | 753,280,264 | 640 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-19 10:14:47 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-11-28 11:26:48 +0100
## Comment 3586
Date: 2005-07-19 10:14:47 +0200
From: Steffen Goeldner <<sgoeldner>>
Test case:
sql>select * from env where name = ( select 'prefix' from
env );
3 querytype
env, env table_name
name, value name
varchar, varchar type
6, 17 length
32, tuplecount
14, id
...
I'd expect a 'cardinality violation' if the scalar subquery
returns more than one row.
## Comment 3587
Date: 2005-11-28 23:26:48 +0100
From: @njnes
Logged In: YES
user_id=43556
Added missing dynamic check for cardinality (ie single value
(possibly one per outer reference))
Added test script
src/test/BugTracker/Tests/cardinality_violation.SF-1240701.sql
## Comment 3588
Date: 2005-12-19 13:20:16 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/select_from_env.SF-1240701.sql
## Comment 3589
Date: 2005-12-19 13:26:03 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: moved test to
select_from_env.SF-1240701.1242164.sql
## Comment 3590
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1240701 at http://sourceforge.net/support/tracker.php?aid=1240701
| SQL: SELECT ... where name = ( select ... ) | https://api.github.com/repos/MonetDB/MonetDB/issues/639/comments | 0 | 2020-11-30T08:46:35Z | 2024-06-27T11:05:18Z | https://github.com/MonetDB/MonetDB/issues/639 | 753,280,232 | 639 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-18 02:22:28 +0200
From: @yzchang
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-07-26 07:43:18 +0200
## Comment 3580
Date: 2005-07-18 14:22:28 +0200
From: @yzchang
Hello,
Execution of several example x-queries causes Mserver
to crash (segmentation fault). I recompiled
MonetDB/XQuery from scratch for my 64-bits system, but
it didn't help.
Regards,
Jennie
===================
error info given by gdb:
(gdb) run
Starting program: /ufs/zhang/install/monetdb/bin/Mserver
[Thread debugging using libthread_db enabled]
[New Thread 46912516581280 (LWP 31444)]
[New Thread 1075841376 (LWP 31447)]
Monet Database Server V4.9.0
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 64bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>module(pathfinder);
MonetDB>pfstart();
[New Thread 1077942624 (LWP 31448)]
MonetDB>[New Thread 1080043872 (LWP 31449)]
[New Thread 1082145120 (LWP 31450)]
MonetDB>mapi_start();
[New Thread 1084246368 (LWP 31451)]
MonetDB>[New Thread 1086347616 (LWP 31452)]
[New Thread 1088448864 (LWP 31453)]
MonetDB>
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1080043872 (LWP 31449)]
0x00002aaab5a03509 in reduce (p=0x2aaab5d3b630,
goalnt=57) at ../../../compiler/core/fs.brg:675
675 rule = PFfs_rule (STATE_LABEL (p), goalnt);
(gdb) where
0 0x00002aaab5a03509 in reduce (p=0x2aaab5d3b630,
goalnt=57) at ../../../compiler/core/fs.brg:675
1 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=56) at ../../../compiler/core/fs.brg:729
2 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=55) at ../../../compiler/core/fs.brg:729
3 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=54) at ../../../compiler/core/fs.brg:729
4 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=53) at ../../../compiler/core/fs.brg:729
5 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=52) at ../../../compiler/core/fs.brg:729
6 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=51) at ../../../compiler/core/fs.brg:729
7 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=50) at ../../../compiler/core/fs.brg:729
8 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=49) at ../../../compiler/core/fs.brg:729
9 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=48) at ../../../compiler/core/fs.brg:729
10 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=47) at ../../../compiler/core/fs.brg:729
11 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=46) at ../../../compiler/core/fs.brg:729
12 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=45) at ../../../compiler/core/fs.brg:729
13 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=26) at ../../../compiler/core/fs.brg:729
14 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=25) at ../../../compiler/core/fs.brg:729
15 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b630,
goalnt=87) at ../../../compiler/core/fs.brg:729
16 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b6f0,
goalnt=87) at ../../../compiler/core/fs.brg:729
17 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=80) at ../../../compiler/core/fs.brg:729
18 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=79) at ../../../compiler/core/fs.brg:729
19 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=65) at ../../../compiler/core/fs.brg:729
20 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=63) at ../../../compiler/core/fs.brg:729
21 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=62) at ../../../compiler/core/fs.brg:729
22 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=60) at ../../../compiler/core/fs.brg:729
23 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=58) at ../../../compiler/core/fs.brg:729
24 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=57) at ../../../compiler/core/fs.brg:729
25 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=56) at ../../../compiler/core/fs.brg:729
26 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=55) at ../../../compiler/core/fs.brg:729
27 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=54) at ../../../compiler/core/fs.brg:729
28 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=53) at ../../../compiler/core/fs.brg:729
29 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=52) at ../../../compiler/core/fs.brg:729
30 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=51) at ../../../compiler/core/fs.brg:729
31 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=50) at ../../../compiler/core/fs.brg:729
32 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=49) at ../../../compiler/core/fs.brg:729
33 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=48) at ../../../compiler/core/fs.brg:729
34 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=47) at ../../../compiler/core/fs.brg:729
35 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=46) at ../../../compiler/core/fs.brg:729
36 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=45) at ../../../compiler/core/fs.brg:729
37 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=26) at ../../../compiler/core/fs.brg:729
38 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=25) at ../../../compiler/core/fs.brg:729
39 0x00002aaab5a03654 in reduce (p=0x2aaab5d3b7d0,
goalnt=87) at ../../../compiler/core/fs.brg:729
40 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c170,
goalnt=87) at ../../../compiler/core/fs.brg:729
41 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c1d0,
goalnt=87) at ../../../compiler/core/fs.brg:729
42 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c2b0,
goalnt=80) at ../../../compiler/core/fs.brg:729
43 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c2b0,
goalnt=79) at ../../../compiler/core/fs.brg:729
44 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c2b0,
goalnt=65) at ../../../compiler/core/fs.brg:729
45 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c2b0,
goalnt=63) at ../../../compiler/core/fs.brg:729
46 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c2b0,
goalnt=62) at ../../../compiler/core/fs.brg:729
47 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=61) at ../../../compiler/core/fs.brg:729
48 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=60) at ../../../compiler/core/fs.brg:729
49 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=58) at ../../../compiler/core/fs.brg:729
50 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=57) at ../../../compiler/core/fs.brg:729
51 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=56) at ../../../compiler/core/fs.brg:729
52 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=55) at ../../../compiler/core/fs.brg:729
53 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=54) at ../../../compiler/core/fs.brg:729
54 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=53) at ../../../compiler/core/fs.brg:729
55 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=52) at ../../../compiler/core/fs.brg:729
56 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=51) at ../../../compiler/core/fs.brg:729
57 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=50) at ../../../compiler/core/fs.brg:729
58 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=49) at ../../../compiler/core/fs.brg:729
59 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=48) at ../../../compiler/core/fs.brg:729
60 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=47) at ../../../compiler/core/fs.brg:729
---Type <return> to continue, or q <return> to quit---
61 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=46) at ../../../compiler/core/fs.brg:729
62 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=45) at ../../../compiler/core/fs.brg:729
63 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=26) at ../../../compiler/core/fs.brg:729
64 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=25) at ../../../compiler/core/fs.brg:729
65 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=24) at ../../../compiler/core/fs.brg:729
66 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=73) at ../../../compiler/core/fs.brg:729
67 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=65) at ../../../compiler/core/fs.brg:729
68 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=63) at ../../../compiler/core/fs.brg:729
69 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c6f0,
goalnt=62) at ../../../compiler/core/fs.brg:729
70 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=61) at ../../../compiler/core/fs.brg:729
71 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=60) at ../../../compiler/core/fs.brg:729
72 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=58) at ../../../compiler/core/fs.brg:729
73 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=57) at ../../../compiler/core/fs.brg:729
74 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=56) at ../../../compiler/core/fs.brg:729
75 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=55) at ../../../compiler/core/fs.brg:729
76 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=54) at ../../../compiler/core/fs.brg:729
77 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=53) at ../../../compiler/core/fs.brg:729
78 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=52) at ../../../compiler/core/fs.brg:729
79 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=51) at ../../../compiler/core/fs.brg:729
80 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=50) at ../../../compiler/core/fs.brg:729
81 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=49) at ../../../compiler/core/fs.brg:729
82 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=48) at ../../../compiler/core/fs.brg:729
83 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=47) at ../../../compiler/core/fs.brg:729
84 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=46) at ../../../compiler/core/fs.brg:729
85 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=45) at ../../../compiler/core/fs.brg:729
86 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=26) at ../../../compiler/core/fs.brg:729
87 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=25) at ../../../compiler/core/fs.brg:729
88 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c630,
goalnt=24) at ../../../compiler/core/fs.brg:729
89 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c7b0,
goalnt=25) at ../../../compiler/core/fs.brg:729
90 0x00002aaab5a03654 in reduce (p=0x2aaab5d3c7b0,
goalnt=87) at ../../../compiler/core/fs.brg:729
91 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d150,
goalnt=87) at ../../../compiler/core/fs.brg:729
92 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d1b0,
goalnt=87) at ../../../compiler/core/fs.brg:729
93 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=80) at ../../../compiler/core/fs.brg:729
94 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=79) at ../../../compiler/core/fs.brg:729
95 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=65) at ../../../compiler/core/fs.brg:729
96 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=63) at ../../../compiler/core/fs.brg:729
97 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=62) at ../../../compiler/core/fs.brg:729
98 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=60) at ../../../compiler/core/fs.brg:729
99 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=58) at ../../../compiler/core/fs.brg:729
100 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=57) at ../../../compiler/core/fs.brg:729
101 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=56) at ../../../compiler/core/fs.brg:729
102 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=55) at ../../../compiler/core/fs.brg:729
103 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=54) at ../../../compiler/core/fs.brg:729
104 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=53) at ../../../compiler/core/fs.brg:729
105 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=52) at ../../../compiler/core/fs.brg:729
106 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=51) at ../../../compiler/core/fs.brg:729
107 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=50) at ../../../compiler/core/fs.brg:729
108 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=49) at ../../../compiler/core/fs.brg:729
109 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=48) at ../../../compiler/core/fs.brg:729
110 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=47) at ../../../compiler/core/fs.brg:729
111 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=46) at ../../../compiler/core/fs.brg:729
112 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=45) at ../../../compiler/core/fs.brg:729
113 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=26) at ../../../compiler/core/fs.brg:729
114 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=25) at ../../../compiler/core/fs.brg:729
115 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d290,
goalnt=24) at ../../../compiler/core/fs.brg:729
116 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d350,
goalnt=25) at ../../../compiler/core/fs.brg:729
117 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d350,
goalnt=24) at ../../../compiler/core/fs.brg:729
118 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d350,
goalnt=23) at ../../../compiler/core/fs.brg:729
119 0x00002aaab5a03654 in reduce (p=0x2aaab5d3d3b0,
goalnt=1) at ../../../compiler/core/fs.brg:729
120 0x00002aaab5a10126 in PFfs (r=0x2aaab5d3d3b0) at
../../../compiler/core/fs.brg:3310
121 0x00002aaab599c822 in pf_compile_MonetDB
(xquery=0x1c68f10 "<example> { \n <doc>\n
<test>Hello World</test>\n </doc>/test/text() \n}
</example>\n \n", mode=0x1c68f08 "dm-mapi",
---Type <return> to continue, or q <return> to quit---
prologue=0x40601cb8, query=0x40601cb0,
epilogue=0x40601ca8) at ../../compiler/compile.c:562
122 0x00002aaab599ad18 in CMDxquery (mode=0x1c68f08
"dm-mapi", xquery=0x1c68f10 "<example> { \n <doc>\n
<test>Hello World</test>\n </doc>/test/text() \n}
</example>\n \n")
at ../../runtime/pathfinder.mx:797
123 0x00002aaab599b74a in CMDxquery_server (stk=3,
lt=0x1bf30c8, res=0x40602180) at
../../runtime/pathfinder.mx:979
124 0x00002aaaaaabe4a0 in interpret (stk=3,
lt=0x1bf30c8, res=0x40602180) at
../../../src/monet/monet_interpreter.mx:845
125 0x00002aaaaaadb3f2 in handleRequest
(t=0x2aaaabbe8e30, q=0x1bf60e8, res=0x40602180) at
../../../src/monet/monet_queue.mx:533
126 0x00002aaaaaadb843 in doRequest (t=0x2aaaabbe8e30,
preference=0x0) at ../../../src/monet/monet_queue.mx:559
127 0x00002aaaaab2fcc0 in monetInterpreter
(status=0x403fdfa4) at
../../../src/monet/monet_process.mx:111
128 0x000000351d50697c in start_thread () from
/lib64/libpthread.so.0
129 0x000000351c8c9c2e in clone () from /lib64/libc.so.6
130 0x0000000000000000 in ?? ()
## Comment 3581
Date: 2005-07-21 12:44:44 +0200
From: @drstmane
Logged In: YES
user_id=572415
This problem seems to occur only with Fedora Core 4's gcc
4.0.0-8 on x86_64; hence, I suspect that this is a compiler
problem/bug.
Actually, it seemt to be worth, the less(!) optimization is
usd, i.e., when configured with "--enable-debug", more tests
fail than when configured with "--enable-optimize" ... !??
## Comment 3582
Date: 2005-07-23 12:04:56 +0200
From: @drstmane
Logged In: YES
user_id=572415
The observed segfaults in the pathfinder compiler (mainly
occuring in brg code) indeed seem to be caused by (a bug
in?) Fedora Core 4's gcc 4.0.0-8 (on x86_64) --- when
compiled with gcc 3.2.3 (configured with "--with-gcc=gcc32")
on the same Fedora Core 4 machine, the pathfinder compiler
and hence all MonetDB/XQuery tests work fine.
However, similar (if not identical) segfaults seem to occur
on IA-64 (Itanium) with Intel's icc 8.1-26 ... !??
... to be continued ...
## Comment 3583
Date: 2005-07-26 07:43:18 +0200
From: @drstmane
Logged In: YES
user_id=572415
Testing confirmed that this seems to be a compiler/64bit
problem that only occurs with Fedora Core 4's gcc 4.0.0-8 on
x86_64 and Intel's icc 8.1 on IA64; for Fedora Core 4, I
recommend to compile pathfinder with gcc 3.2.3 (configure
with "--with-gcc=gcc32") instead of with the default gcc 4.0.0.
## Comment 3584
Date: 2005-12-19 13:15:59 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
NBugDay_2005-12-19, sjoerd: O TEST / COMPILATION
## Comment 3585
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1240211 at http://sourceforge.net/support/tracker.php?aid=1240211
| Segmentation fault at Mserver by example XQeury queries | https://api.github.com/repos/MonetDB/MonetDB/issues/638/comments | 0 | 2020-11-30T08:46:31Z | 2024-06-28T13:34:43Z | https://github.com/MonetDB/MonetDB/issues/638 | 753,280,194 | 638 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-18 12:34:01 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-07-19 09:54:57 +0200
## Comment 3576
Date: 2005-07-18 12:34:01 +0200
From: Steffen Goeldner <<sgoeldner>>
The following results in a server crash:
sql>create table x( a int, b varchar(9));
6 querytype
sql>delete from x where 1 = 1;
MAPI = monetdb@localhost:45123
ACTION= read_line
QUERY = delete from x where 1 = 1;
ERROR = Connection terminated
## Comment 3577
Date: 2005-07-18 15:41:03 +0200
From: @njnes
Logged In: YES
user_id=43556
delete now works with predicates which do not use the table.
## Comment 3578
Date: 2005-12-19 13:14:04 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/test/BugDay_2005-12-19_2.9.3/Tests/delete_where_1_equals_1.SF-1240162.sql
## Comment 3579
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1240162 at http://sourceforge.net/support/tracker.php?aid=1240162
| SQL: DELETE ... where 1 = 1 | https://api.github.com/repos/MonetDB/MonetDB/issues/637/comments | 0 | 2020-11-30T08:46:28Z | 2024-06-27T11:05:16Z | https://github.com/MonetDB/MonetDB/issues/637 | 753,280,155 | 637 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-18 10:41:00 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-07-19 09:54:36 +0200
## Comment 3572
Date: 2005-07-18 10:41:00 +0200
From: Steffen Goeldner <<sgoeldner>>
The following results in a server crash:
sql>prepare select * from env where name in (?);
MAPI = monetdb@localhost:45123
ACTION= read_line
QUERY = prepare select * from env where name in (?);
ERROR = Connection terminated
## Comment 3573
Date: 2005-07-18 11:13:57 +0200
From: @njnes
Logged In: YES
user_id=43556
moved set_type_param into check_types. This should fix the
remaining prepare - type check related crashes.
## Comment 3574
Date: 2005-12-19 13:04:32 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/tests/BugDay_2005-12-19_2.9.3/Tests/prepare-where.SF-1238867.1238959.1238965.1240124.sql
## Comment 3575
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1240124 at http://sourceforge.net/support/tracker.php?aid=1240124
| SQL: PREPARE select ... where ... in (?) | https://api.github.com/repos/MonetDB/MonetDB/issues/636/comments | 0 | 2020-11-30T08:46:25Z | 2024-06-27T11:05:15Z | https://github.com/MonetDB/MonetDB/issues/636 | 753,280,124 | 636 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-15 04:54:21 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-07-18 11:16:34 +0200
## Comment 3567
Date: 2005-07-15 16:54:21 +0200
From: Steffen Goeldner <<sgoeldner>>
Sorry again for this esoteric test case! It results in a
server crash:
sql>prepare select ? from env;
MAPI = monetdb@localhost:45123
ACTION= read_line
QUERY = prepare select ? from env;
ERROR = Connection terminated
Dunno whether or not this should be allowed.
## Comment 3568
Date: 2005-07-15 17:10:54 +0200
From: @grobian
Logged In: YES
user_id=963970
no, it shouldn't crash
no, it is not allowed, for the prepare cannot know what to
expect for the ?, ie. type, size, etc.
## Comment 3569
Date: 2005-07-15 18:01:29 +0200
From: @njnes
Logged In: YES
user_id=43556
Result columns need to have a type, this is enforced now.
## Comment 3570
Date: 2005-12-19 13:04:30 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/tests/BugDay_2005-12-19_2.9.3/Tests/prepare-where.SF-1238867.1238959.1238965.1240124.sql
## Comment 3571
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1238965 at http://sourceforge.net/support/tracker.php?aid=1238965
| SQL: PREPARE select ? ... | https://api.github.com/repos/MonetDB/MonetDB/issues/635/comments | 0 | 2020-11-30T08:46:22Z | 2024-06-27T11:05:14Z | https://github.com/MonetDB/MonetDB/issues/635 | 753,280,077 | 635 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-15 04:41:26 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-07-18 11:15:20 +0200
## Comment 3563
Date: 2005-07-15 16:41:26 +0200
From: Steffen Goeldner <<sgoeldner>>
Maybe it's related to http://sourceforge.net/support/tracker.php?aid=1238867, but even more
esoteric:
sql>prepare select * from env where ? = ?;
MAPI = monetdb@localhost:45123
QUERY = prepare select * from env where ? = ?;
ERROR = !subquery result missing
I guess there is an additional difficulty in type
inference.
## Comment 3564
Date: 2005-07-15 18:00:36 +0200
From: @njnes
Logged In: YES
user_id=43556
We now enforce atleast one known type on expressions.
## Comment 3565
Date: 2005-12-19 13:04:26 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/tests/BugDay_2005-12-19_2.9.3/Tests/prepare-where.SF-1238867.1238959.1238965.1240124.sql
## Comment 3566
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1238959 at http://sourceforge.net/support/tracker.php?aid=1238959
| SQL: PREPARE ... where ? = ? | https://api.github.com/repos/MonetDB/MonetDB/issues/634/comments | 0 | 2020-11-30T08:46:19Z | 2024-06-27T11:05:13Z | https://github.com/MonetDB/MonetDB/issues/634 | 753,280,046 | 634 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-15 04:08:19 +0200
From: @swingbit
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-07-21 07:33:14 +0200
## Comment 3556
Date: 2005-07-15 16:08:19 +0200
From: @swingbit
casting bats makes them writable.
var b:=new(oid,int).insert(0@0,10).insert(1@0,
10).access(BAT_READ);
b.access().print();
[1]
b.[dbl]().access().print();
[0]
## Comment 3557
Date: 2005-07-15 16:21:19 +0200
From: @swingbit
Logged In: YES
user_id=883451
Apparently the same happens more in general for multiplexed
operations (not for all of them though).
Is this a bug or a feature?
## Comment 3558
Date: 2005-07-21 12:47:52 +0200
From: @drstmane
Logged In: YES
user_id=572415
I'd say it's a BUG, since neither bat casts nor (general)
multiplexes should return a writeable BAT (not even it the
input BAT was writeable!).
Do you by any chance have a more specific analysis, which
casts/multiplexes do return a writeable BAT and which don't?
## Comment 3559
Date: 2005-07-21 17:57:11 +0200
From: @swingbit
Logged In: YES
user_id=883451
mhh... I cannot reproduce it on normal multiplexes... maybe
it applies only to casts.
But I confirm that it happens to all the casts.
Notice that it does not happens, for istance, when casting
between oid and 64bits types (for 64bits machines).
## Comment 3560
Date: 2005-07-21 19:33:14 +0200
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS (src/monet/monet_interpreter.mx):
when interpret_batcast() returns a new BAT (instead of a view),
make sure, the new BAT is marked BAT_READ;
(in case of TOK_BATCAST, we do (can?)not use CATCHVAL to
take care of this...)
## Comment 3561
Date: 2005-12-19 13:08:46 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
MonetDB/tests/BugDay_2005-12-19_4.9.3/Tests/bat_cast_makes_writable.SF-1238930.milS
## Comment 3562
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1238930 at http://sourceforge.net/support/tracker.php?aid=1238930
| bat cast makes bats writable | https://api.github.com/repos/MonetDB/MonetDB/issues/633/comments | 0 | 2020-11-30T08:46:16Z | 2024-06-27T11:05:12Z | https://github.com/MonetDB/MonetDB/issues/633 | 753,280,026 | 633 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-15 02:26:42 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-07-18 11:15:01 +0200
## Comment 3552
Date: 2005-07-15 14:26:42 +0200
From: Steffen Goeldner <<sgoeldner>>
Sorry for this esoteric test case! It results in a server
crash:
sql>prepare select * from env where 1 = ?;
MAPI = monetdb@localhost:45123
ACTION= read_line
QUERY = prepare select * from env where 1 = ?;
ERROR = Connection terminated
## Comment 3553
Date: 2005-07-15 15:12:03 +0200
From: @njnes
Logged In: YES
user_id=43556
The type checks for the simple predicate case were to late
(the type_check code automatically updates the types of
parameters (?)).
## Comment 3554
Date: 2005-12-19 13:04:17 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
sql/src/tests/BugDay_2005-12-19_2.9.3/Tests/prepare-where.SF-1238867.1238959.1238965.1240124.sql
## Comment 3555
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1238867 at http://sourceforge.net/support/tracker.php?aid=1238867
| SQL: PREPARE ... where 1 = ? | https://api.github.com/repos/MonetDB/MonetDB/issues/632/comments | 0 | 2020-11-30T08:46:14Z | 2024-06-27T11:05:11Z | https://github.com/MonetDB/MonetDB/issues/632 | 753,280,007 | 632 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-14 07:43:51 +0200
From: Wouter Alink <<alink>>
To: @njnes
Version: -- development
Last updated: 2005-07-21 05:03:19 +0200
## Comment 3546
Date: 2005-07-14 19:43:51 +0200
From: Wouter Alink <<alink>>
When shredding a document using MapiClient in a
Cygwin environment (compiled with MinGW):
MapiClient -lmil -q -s "shred_doc
(\"c:\\\\cygwin\\\\home\\\\Wouter\\\\test\\\\temp.xml\",\"tem
p.xml\");"
MapiClient shreds the document successfully, BUT,
the file (/home/Wouter/test/temp.xml) will be
kept 'locked' by Mserver. I found this out, because I
wanted to delete the file, but I first had to kill Mserver,
before the lock on the file was released.
(i'm using MonetDB 4.9.0 (latest CVS) + Pathfinder
0.9.0 (latest CVS))
## Comment 3547
Date: 2005-07-15 14:13:55 +0200
From: @njnes
Logged In: YES
user_id=43556
added code to cleanup the xml2 parser context (I hope this
will close the
file). Could you test this?
## Comment 3548
Date: 2005-07-21 12:50:46 +0200
From: @drstmane
Logged In: YES
user_id=572415
Wouter,
do Niels' changes solve the problem?
## Comment 3549
Date: 2005-07-21 17:03:19 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
bug is fixed, thanx!
## Comment 3550
Date: 2005-12-19 12:52:44 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
BugDay_2005-12-19, sjoerd: TEST ADDED / SUCCESS
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/file_locked_after_shredding.SF-1238352.*
## Comment 3551
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1238352 at http://sourceforge.net/support/tracker.php?aid=1238352
| XQuery: shredder keeps lock on files | https://api.github.com/repos/MonetDB/MonetDB/issues/631/comments | 0 | 2020-11-30T08:46:10Z | 2024-06-27T11:05:10Z | https://github.com/MonetDB/MonetDB/issues/631 | 753,279,964 | 631 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-14 04:13:16 +0200
From: Wouter Alink <<alink>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-07-14 05:45:07 +0200
## Comment 3542
Date: 2005-07-14 16:13:16 +0200
From: Wouter Alink <<alink>>
Excuse me for being new to compiling in a cygwin
environment, but I'm not able to compile Pathfinder
(from CVS) using the HowToStart-MinGW (see error
message below, which occurs during 'make'). I tried to
find what i did wrong, and came across:
"The values for the --with-mel and --with-mx options
should be the full (Cygwin) path to the respective
binaries, which should be compiled for
Cygwin (i.e. not using the MinGW tools)."
And maybe this is the reason, and if so, how do I
compile Mx and mel seperately?
-------------------
...
...
...
Making all in runtime
make[2]: Entering directory
`/home/Wouter/src/pathfinder-0.8/MinGW/runtime'
/home/Wouter/MinGW/bin/Mx.exe -x
m ../../runtime/pf_support.mx
../../runtime/pf_support.mx: .\pf_support.m - not modified
/home/Wouter/MinGW/bin/mel.exe -I../../runtime -
I/home/Wouter/MinGW/include/MonetDB -
I/home/Wouter/MinGW/include/MonetDB
/common -
I/home/Wouter/MinGW/include/MonetDB/gdk -
I/home/Wouter/MinGW/include/MonetDB/monet -
I/home/Wouter/MinGW/include
/MonetDB/plain -
I/home/Wouter/MinGW/include/MonetDB/contrib -
I/home/Wouter/MinGW/include/MonetDB/mapi -
I/home/Wouter/Min
GW/include/MonetDB/C -
I/home/Wouter/MinGW/include/libxml2 -
I../compiler/include -I../../runtime/../compiler/include -pr
oto pf_support.m > pf_support.proto.h
C:\cygwin\home\Wouter\MinGW\bin\mel.exe: "../../runtim
e/pf_support.mx", line 29, Module not found at "9" of
monettime.
C:\cygwin\home\Wouter\MinGW\bin\mel.exe: "../../runtim
e/pf_support.mx", line 29, syntax error at "1" of ;.
Found 2 syntax errors.
make[2]: *** [pf_support.proto.h] Fout 1
make[2]: *** Bestand `pf_support.proto.h' wordt
verwijderd
make[2]: Leaving directory
`/home/Wouter/src/pathfinder-0.8/MinGW/runtime'
make[1]: *** [all-recursive] Fout 1
make[1]: Leaving directory
`/home/Wouter/src/pathfinder-0.8/MinGW'
make: *** [all] Fout 2
## Comment 3543
Date: 2005-07-14 17:45:06 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Where is the bug? If you don't follow the instructions that
are given, the error is not with the instructions.
As to your question, you either:
- compile MonetDB using Cygwin (in addition to compiling it
using MinGW);
- check out buildtools from MonetDB SourceForge and follow
the instructions.
## Comment 3544
Date: 2005-12-19 17:10:10 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: NO TEST / COMPILATION
## Comment 3545
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1238228 at http://sourceforge.net/support/tracker.php?aid=1238228
| compiling PF using MinGW | https://api.github.com/repos/MonetDB/MonetDB/issues/630/comments | 0 | 2020-11-30T08:46:08Z | 2024-06-28T13:34:43Z | https://github.com/MonetDB/MonetDB/issues/630 | 753,279,934 | 630 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-13 01:22:50 +0200
From: Wouter Alink <<alink>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-07-14 03:55:23 +0200
## Comment 3537
Date: 2005-07-13 13:22:50 +0200
From: Wouter Alink <<alink>>
When issuing the following query:
doc("video.xml")//*/self::scene
in a document which looks like:
<video>
<scene id="1"/>
<scene id="2"/>
<scene id="3"/>
</video>
the result is:
WARNING: BBPdecref: range error 450
-------------
NOTE: i still use a version (0.8) of pathfinder from a few
weeks ago, so if it is already fixed, please ignore this
bugreport
## Comment 3538
Date: 2005-07-13 13:53:57 +0200
From: @drstmane
Logged In: YES
user_id=572415
[Sorry, couldn't resist...]
Cetero Censeo ...:
You can easily have this monitored automatically every night
with the latest version by adding the respective test to the
CVS repository ... then, you'll also know, when which bugs
are fixed, and hence, when you might want to upgrade your
version ... ;-))
(ps: "you" does not only refert to you, Wouter, but to
everyone who reports bugs ... ;-))
## Comment 3539
Date: 2005-07-14 15:55:23 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed in CVS (XQuery_0-8 release branch):
There is no reason for mvaljoin to require that the BATs in
ws_item have dense heads.
Thanks to Wouter for submitting the bug report and providing
the test.
Thanks to JanR for pointing me to the origin of this bug.
## Comment 3540
Date: 2005-12-19 18:53:46 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: TEST ADDED / SUCCESS
src/pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/self_axis_not_supported.SF-1237444.milM
## Comment 3541
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1237444 at http://sourceforge.net/support/tracker.php?aid=1237444
| XQuery: self-axis not supported? | https://api.github.com/repos/MonetDB/MonetDB/issues/629/comments | 0 | 2020-11-30T08:46:05Z | 2024-06-27T11:05:08Z | https://github.com/MonetDB/MonetDB/issues/629 | 753,279,902 | 629 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-12 03:42:55 +0200
From: thijs <<thijswes>>
To: @njnes
Version: -- development
Last updated: 2007-12-21 04:20:03 +0100
## Comment 3532
Date: 2005-07-12 15:42:55 +0200
From: thijs <<thijswes>>
The PF documentation, ie. HowToStart-PF, states that
the document cache keeps track of the "last-modified"
time. This is supposed to be checked on each request.
However, when a file is modified, pathfinder does not
update it's cache.
Suppose we have a document test.xml that looks like this:
<?xml version="1.0"?>
<A/>
Then
echo 'doc("file://test.xml")'|MapiClient -lx
produces
<A/>
(and so it should).
But if we then modify test.xml to read:
<?xml version="1.0"?>
<B/>
and run the same command again
echo 'doc("file://test.xml")'|MapiClient -lx
the output is still:
<A/>
## Comment 3533
Date: 2005-12-19 18:03:31 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: RELEASE
Don't know how to write test script, sorry!
## Comment 3534
Date: 2006-12-20 04:20:04 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 3535
Date: 2007-12-21 04:20:03 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 3536
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1236711 at http://sourceforge.net/support/tracker.php?aid=1236711
| \"last-modified\" time ignored in document cache | https://api.github.com/repos/MonetDB/MonetDB/issues/628/comments | 0 | 2020-11-30T08:46:02Z | 2024-06-27T11:05:07Z | https://github.com/MonetDB/MonetDB/issues/628 | 753,279,865 | 628 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-07 04:46:33 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-07-18 11:14:26 +0200
## Comment 3526
Date: 2005-07-07 16:46:33 +0200
From: Steffen Goeldner <<sgoeldner>>
PREPARE doesn't work for queries with like predicates:
sql>prepare select * from env where name like ?;
MAPI = monetdb@localhost:45123
QUERY = prepare select * from env where name like ?;
ERROR = !parse error in: "prepare select * from env
where name like ?"
sql>prepare select * from env where name like 'm%';
10 querytype
prepare, prepare, prepare table_name
type, digits, scale name
varchar, int, int type
0, 0, 0 length
0 tuplecount
0 id
sql>execute 0 ('m%');
3 querytype
env, env table_name
name, value name
varchar, varchar type
14, 55 length
11, tuplecount
5, id
[ "monet_pid", "456" ]
...
sql>execute 0 ();
MAPI = monetdb@localhost:45123
QUERY = execute 0 ();
ERROR = !wrong number of arguments for prepared
statement
## Comment 3527
Date: 2005-07-15 10:06:18 +0200
From: @njnes
Logged In: YES
user_id=43556
support for ? in like statements is now added.
## Comment 3528
Date: 2005-07-15 14:20:27 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
I see Resolution is 'Fixed'. Please don't forget the second
test case:
prepare select * from env where name like 'm%';
## Comment 3529
Date: 2005-07-15 15:18:01 +0200
From: @njnes
Logged In: YES
user_id=43556
The wrong number of arguments bug is related to the 'm%'
being treated as an literal which is put in an argument
list. We do this to improve query reuse (query caching).
Queries where only the literals differ are handled by the
same mil proc. But for prepared statement we shouldn't do
this literal replacement. This is now fixed in the parser.
## Comment 3530
Date: 2005-12-19 15:10:36 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: TEST ADDED / SUCCESS
src/sql/src/test/BugDay_2005-12-19_2.9.3/Tests/prepare_doesnot_like_LIKE.SF-1234205.sql
Note: the test script is added to show "prepare select *
from env where name like ?" works (correctly). The ID of
the prepared statement is hard coded in the test script, so
the test can fail if the prepared statement gets another ID.
A test script in application level should be added to
dynamically retrieve the ID of the prepared statement and
call 'execute' with the right ID. But this is out of the
scoope of the BugDay.
## Comment 3531
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1234205 at http://sourceforge.net/support/tracker.php?aid=1234205
| SQL: PREPARE doesn't like LIKE | https://api.github.com/repos/MonetDB/MonetDB/issues/627/comments | 0 | 2020-11-30T08:45:58Z | 2024-06-27T11:05:06Z | https://github.com/MonetDB/MonetDB/issues/627 | 753,279,835 | 627 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-07-04 03:22:03 +0200
From: rimluga <<rimluga>>
To: @sjoerdmullender
Version: -- development
Last updated: 2005-07-04 04:39:12 +0200
## Comment 3521
Date: 2005-07-04 15:22:03 +0200
From: rimluga <<rimluga>>
Hi,
am sorry to report this request for information here, I
tried to email you, and I didn't succeed.
I would like to know if Monet allows automatic key
generation,
Best,
## Comment 3522
Date: 2005-07-04 16:39:12 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
MonetDB does not allow automatic key generation.
I don't know how you tried to email us. You should be able
to reach us through any of the following addresses:
monet@cwi.nl
monetdb-users@lists.sourceforge.net (MonetDB users list)
monetdb-developers@lists.sourceforge.net (MonetDB developers)
## Comment 3523
Date: 2005-12-19 13:18:58 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: NO TEST / DOCUMENTATION
## Comment 3524
Date: 2005-12-19 13:23:30 +0100
From: @grobian
Logged In: YES
user_id=963970
FYI:
The CVS version of MonetDB/SQL supports sequences, such that
auto_increment, serial and identities can be used.
## Comment 3525
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1232208 at http://sourceforge.net/support/tracker.php?aid=1232208
| key autogeneration | https://api.github.com/repos/MonetDB/MonetDB/issues/626/comments | 0 | 2020-11-30T08:45:56Z | 2024-06-28T13:34:42Z | https://github.com/MonetDB/MonetDB/issues/626 | 753,279,806 | 626 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-30 02:17:37 +0200
From: @swingbit
To: GDK devs <<bugs-common>>
Version: -- development
Last updated: 2005-07-15 11:21:59 +0200
## Comment 3518
Date: 2005-06-30 14:17:37 +0200
From: @swingbit
bat casting from lng to oid fails with assertion on a
32 bits machine:
var a:=new(void,lng).seqbase(0@0);
a.insert(nil,0LL).insert(nil,1LL).insert(nil,2LL);
var b:=[oid](a);
Mserver:
/net/db2/export/data1/roberto/MonetDB/src/MonetDB/src/monet/monet_interpreter.mx:1493:
interpret_batcast: Assertion `b->dims.tailtype != 6 &&
tpe != 6' failed.
Aborted
Notice that the following does NOT fail. The only
difference is that the bat 'a' is not sorted on its tail:
var a:=new(void,lng).seqbase(0@0);
a.insert(nil,0LL).insert(nil,2LL).insert(nil,1LL);
var b:=[oid](a);
## Comment 3519
Date: 2005-12-19 14:34:59 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: TEST ADDED / SUCCESS
src/MonetDB/tests/BugDay_2005-12-19_4.9.3/Tests/assertion_on_bat_cast.SF-1230320.milS
## Comment 3520
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1230320 at http://sourceforge.net/support/tracker.php?aid=1230320
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-common@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| Assertion on bat cast | https://api.github.com/repos/MonetDB/MonetDB/issues/625/comments | 0 | 2020-11-30T08:45:53Z | 2024-06-27T11:05:04Z | https://github.com/MonetDB/MonetDB/issues/625 | 753,279,769 | 625 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-29 03:07:29 +0200
From: @swingbit
To: @sjoerdmullender
Version: -- development
Last updated: 2005-07-04 05:02:30 +0200
## Comment 3514
Date: 2005-06-29 15:07:29 +0200
From: @swingbit
.COMMAND count ( BAT[void,void] b, lng cnt ) : lng =
CMDsetcount_bat;
"Returns the current size (in number of elements) of a
BAT."
I guess this should be:
"Sets the size (in number of elements) of a BAT."
Also, this text -is- in bat.mx:
The sole reason not to move it to
the bat.mx module is that overall the kernel does not
support bat[void,void]
## Comment 3515
Date: 2005-07-04 17:02:30 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Fixed.
## Comment 3516
Date: 2005-12-19 12:43:01 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: NO TEST / DOCUMENTATION
## Comment 3517
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1229685 at http://sourceforge.net/support/tracker.php?aid=1229685
| Too quick copy and paste in bat.mx | https://api.github.com/repos/MonetDB/MonetDB/issues/624/comments | 0 | 2020-11-30T08:45:50Z | 2024-06-27T11:05:03Z | https://github.com/MonetDB/MonetDB/issues/624 | 753,279,743 | 624 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-29 10:46:03 +0200
From: @swingbit
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-06-29 01:50:33 +0200
## Comment 3510
Date: 2005-06-29 10:46:03 +0200
From: @swingbit
In the MIL documentation:
http://monetdb.cwi.nl/TechDocs/FrontEnds/mil/index.html
section "2.1.5 Unique OIDs", the following link is dead:
http://monetdb.cwi.nl/TechDocs/Core/Modules/builtin/index.htmlmod_2_2_0
as the 'builtin' module has been replaced by 'kernel',
I suppose.
Also, the 'kernel' module documentation on the website
gives the following:
[an error occurred while processing this directive]
## Comment 3511
Date: 2005-06-29 13:50:33 +0200
From: @drstmane
Logged In: YES
user_id=572415
Fixed monetweb: (quite some time ago) module "kernel" had
been renamed to "builtin"...
## Comment 3512
Date: 2005-12-19 12:37:11 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennnie : NO TEST / DOCUMENTATION
## Comment 3513
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1229552 at http://sourceforge.net/support/tracker.php?aid=1229552
| dead link in MIL docs | https://api.github.com/repos/MonetDB/MonetDB/issues/623/comments | 0 | 2020-11-30T08:45:47Z | 2024-06-27T11:05:02Z | https://github.com/MonetDB/MonetDB/issues/623 | 753,279,720 | 623 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-28 09:03:51 +0200
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2005-07-24 06:25:40 +0200
## Comment 3500
Date: 2005-06-28 21:03:51 +0200
From: @grobian
When compiling monetdb with gcc 4.0 and readline 5.0 on
OSX Tiger, the compilation dies:
cc1: warnings being treated as errors
../../../../../src/mapi/clients/C/Mapi.mx: In function
'read_line':
../../../../../src/mapi/clients/C/Mapi.mx:2683:
warning: format '%d' expects type 'int', but argument 2
has type 'ssize_t'
../../../../../src/mapi/clients/C/Mapi.mx: In function
'mapi_execute_internal':
../../../../../src/mapi/clients/C/Mapi.mx:3125:
warning: format '%u' expects type 'unsigned int', but
argument 2 has type 'size_t'
../../../../../src/mapi/clients/C/Mapi.mx: In function
'mapi_query_part':
../../../../../src/mapi/clients/C/Mapi.mx:3296:
warning: format '%u' expects type 'unsigned int', but
argument 2 has type 'size_t'
make[7]: *** [libMapi_la-Mapi.lo] Error 1
make[6]: *** [all] Error 2
make[5]: *** [all-recursive] Error 1
make[4]: *** [all-recursive] Error 1
make[3]: *** [all] Error 2
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
[aphrodite:current/MonetDB/build] %
The build is using 32 bits
## Comment 3501
Date: 2005-07-15 10:40:03 +0200
From: @njnes
Logged In: YES
user_id=43556
is this bug still open? Could you try again as the mapi code
is changed since
the bug was reported, ie line numbers aren't valid anymore.
## Comment 3502
Date: 2005-07-23 10:31:46 +0200
From: @drstmane
Logged In: YES
user_id=572415
Just for info:
compilation works fine with gcc 4.0.0 and readline 5.0 on
Fedora Core 4.
## Comment 3503
Date: 2005-07-23 20:37:07 +0200
From: @grobian
Logged In: YES
user_id=963970
Still the same errors:
cc1: warnings being treated as errors
../../../../../src/mapi/clients/C/Mapi.mx: In function
'read_line':
../../../../../src/mapi/clients/C/Mapi.mx:2683: warning:
format '%d' expects type 'int', but argument 2 has type
'ssize_t'
../../../../../src/mapi/clients/C/Mapi.mx: In function
'mapi_execute_internal':
../../../../../src/mapi/clients/C/Mapi.mx:3125: warning:
format '%u' expects type 'unsigned int', but argument 2 has
type 'size_t'
../../../../../src/mapi/clients/C/Mapi.mx: In function
'mapi_query_part':
../../../../../src/mapi/clients/C/Mapi.mx:3296: warning:
format '%u' expects type 'unsigned int', but argument 2 has
type 'size_t'
make[7]: *** [libMapi_la-Mapi.lo] Error 1
See attached path for a fix for this issue. "%zu" and "%zd"
should be used on Darwin according to the printf(3) man-page.
With the attached patch readline works as expected
## Comment 3504
Date: 2005-07-24 01:17:35 +0200
From: @drstmane
Logged In: YES
user_id=572415
Your patch indicates, that your using revision 1.160 (Jun 5
2005) of src/mapi/clients/C/Mapi.mx;
however, the problematic definitions of SSZFMT have been
removed from src/mapi/clients/C/Mapi.mx as of revision
1.156.2.5 (Jun 7 2005, MonetDB_4-8 release branch)
respectively revision 1.161 (Jun 12 2005, development trunk,
aka. CVS HEAD).
Hence, please update your source code, and try again.
The up-to-date code contains a fixed definition of SSZFMT
(now centralized in src/common/monet_utils.mx) that is
compliant with your patch and should work fine in your case...
## Comment 3505
Date: 2005-07-24 10:16:50 +0200
From: @grobian
Logged In: YES
user_id=963970
thanks CVS for ignoring my -A and -C
The latest news is to get over 10000 erors on
gdk_scanselect, but maybe cvs has not updated the sources
there as well...
to be continued.
## Comment 3506
Date: 2005-07-24 13:28:11 +0200
From: @drstmane
Logged In: YES
user_id=572415
In case youre not sure, whether your current working
directory has the latest source even after a 'cvs up -dPA',
you might want to consider getting a fresh check-out ('cvs
-d... co -P MonetDB') in a new location, and compare both
tree, e.g., via 'diff -r [-xCVS]' ...
Further, unless you already did this, you might want to redo
all compilation from scratch, i.e.,
cd <sourcedir>
./de-bootstrap
cvs up -dPA
./bootstrap
rm -rf <builddir>
mkdir <builddir>
cd <builddir>
<sourcedir>/configure ...
make
...
## Comment 3507
Date: 2005-07-24 14:57:47 +0200
From: @grobian
Logged In: YES
user_id=963970
I've done this 'clean' compile and let it simmer when I was
bicycling.
No errors whatsoever found, so this bug report can be closed
as invalid.
Sorry for the hassle!
## Comment 3508
Date: 2005-12-19 11:04:06 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDay_2005-12-19, Jennie: NO TEST / COMPILATION
## Comment 3509
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1229245 at http://sourceforge.net/support/tracker.php?aid=1229245
| Compilation error Darwin readline5 gcc4 | https://api.github.com/repos/MonetDB/MonetDB/issues/622/comments | 0 | 2020-11-30T08:45:44Z | 2024-06-27T11:05:01Z | https://github.com/MonetDB/MonetDB/issues/622 | 753,279,694 | 622 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-28 04:02:58 +0200
From: Steffen Goeldner <<sgoeldner>>
To: Steffen Goeldner <<sgoeldner>>
Version: -- development
Last updated: 2007-12-21 04:20:04 +0100
## Comment 3486
Date: 2005-06-28 16:02:58 +0200
From: Steffen Goeldner <<sgoeldner>>
Running
use DBI();
$dsn = 'dbi:ODBC:driver=MonetDB ODBC Driver';
$dbh = DBI->connect( $dsn ) or die $DBI::errstr;
$tbl = 'perl_dbd_test';
$dbh->do("drop table $tbl");
$dbh->do("create table $tbl( A int, B varchar(50), C
varchar(300), D date )");
$sth = $dbh->prepare("insert into $tbl (A, B, C, D)
VALUES (?, ?, ?, ?)");
$sth->bind_param( 1, 1 );
$sth->bind_param( 2,'foo');
$sth->bind_param( 3,'me' x 120 );
$sth->bind_param( 4,'1998-05-13');
$sth->execute;
shows
DBD::ODBC::st execute failed: [MonetDB][ODBC
Driver 1.0]!unexpected end of input (SQL-42000)
## Comment 3487
Date: 2005-06-28 21:05:02 +0200
From: @grobian
Logged In: YES
user_id=963970
does the server survive this?
## Comment 3488
Date: 2005-06-29 08:40:58 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
Yes, it survives. Is it possible to trace the input it receives?
## Comment 3489
Date: 2005-06-29 09:21:45 +0200
From: @grobian
Logged In: YES
user_id=963970
if you don't have strace available for your platform, I
think tools like ethereal are your only solution if you
don't want to hack sources.
## Comment 3490
Date: 2005-06-30 16:22:33 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
I added traces on the client-side (src/mapi/clients/C/Mapi.c):
@@ -1800,7 +1800,9 @@
p = q + 1;
q = strchr(p, PLACEHOLDER);
}
+ printf("***: =%d %d ... %s\n", lim, strlen(hdl->query),
hdl->query + k - 10);
checkSpace(strlen(p));
+ printf("***: =%d %d ... %s\n", lim, strlen(hdl->query),
hdl->query + k - 10);
strcpy(hdl->query + k, p);
if (hdl->mid->trace)
printf("param_store: result=%s\n", hdl->query);
The output is:
***: =314 315 ... 998-05-13'
***: =572 314 ... 998-05-13
I.e., lim < strlen(hdl->query), which means a buffer overrun.
Trying to make room for \' and \0 seems to work:
@@ -1779,13 +1779,13 @@
buf[0] = *(char *) src;
buf[1] = 0;
val = mapi_quote(buf, 1);
- checkSpace(strlen(val));
+ checkSpace(strlen(val)+3);
sprintf(hdl->query + k, "'%s'", val);
free(val);
break;
case MAPI_VARCHAR:
val = mapi_quote((char *) src, hdl-
>params[i].sizeptr ? *hdl->params[i].sizeptr : -1);
- checkSpace(strlen(val));
+ checkSpace(strlen(val)+3);
sprintf(hdl->query + k, "'%s'", val);
free(val);
break;
@@ -1800,7 +1800,7 @@
p = q + 1;
q = strchr(p, PLACEHOLDER);
}
- checkSpace(strlen(p));
+ checkSpace(strlen(p)+1);
strcpy(hdl->query + k, p);
if (hdl->mid->trace)
printf("param_store: result=%s\n", hdl->query);
O.k.?
## Comment 3491
Date: 2005-06-30 17:22:41 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Looks good. Can you check this in in the stable branch?
## Comment 3492
Date: 2005-07-01 14:36:48 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
O.k., fixed:
<http://cvs.sourceforge.net/viewcvs.py/monetdb/MonetDB/src/
mapi/clients/C/Mapi.mx?r1=1.156.2.7&r2=1.156.2.8>
## Comment 3493
Date: 2005-12-19 16:51:15 +0100
From: @yzchang
Logged In: YES
user_id=341633
BugDag_2005-12-19, Jennie: REASSIGN to Steffen or Sjoerd.
Don't know how to make test script for Perl, sorry!
## Comment 3494
Date: 2005-12-19 18:12:22 +0100
From: @drstmane
Logged In: YES
user_id=572415
see also
https://sourceforge.net/tracker/index.php?func=detail&aid=1225455&group_id=56967&atid=482468
## Comment 3495
Date: 2006-01-15 15:14:21 +0100
From: @grobian
Logged In: YES
user_id=963970
is this bug pending because there is no test available?
## Comment 3496
Date: 2006-01-15 15:39:08 +0100
From: @drstmane
Logged In: YES
user_id=572415
right, all "pending" bugs still don't have a test script,
although one could create one --- usualy requires "more
effort" and/or new functionallity in Mtest.py and/or
detailed checks whether to add a test script at all...
## Comment 3497
Date: 2006-12-20 04:20:04 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 3498
Date: 2007-12-21 04:20:04 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 3499
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1229018 at http://sourceforge.net/support/tracker.php?aid=1229018
| !unexpected end of input | https://api.github.com/repos/MonetDB/MonetDB/issues/621/comments | 0 | 2020-11-30T08:45:41Z | 2024-06-27T11:05:00Z | https://github.com/MonetDB/MonetDB/issues/621 | 753,279,652 | 621 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-25 10:13:01 +0200
From: @grobian
To: @sjoerdmullender
Version: -- development
Last updated: 2005-06-25 10:49:28 +0200
## Comment 3480
Date: 2005-06-25 22:13:01 +0200
From: @grobian
On behalf of:
I don"t know whether MonetDB.conf in CWI\MonetDB\etc is
used, but it contains the following lines:
Where MonetDB is installed
prefix=C:\Documents and Settings\sjoerd\My
Documents\src\MonetDB\NT
exec_prefix=C:\Documents and Settings\sjoerd\My
Documents\src\MonetDB\NT
Roger
## Comment 3481
Date: 2005-06-25 22:49:28 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
This is known and worked around by using BAT scripts that
override the prefix variable from the MonetDB.conf file.
The BAT scripts are installed in the top-most folder in the
Program Files folder whereas the EXE files which they call
are "hidden" in a subfolder.
The reason for doing it this way is that it is too much
effort (because of the way MonetDB is built) to do it
differently. Note that the location of the MonetDB.conf
file is built-in and wrong, so this too is overridden by the
BAT scripts.
## Comment 3482
Date: 2005-06-25 22:55:40 +0200
From: @grobian
Logged In: YES
user_id=963970
isn't it better to leave it out in the distribution? It
looks quite confusing.
## Comment 3483
Date: 2005-06-25 23:01:46 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Perhaps, but then submit it as an RFE. The way it is now is
not a bug.
## Comment 3484
Date: 2005-12-19 11:50:22 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: NO TEST / CONFIGURATION
## Comment 3485
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1227493 at http://sourceforge.net/support/tracker.php?aid=1227493
| Windows: config file contains developer's paths | https://api.github.com/repos/MonetDB/MonetDB/issues/620/comments | 0 | 2020-11-30T08:45:38Z | 2024-06-28T13:35:23Z | https://github.com/MonetDB/MonetDB/issues/620 | 753,279,620 | 620 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-25 08:29:07 +0200
From: @MarcinZukowski
To: @mlkersten
Version: -- development
Last updated: 2005-06-25 08:39:41 +0200
## Comment 3474
Date: 2005-06-25 20:29:07 +0200
From: @MarcinZukowski
Try this:
var int := 1 ;
printf("int=%d\n",int);
it prints:
int=5
I think the type names should be reserved keywords.
## Comment 3475
Date: 2005-06-25 20:38:52 +0200
From: @grobian
Logged In: YES
user_id=963970
This is already 'fixed' in the 'next release'.
Monet Database Server V4.99.19
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for i686-redhat-linux-gnu/32bit; dynamically linked.
config:/ufs/fabian/scratch/monet5/program/etc/MonetDB5.conf
dbfarm:/ufs/fabian/scratch/monet5/program/var/MonetDB5/dbfarm
dbname:demo
Visit http://monetdb.cwi.nl/ for further information.
>int := 1;
>io.printf("int=%d\n", int);
int=1
>
## Comment 3476
Date: 2005-06-25 20:39:40 +0200
From: @mlkersten
Logged In: YES
user_id=490798
This a remnant of a design decision taken in MIL and is used
to get the underlying type index. It won;t be resolved
before MonetDB version 5, where a clear distinction is made
between variable identifiers and type identifiers.
## Comment 3477
Date: 2005-06-25 23:14:50 +0200
From: @mlkersten
Logged In: YES
user_id=490798
This a remnant of a design decision taken in MIL and is used
to get the underlying type index. It won;t be resolved
before MonetDB version 5, where a clear distinction is made
between variable identifiers and type identifiers.
## Comment 3478
Date: 2005-12-19 11:48:58 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: TEST ADDED / WONT FIX
MonetDB/tests/BugDay_2005-12-19_4.9.3/Tests/type_names_are_not_keywords.SF-1227464.*
Note, this test fails, hence should probably not be enabled
## Comment 3479
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1227464 at http://sourceforge.net/support/tracker.php?aid=1227464
| type names are not keywords | https://api.github.com/repos/MonetDB/MonetDB/issues/619/comments | 0 | 2020-11-30T08:45:35Z | 2024-06-28T13:35:22Z | https://github.com/MonetDB/MonetDB/issues/619 | 753,279,580 | 619 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-22 01:42:26 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @sjoerdmullender
Version: -- development
Last updated: 2006-06-06 08:51:31 +0200
## Comment 3456
Date: 2005-06-22 13:42:26 +0200
From: Steffen Goeldner <<sgoeldner>>
Running
use DBI();
my $dsn = 'dbi:ODBC:driver=MonetDB ODBC Driver';
my $dbh = DBI->connect( $dsn ) or die $DBI::errstr;
$dbh->{RaiseError} = 1;
$dbh->{PrintError} = 0;
my $sql = 'select * from env where name = ?';
my $sth = $dbh->prepare( $sql );
$sth->bind_param( 1, undef ); NULL
$sth->execute;
shows
DBD::ODBC::st execute failed:
[MonetDB][ODBC Driver 1.0]
!cannot have parameters (?) in normal queries,
use PREPARE; (SQL-42000)
Looks like Mapi is unable to bind a NULL;
mapi_param_store(MapiHdl hdl)
...
if (hdl->params[i].inparam == 0) {
hdl->query[k++] = PLACEHOLDER;
...
## Comment 3457
Date: 2005-06-23 13:59:31 +0200
From: @grobian
Logged In: YES
user_id=963970
ok.
I suppose you somehow add type information to the NULL value
you send out.
If you do so, however, you get a crash anyway:
exec 1(cast(null as varchar(1024)));
MonetDB>Segmentation fault
work in progress ;)
## Comment 3458
Date: 2005-06-23 15:42:54 +0200
From: @grobian
Logged In: YES
user_id=963970
I think this is now fixed in CVS current:
% MapiClient -ls
sql>prepare select * from env where name = ?;
10 querytype
prepare, prepare, prepare table_name
type, digits, scale name
varchar, int, int type
0, 0, 0 length
1 tuplecount
1 id
[ "varchar", 1024, 0 ]
sql>exec 1(null);
3 querytype
env, env table_name
name, value name
varchar, varchar type
0, 0 length
0, tuplecount
1, id
sql>
## Comment 3459
Date: 2005-06-23 17:49:21 +0200
From: @grobian
Logged In: YES
user_id=963970
This bug doesn't deal with prepare/exec functionality of the
backend. As such, by coincidence a bug with the same effect
on the output was fixed in the backend. The status of this
bug is unaffected.
## Comment 3460
Date: 2005-06-24 17:06:30 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
IMO, mapi_param_store should substitute the
PLACEHOLDER with NULL/nil, see the attached patch.
## Comment 3461
Date: 2005-07-04 14:18:49 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
In the light of http://sourceforge.net/support/tracker.php?aid=1229018, I suggest a revised patch:
--- src/mapi/clients/C/Mapi.mx 4 Jul 2005 10:59:22 -0000
1.156.2.10
+++ src/mapi/clients/C/Mapi.mx 4 Jul 2005 12:10:46 -0000
@@ -2519,8 +2519,8 @@
hdl->query[k] = 0;
if (hdl->params[i].inparam == 0) {
- hdl->query[k++] = PLACEHOLDER;
- hdl->query[k] = 0;
+ checkSpace(5);
+ strcpy(hdl->query + k, hdl->mid->languageId
== LANG_SQL ? "NULL" : "nil");
} else {
void *src = hdl->params[i].inparam; /*
abbrev */
@@ -2604,9 +2604,8 @@
strcpy(hdl->query + k, src);
break;
}
- k += strlen(hdl->query + k);
}
-
+ k += strlen(hdl->query + k);
i++;
p = q + 1;
q = strchr(p, PLACEHOLDER);
## Comment 3462
Date: 2005-07-04 15:39:19 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
The problem is that currently ODBC (through the good offices
of Mapi) does the parameter stuff itself instead of having
the server do the work.
As to your patch, does that really work? If I try this query:
select * from tables where query = NULL;
I get an error from the server, whereas when I try this query:
select * from tables where query = cast(NULL as varchar(1));
I get the expected result (the 1 can be another number
also). It seems to me that your patch produces the first of
these two queries (when you try select * from tables where
query = ? with parameter value NULL).
What should happen is that the NULL is translated into the
correct cast(NULL as <TYPE>) depending on the type of the
parameter. This is really something that ODBC should do,
not Mapi, but currently the translation from ? to value
happens in Mapi. (Actually, this is something that the
server should do, but we're not there yet. That requires a
change to the protocol. We're working on that, but it's
slow going with all the other stuff that needs doing.)
## Comment 3463
Date: 2005-07-05 09:21:15 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
It works (and should work, IMHO) for INSERT/UPDATE:
insert into x values( 1, null );
update x set s = null where i = 2;
I choose the SELECT example as the smallest test case that
demonstrates the current behavior (where NULLs are not
bound).
Of course, I don't expect that 'WHERE s = NULL' works (at
least in a portable way), even if the parser would accept it.
DBI devotes quite a large section on this topic:
<http://search.cpan.org/~timb/DBI-
1.48/DBI.pmPlaceholders_and_Bind_Values>
## Comment 3464
Date: 2005-07-05 09:41:48 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
Last night I came up with the idea that the ODBC driver
could use the server to figure out information about the
parameters. The server now understands the PREPARE
statement (you just put "prepare" in front of a valid SQL
statement-with-parameters) and it returns a table with
information about the parameters. Most importantly, it
returns the expected type.
Even if we then do the filling in of the parameters
ourselves (in ODBC or Mapi), we at least know the correct
type and can convert any values to that type.
But then again, this may be overkill in light of the text
you referred to.
So perhaps for now we should just accept your fix.
## Comment 3465
Date: 2005-07-15 11:05:14 +0200
From: @njnes
Logged In: YES
user_id=43556
patch applied. Could you test the result?
## Comment 3466
Date: 2005-07-15 13:52:04 +0200
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
parse error in: "select * from env where name = ;"
You omitted one line:
@@ -2521,7 +2521,6 @@
if (hdl->params[i].inparam == 0) {
checkSpace(5);
strcpy(hdl->query + k, hdl-
>mid->languageId == LANG_SQL ? "NULL" : "nil");
- hdl->query[k] = 0;
} else {
void *src = hdl->params
[i].inparam; /* abbrev */
## Comment 3467
Date: 2005-12-19 11:41:33 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: NO TEST / Perl-DBD-ODBC -> NO CLUE
Is it possible to have an ODBC thing running without having
to register it somewhere? Is there a HAVE_PERL or HAVE_DBD
thing for TestWeb? Maybe a new ODBC C sample should be
written to cover this.
## Comment 3468
Date: 2005-12-19 18:12:31 +0100
From: @drstmane
Logged In: YES
user_id=572415
see also
https://sourceforge.net/tracker/?func=detail&atid=482468&aid=1229018&group_id=56967
## Comment 3469
Date: 2006-01-15 15:14:40 +0100
From: @grobian
Logged In: YES
user_id=963970
Is this bug fixed in the meanwhile?
## Comment 3470
Date: 2006-02-14 10:41:21 +0100
From: @grobian
Logged In: YES
user_id=963970
ping/bump
## Comment 3471
Date: 2006-02-14 11:52:42 +0100
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
Works for me:
my $sql = q(insert into env values('x',?));
my $sth = $dbh->prepare( $sql );
$sth->bind_param( 1, undef ); NULL
$sth->execute;
$sql = 'select * from env where value = ?';
$sth = $dbh->prepare( $sql );
$sth->bind_param( 1, undef ); NULL
$sth->execute;
while ( my $r = $sth->fetch ) { print @$r, "\n"; }
gives:
x
## Comment 3472
Date: 2006-06-06 20:51:31 +0200
From: @grobian
Logged In: YES
user_id=963970
bug is fixed. No idea on how to add a test as we now use
server side prepares.
## Comment 3473
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1225455 at http://sourceforge.net/support/tracker.php?aid=1225455
| Unable to bind NULL | https://api.github.com/repos/MonetDB/MonetDB/issues/618/comments | 0 | 2020-11-30T08:45:31Z | 2024-06-27T11:04:56Z | https://github.com/MonetDB/MonetDB/issues/618 | 753,279,541 | 618 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-16 10:15:57 +0200
From: Jan Rittinger <<jan.rittinger>>
To: @njnes
Version: -- development
Last updated: 2007-05-29 02:05:54 +0200
## Comment 3450
Date: 2005-06-16 22:15:57 +0200
From: Jan Rittinger <<jan.rittinger>>
I shredded a 1GB XML document. If I shut down Mserver
and restart it the committed, non-empty BATs of the 1
GB document are set to 'diffs' in the dirty field.
(suffixes: 0 = 11MB, 1 = 110MB, 2 = 3 = 1 GB)
21:49:16 rittinge@sn-1:~/pathfinder/Linux$ echo 'ls();' |
Mserver
Monet Database Server V4.9.0
Copyright (c) 1993-2005, CWI. All rights reserved.
Compiled for x86_64-redhat-linux-gnu/64bit with 32bit
OIDs; dynamically linked.
Visit http://monetdb.cwi.nl/ for further information.
MonetDB>-----------------------------------------------------------------
--------------------------------------------------------
name htype ttype count heat
dirty status kind refcnt lrefcnt name
str str str lng int str
str str int int type
---------------------------------------------------------------------------------
----------------------------------------
[ "attr_own0", "void", "oid", 38266, 0,
"clean", "disk", "pers", 0, 1 ]
[ "attr_own1", "void", "oid", 381870, 0,
"clean", "disk", "pers", 0, 1 ]
[ "attr_own2", "void", "oid", 3829768, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "attr_own3", "void", "oid", 3829768, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "attr_prop0", "void", "oid", 38266, 0,
"clean", "disk", "pers", 0, 1 ]
[ "attr_prop1", "void", "oid", 381870, 0,
"clean", "disk", "pers", 0, 1 ]
[ "attr_prop2", "void", "oid", 3829768, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "attr_prop3", "void", "oid", 3829768, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "attr_qn0", "void", "oid", 38266, 0,
"clean", "disk", "pers", 0, 1 ]
[ "attr_qn1", "void", "oid", 381870, 0,
"clean", "disk", "pers", 0, 1 ]
[ "attr_qn2", "void", "oid", 3829768, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "attr_qn3", "void", "oid", 3829768, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "doc_height", "oid", "int", 4, 0,
"clean", "disk", "pers", 0, 1 ]
[ "doc_location", "oid", "str", 4, 0,
"clean", "disk", "pers", 0, 1 ]
[ "doc_name", "oid", "str", 4, 0,
"clean", "disk", "pers", 0, 1 ]
[ "doc_size", "oid", "lng", 4, 0,
"clean", "disk", "pers", 0, 1 ]
[ "doc_timestamp", "oid", "null", 4, 0,
"clean", "disk", "pers", 0, 1 ]
[ "id_pre0", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "id_pre1", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "id_pre2", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "id_pre3", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "idref_pre0", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "idref_pre1", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "idref_pre2", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "idref_pre3", "str", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_00", "void", "oid", 167864, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_01", "void", "oid", 1666310, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_02", "void", "oid", 16703210, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "kind_pre_03", "void", "oid", 16703210, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "kind_pre_10", "void", "oid", 304816, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_11", "void", "oid", 3024328, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_12", "void", "oid", 30312003, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "kind_pre_13", "void", "oid", 30312003, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "kind_pre_20", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_21", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_22", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_23", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_30", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_31", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_32", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_33", "void", "oid", 0, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_40", "void", "oid", 1, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_41", "void", "oid", 1, 0,
"clean", "disk", "pers", 0, 1 ]
[ "kind_pre_42", "void", "oid", 1, 0,
"diffs", "disk", "pers", 0, 1 ]
[ "kind_pre_43", "void", "oid", 1, 0,
"diffs", "disk", "pers", 0, 1 ]
...
## Comment 3451
Date: 2005-07-15 13:58:50 +0200
From: @njnes
Logged In: YES
user_id=43556
A bbp.dir file will now only save 'persistent' information.
So the incorrect BBPSWAPPED isn't used on the next run, ie
making all bats correctly 'on disk' again.
## Comment 3452
Date: 2005-12-19 11:32:23 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: NO TEST / NO CLUE
Please someone else with a clue on how to test this
properly, do so.
## Comment 3453
Date: 2006-12-20 04:20:04 +0100
From: Pseudo user for Sourceforge import <<sfimport>>
---- Original comment by: sf-robot@users.sourceforge.net ----
Logged In: YES
user_id=1312539
Originator: NO
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 365 days (the time period specified by
the administrator of this Tracker).
## Comment 3454
Date: 2007-05-29 14:05:54 +0200
From: @njnes
Logged In: YES
user_id=43556
Originator: NO
many tests with ls() output exist, ie tests exist, closing now.
## Comment 3455
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1222165 at http://sourceforge.net/support/tracker.php?aid=1222165
| property dirty is set to \'diffs\' directly after startup | https://api.github.com/repos/MonetDB/MonetDB/issues/617/comments | 0 | 2020-11-30T08:45:28Z | 2024-06-27T11:04:55Z | https://github.com/MonetDB/MonetDB/issues/617 | 753,279,513 | 617 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-16 06:02:49 +0200
From: Wouter Alink <<alink>>
To: XQuery devs <<bugs-xquery>>
Version: -- development
Last updated: 2005-07-21 02:47:17 +0200
## Comment 3444
Date: 2005-06-16 18:02:49 +0200
From: Wouter Alink <<alink>>
I haven't seen this one before:
WARNING: BATpropcheck: BAT tmp_732 was incorrectly
marked keyed!
It was generated by using the following query:
element { "result" } {
for $q in doc("result.xml")//folder except
doc("result.xml")//folder//folder
where $q//*[contains(.,"aap")]
return $q/@name
}
when it is rewritten as:
element { "result" } {
let $a := doc("result.xml")
for $q in $a//folder except $a//folder//folder
where $q//*[contains(.,"aap")]
return $q/@name
}
it returns:
ERROR: err:XQDY0025: attribute names are not unique in
constructed element 'result'.
which is probably correct.
## Comment 3445
Date: 2005-06-16 22:59:11 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
The document result.xml could be useful to find the bug :)
## Comment 3446
Date: 2005-07-21 12:52:11 +0200
From: @drstmane
Logged In: YES
user_id=572415
I totally agree with Jan ... ;-)
## Comment 3447
Date: 2005-07-21 14:47:17 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
i tried the queries again on the test-set and it seemed like
the error has disappeared. maybe some other bug-fix fixed
this bug?
i close it for now
(i could not provide the test-document, because of privacy...)
## Comment 3448
Date: 2005-12-19 11:29:38 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: NO TEST / INVALID
## Comment 3449
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1222021 at http://sourceforge.net/support/tracker.php?aid=1222021
The original assignee of this bug does not have
an account here. Reassigning to the default assignee
for the component, bugs-xquery@monetdb.org.
Previous assignee was nobody@users.sourceforge.net.
| XQuery: duplicate attributes | https://api.github.com/repos/MonetDB/MonetDB/issues/616/comments | 0 | 2020-11-30T08:45:25Z | 2024-06-28T13:35:22Z | https://github.com/MonetDB/MonetDB/issues/616 | 753,279,485 | 616 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-16 04:51:48 +0200
From: Wouter Alink <<alink>>
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2006-01-18 01:27:21 +0100
## Comment 3434
Date: 2005-06-16 16:51:48 +0200
From: Wouter Alink <<alink>>
element { " " } { "test" }
returns:
< >test</ >
## Comment 3435
Date: 2005-06-16 17:06:29 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
after a short tour over the keyboard:
element {
"/:~`',.;[]{}|<>+_-=!@$%^*()&"'><" } {
"test" }
## Comment 3436
Date: 2005-06-16 17:40:24 +0200
From: @grobian
Logged In: YES
user_id=963970
you forgot the most interesting output:
</:~`',.;[]{}|<>+_-=!@$%^*()&"'><>test<//:~`',.;[]{}|<>+_-=!@$%^*()&"'><>
:)
## Comment 3437
Date: 2005-06-16 21:41:06 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
As there is no validity check for tagnames there is nothing
surprising here.
If someone is willing to add a check, the function castQName
in line 2466 of milprint_summer.c is the right place :)
## Comment 3438
Date: 2005-11-25 02:06:10 +0100
From: @peterboncz
Logged In: YES
user_id=591107
fixed; implemented invalid_qname() test
## Comment 3439
Date: 2005-12-19 11:28:14 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: TEST ADDED / NO CLUE
xquery/tests/BugDay_2005-12-19_0.9.3/Tests/tag_name_validation.SF-1221984.*
Using MapiClient by hand, the added test results in:
% MapiClient -lxq tag_name_validation.SF-1221984.xq
ERROR: err:XPTY0004: illegal qname ' '
using mtest this error doesn't show up anywhere. I have no
idea whether the above error is correct, given the bug
report, I do suspect so... Someone needs to check the
output and check in the correct stable output.
## Comment 3440
Date: 2006-01-14 21:15:52 +0100
From: @drstmane
Logged In: YES
user_id=572415
Question for the XQuery experts:
Is(n't?) the error message the correct/expected output for
this query?
If so (my layman's guess), we can approve the current output
as stable, and close this report.
If not, what is the correct/expected output, and who's going
to fix it?
see also
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_BugDay_2005-12-19_0.9.3/tag_name_validation.SF-1221984.err.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_BugDay_2005-12-19_0.9.3/tag_name_validation.SF-1221984.out.00.html
## Comment 3441
Date: 2006-01-18 09:47:11 +0100
From: @drstmane
Logged In: YES
user_id=572415
Sorry, I provided the wrong URLs.
The correct ones are
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora4/tests_BugDay_2005-12-19_0.9.3/tag_name_validation.SF-1221984.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.64.64.d-Fedora4/tests_BugDay_2005-12-19_0.9.3/tag_name_validation.SF-1221984.err.00.html
... any verdict on the correctness of the currect output ?
## Comment 3442
Date: 2006-01-18 13:27:21 +0100
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
approved output
## Comment 3443
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1221984 at http://sourceforge.net/support/tracker.php?aid=1221984
| XQuery: tag name validation | https://api.github.com/repos/MonetDB/MonetDB/issues/615/comments | 0 | 2020-11-30T08:45:22Z | 2024-06-27T11:04:53Z | https://github.com/MonetDB/MonetDB/issues/615 | 753,279,455 | 615 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-15 05:59:20 +0200
From: Wouter Alink <<alink>>
To: Jan Rittinger <<jan.rittinger>>
Version: -- development
Last updated: 2006-01-15 11:45:51 +0100
## Comment 3423
Date: 2005-06-15 17:59:20 +0200
From: Wouter Alink <<alink>>
Using Pathfinder 0.8 CVS
ERROR: merged_union: tail of first BAT must be sorted.
let $a := (<a/>,<a/>)
for $i in 0 to 3
let $b := zero-or-one($a[$i])
return element { name($b) } { $b/@* }
## Comment 3424
Date: 2005-06-16 13:58:17 +0200
From: @drstmane
Logged In: YES
user_id=572415
Wouter,
this is not an merged_union problem, but a property
problem "somewhere" else in MonetDB.
I can only check it once I'm back in A'dam last week of June.
For now, please try to run Mserver with property checking
enabled; i.e., `Mserver --debug=10`.
Stefan.
## Comment 3425
Date: 2005-06-16 16:50:58 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
If I use the option: --debug=10
instead of the error, the following result is returned:
</>,
<a/>,
<a/>,
</>
this is the correct answer, although it is invalid. i am not
sure what the problem is, but i guess that the
element-constructor should not accept the empty string (or a
string with spaces) as the xml-tag-name.
## Comment 3426
Date: 2005-06-16 23:02:39 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
it is now fixed by adding more chk_order calls, and still throws
no error message because of empty element name
This bug could have been avoided if order properties are
exploited even more :)
x: sorted|sorted
y: sorted|sorted
x.join(y): sorted|sorted
## Comment 3427
Date: 2005-07-21 13:06:45 +0200
From: @drstmane
Logged In: YES
user_id=572415
Even in this case, correct property propagation is a bit
more compilcated than it might seem:
x: sorted|sorted
y: sorted|sorted
==>
x.join(y): sorted|sorted
only holds if
(1)
the respective join algorithm scans one of the BATs (say x)
sequentially, looking up the matches in the other BAT (say y);
and
(2)
with duplicate values in the join column/attribute of the
looked-up BAT (y), results are produced in the original order;
and
(3)
no join-attribute value occurs multiple times in both BATs
(x & y).
In our case (MonetDB), (1) hold for all join algorithms
(AFAIK), (2) holds for mreg-join, but not for hash-join, and
(3) could only be approximated by "(3') (at least) one of
the two join columns is unique (key)"...
... I'll check, whether our join algorithms do exploit (3') ...
## Comment 3428
Date: 2005-11-25 00:23:29 +0100
From: @peterboncz
Logged In: YES
user_id=591107
it seems this is fixed. Wouter, can you confirm this?
## Comment 3429
Date: 2005-12-19 11:10:18 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: TEST ADDED / NO CLUE
pathfinder/tests/BugDay_2005-12-19_0.9.3/Tests/zero-or-one_merged-union.SF-1221336.*
Wouter, can you check the output of this test? It returns
nothing for me.
## Comment 3430
Date: 2006-01-14 21:36:14 +0100
From: @drstmane
Logged In: YES
user_id=572415
The original version of the query results in a
ERROR: err:XPTY0004: illegal qname ''
which is IMHO correct, since in the first iteration ($i==0),
$b=zero-or-one($a[0]) is empty (XQuery indices start at 1,
not at 0 ??), hence, name($b) is empty, which is indeed not
a legal qname.
I added a second test query that iterates only over the
existing items in $a (1 to 2)
let $a := (<a/>,<a/>)
for $i in 1 to 2
let $b := zero-or-one($a[$i])
return element { name($b) } { $b/@* }
and that one nicely (and correctly, right?) returns
<?xml version="1.0" encoding="utf-8"?>
<XQueryResult>
<a/><a/>
</XQueryResult>
Do the XQuery expert agree??
If so, we should approve
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_BugDay_2005-12-19_0.9.3/zero-or-one_merged-union.SF-1221336.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/pathfinder/.mTests103/GNU.32.32.d-Fedora4/tests_BugDay_2005-12-19_0.9.3/zero-or-one_merged-union.SF-1221336.err.00.html
and close this report...
## Comment 3431
Date: 2006-01-15 11:22:47 +0100
From: @Teggy
Logged In: YES
user_id=642632
Yes, the output looks OK to me. (In an XQuery sequence $s
of the form (x,y,z), item x is $s[1], not $s[0].)
--Teggy
## Comment 3432
Date: 2006-01-15 11:45:51 +0100
From: @drstmane
Logged In: YES
user_id=572415
The exquery expert does agree; hence, we can close this one.
## Comment 3433
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1221336 at http://sourceforge.net/support/tracker.php?aid=1221336
| XQuery: merged_union | https://api.github.com/repos/MonetDB/MonetDB/issues/614/comments | 0 | 2020-11-30T08:45:19Z | 2024-06-27T11:04:52Z | https://github.com/MonetDB/MonetDB/issues/614 | 753,279,416 | 614 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-15 03:00:16 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @njnes
Version: -- development
Last updated: 2005-06-24 02:50:23 +0200
## Comment 3419
Date: 2005-06-15 15:00:16 +0200
From: Steffen Goeldner <<sgoeldner>>
Running
create table t(i int);
prepare insert into t(i) values (?);
results in a server crash.
## Comment 3420
Date: 2005-06-24 14:50:23 +0200
From: @grobian
Logged In: YES
user_id=963970
On CVS HEAD this currently works fine:
% MapiClient -ls
sql> create table t(i int);
6 querytype
sql> prepare insert into t(i) values (?);
10 querytype
prepare, prepare, prepare table_name
type, digits, scale name
varchar, int, int type
0, 0, 0 length
1 tuplecount
0 id
[ "int", 9, 0 ]
sql>exec 0(1);
4 querytype
[ 1 ]
sql>
So I assume this has been fixed in the mean time.
## Comment 3421
Date: 2005-12-19 10:47:06 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: ALREADY IN TESTWEB
amongst others:
sql/src/test/BugTracker/Tests/large_prepare.SF-1363729.*
## Comment 3422
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1221151 at http://sourceforge.net/support/tracker.php?aid=1221151
| Crash: prepare insert into ... | https://api.github.com/repos/MonetDB/MonetDB/issues/613/comments | 0 | 2020-11-30T08:45:16Z | 2024-06-27T11:04:51Z | https://github.com/MonetDB/MonetDB/issues/613 | 753,279,379 | 613 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-15 01:44:02 +0200
From: Jan Rittinger <<jan.rittinger>>
To: Stefan Manegold <<Stefan.Manegold>>
Version: -- development
Last updated: 2005-07-04 10:59:08 +0200
## Comment 3407
Date: 2005-06-15 13:44:02 +0200
From: Jan Rittinger <<jan.rittinger>>
When I tried to load a 1GB document (xmlgen -f10) with
shred_doc() it does not finish. I tested it on the 64-bit
machine at CWI (since I first thought it could be a ram
problem) and on the 32-bit suse machine in KN.
In both cases Mserver starts with 100% CPU utilization
and falls back to 1% after a minute. After half an hour it
then wrote something to disk (and kept idling?).
Could bug [ 1207707 ] 'XQuery: shredding attributes
takes time' be the reason?
## Comment 3408
Date: 2005-06-15 13:50:36 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
it just finished:
Shredded XML doc("auction10MB.xml"), total time after
commit=0.859s
Shredded XML doc("auction100MB.xml"), total time after
commit=8.039s
Shredded XML doc("auction1GB.xml"), total time after
commit=1506.999s
... it just finished on the faster 64-bit machine
## Comment 3409
Date: 2005-06-15 13:50:48 +0200
From: Wouter Alink <<alink>>
Logged In: YES
user_id=621590
I don't think the "shredding attributes"-bug is related, but
maybe the "bus error"-bug, because the "Bus Error" message
was only raised after several minutes (of which
CPU-utilization is only 100% in the first few seconds)
## Comment 3410
Date: 2005-06-16 11:48:02 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
I had a quick look at the problem and I had the same
symptons. I am not really an expert in virtual memory but I
first thought the system was swapping. But when I ran vmstat
it showed no pageswaps in/or out. I only saw a modest amount
of bi (blocks in) beween 400 and 700. Furthermore vmstat
said that of my 1G memory 250M was active and 750M was
inactive!!!! In the coming days I will have a deeper look at
the problem but I think the problem could be caused by Monet
memory management and especially the way we (I?) use it!
## Comment 3411
Date: 2005-06-16 13:55:05 +0200
From: @drstmane
Logged In: YES
user_id=572415
Jan^2,
could you please change the memory settings of MonetDB
and try again:
Before calling shred_doc, call the following
on a 1 GB machine:
vm_minsize ( 500000000LL);
mem_bigsize( 500000000LL);
mem_maxsize(1000000000LL);
and on a 8GB machine:
vm_minsize(4000000000LL);
mem_bigsize(4000000000LL);
mem_maxsize(8000000000LL);
Please let us know, whether this helps.
Stefan.
## Comment 3412
Date: 2005-06-16 14:29:10 +0200
From: Jan Flokstra <<jflokstra>>
Logged In: YES
user_id=1054297
THIS WORKS! For me this does the trick. It even starts
swapping after a while. My figures are on a 1G machine:
auction100M=6s, auction1G=260s. Hope it also works for
sqrt(Jan^2).
What should we do/conclude concerning this bug? A message in
the manual about working with very large data sets??
## Comment 3413
Date: 2005-06-16 16:56:54 +0200
From: @drstmane
Logged In: YES
user_id=572415
Yes, for now, a hint in the docu/HowToStart might be good.
Eventually, we should make Mserver set/use the respective
values by default:
vm_minsize & mem_bigsize: ~50% of the physically
available memory;
mem_maxsize: ~100% of the physically available memory.
... I should have done this before the release, but it
apparently got lost ...
## Comment 3414
Date: 2005-06-16 16:58:24 +0200
From: @drstmane
Logged In: YES
user_id=572415
ps:
We would actually have to verify, first, whether these settings
also work fine for other OSs than Linux...
## Comment 3415
Date: 2005-06-16 22:09:17 +0200
From: Jan Rittinger <<jan.rittinger>>
Logged In: YES
user_id=993208
Stefan,
your comment helped:
Shredded XML doc("auction1GB.xml2"), total time after
commit=88.563s
But there is another open point. The same problem happens
for querying the 1GB document and we can't ask users to set
this property. Somebody who just uses our system will
conclude that the system is NOT able to evaluate larger
documents and therefore also does NOT scale :(
The default loading of this values therefore is important (in my
opinion).
## Comment 3416
Date: 2005-07-04 10:59:08 +0200
From: @drstmane
Logged In: YES
user_id=572415
fixed in CVS (MonetDB_4-8 release branch):
by default, Mserver now sets
mem_maxsize to 95% of the physically available amount of
main-memory, and
vm_minsize and mem_bigsize to 1/2 of mem_maxsize.
## Comment 3417
Date: 2005-12-19 10:42:15 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: NO TEST / TOO LARGE
## Comment 3418
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1221107 at http://sourceforge.net/support/tracker.php?aid=1221107
| XQuery: shred_doc does take to long for 1GB document | https://api.github.com/repos/MonetDB/MonetDB/issues/612/comments | 0 | 2020-11-30T08:45:13Z | 2024-06-27T11:04:50Z | https://github.com/MonetDB/MonetDB/issues/612 | 753,279,340 | 612 |
[
"MonetDB",
"MonetDB"
] | Date: 2005-06-10 05:14:28 +0200
From: Steffen Goeldner <<sgoeldner>>
To: @sjoerdmullender
Version: -- development
Last updated: 2006-01-20 01:26:36 +0100
## Comment 3398
Date: 2005-06-10 17:14:28 +0200
From: Steffen Goeldner <<sgoeldner>>
OS: Windows 2000
Compiler: MS VC++ Toolkit 2003
Parser tools: UnxUtils
Start the server:
bin\Mserver.exe -c etc\MonetDB.conf --dbinit="module
(mapi); mapi_start();" --set gdk_debug=128
Start a client:
bin\MapiClient.exe -c etc\MonetDB.conf -s quit();
The server exits with:
...
MAPIlisten:client accepted 704
------------------------------------------------------------------
[ 0, 2, running, 0, "listen(get_mapi_port
(),5," ]
[ 1, 1, running, 0, "unknown(0)
" ]
[ 2, 0, running, 0, "unknown(1)
" ]
MIget = unknown(1)
setchannel Mserver mapi read Mserver mapi write
------------------------------------------------------------------
req | tid | status | wakeup | action
------------------------------------------------------------------
[ 0, 2, running, 0, "listen(get_mapi_port
(),5," ]
[ 1, 1, running, 0, "unknown(0)
" ]
[ 3, 0, 1, 0, "unknown(1)
" ]
[ 4, 0, running, 3, "quit()
" ]
MIget = quit()
setchannel Mserver mapi read Mserver mapi write
Assertion failed: s->access == ST_WRITE,
file .\..\..\..\src\common\stream.mx, line 414
## Comment 3399
Date: 2005-06-24 14:54:53 +0200
From: @grobian
Logged In: YES
user_id=963970
I can not reproduce this problem on Linux using todats
current branch
## Comment 3400
Date: 2005-07-05 10:04:12 +0200
From: @sjoerdmullender
Logged In: YES
user_id=43607
I can reproduce this. Are you sure you have assertions enabled?
The problem is a bit tough to fix, though. When the client
exits, the server thread is destroyed, and with it the
streams to and from the client. But after this happens
(call to freeClient() in closeClient()), a request is still
sent to the thread to kill itself (putkillRequest()). This
latter call tries to print to the stream that was just
closed if PARDEBUG is set (debugmask=128).
## Comment 3401
Date: 2005-12-19 10:41:33 +0100
From: @grobian
Logged In: YES
user_id=963970
BugDay_2005-12-19, meltdown: TEST ADDED / NO CLUE
MonetDB/tests/BugDay_2005-12-19_4.9.3/Tests/stream_assert_quit.SF-1218291.*
What's the status of this bug?
## Comment 3402
Date: 2006-01-14 21:48:32 +0100
From: @drstmane
Logged In: YES
user_id=572415
Any news?
The respective test seems to work fine on all our testing
platforms:
http://monetdb.cwi.nl/testing/projects/monetdb/Current/MonetDB/.mTests103/GNU.64.64.d-Fedora4/tests_BugDay_2005-12-19_4.9.3/stream_assert_quit.SF-1218291.out.00.html
http://monetdb.cwi.nl/testing/projects/monetdb/Current/MonetDB/.mTests103/GNU.64.64.d-Fedora4/tests_BugDay_2005-12-19_4.9.3/stream_assert_quit.SF-1218291.err.00.html
## Comment 3403
Date: 2006-01-18 09:49:16 +0100
From: @drstmane
Logged In: YES
user_id=572415
Steffen,
does this bug still occur with you?
## Comment 3404
Date: 2006-01-18 16:47:06 +0100
From: Steffen Goeldner <<sgoeldner>>
Logged In: YES
user_id=1174087
Yes:
Assertion failed: s->access == ST_WRITE,
file .\..\..\..\src\common\stream.mx, line 421
## Comment 3405
Date: 2006-01-20 13:26:36 +0100
From: @sjoerdmullender
Logged In: YES
user_id=43607
After Stefan's change to write 0xDB in freed memory, this
test now crashes.
The fix is to not attempt to print the queue after the
thread was killed.
Fixed in CVS.
## Comment 3406
Date: 2010-05-04 09:32:09 +0200
From: Pseudo user for Sourceforge import <<sfimport>>
This bug was previously known as tracker item 1218291 at http://sourceforge.net/support/tracker.php?aid=1218291
| Assertion failed: s->access == ST_WRITE: stream.mx, line 414 | https://api.github.com/repos/MonetDB/MonetDB/issues/611/comments | 0 | 2020-11-30T08:45:10Z | 2024-06-27T11:04:49Z | https://github.com/MonetDB/MonetDB/issues/611 | 753,279,317 | 611 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.