before_code stringlengths 30 149k | reviewer_comment stringlengths 16 20.6k | after_code stringlengths 30 163k | diff_context stringlengths 0 22.4k | file_path stringlengths 6 155 | comment_line int64 0 26 | language stringclasses 19
values | quality_score float64 0.07 1 | comment_type stringclasses 9
values | comment_length int64 16 20.6k | before_lines int64 2 4.61k | after_lines int64 2 4.4k | is_negative bool 2
classes | pr_title stringlengths 7 140 | pr_number int64 18 61.2k | repo_name stringclasses 29
values | repo_stars int64 321 56k | repo_language stringclasses 14
values | reviewer_username stringclasses 378
values | author_username stringclasses 739
values | user stringlengths 361 200k | assistant stringlengths 16 20.6k | system stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
/* outputs */
/* .min_valid = */ *min_valid,
/* .max_valid = */ *max_valid,
/* .ti = */ NULL,
/* .env = */ jl_emptysvec,
/* .issubty = */ 0
},
/* inputs */
intersections,
world,
lim,
include_ambiguou... | ```suggestion
// first check the leaf cache if the type might have been put in there
```
There are lot of different definitions of type "size", so it is not a useful term. It is legal and valid to check any type here against the cache, but since we (currently) only construct it with dispatch tuples, it is sl... | }
if (entry) {
// mfunc was found in slow path, so log --trace-dispatch
jl_method_instance_t *mfunc = entry->func.linfo;
record_dispatch_statement_on_first_dispatch(mfunc);
}
}
jl_method_instance_t *mfunc;
if (entry) {
have_entry:
mfunc = ... | @@ -4223,21 +4254,47 @@ static jl_value_t *ml_matches(jl_methtable_t *mt,
else
va = NULL;
}
- struct ml_matches_env env = {{ml_matches_visitor, (jl_value_t*)type, va, /* .search_slurp = */ 0,
- /* .min_valid = */ *min_valid, /* .max_valid = */ *max_valid,
- /* .ti = *... | src/gf.c | 26 | C | 1 | performance | 381 | 51 | 51 | false | gf.c: Add some comments to `ml_matches` (NFC) | 58,304 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: C
File: src/gf.c
PR Title: gf.c: Add some comments to `ml_matches` (NFC)
BEFORE CODE:
/* outputs */
/* .min_valid = */ *min_valid,
/* .max_valid = */ *max_valid,
/* .ti = */ NULL,
/* .env = */ jl_emptysvec,
... | ```suggestion
// first check the leaf cache if the type might have been put in there
```
There are lot of different definitions of type "size", so it is not a useful term. It is legal and valid to check any type here against the cache, but since we (currently) only construct it with dispatch tuples, it is sl... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
env.match.ti = unw;
}
else if (jl_egal((jl_value_t*)type, mi->specTypes)) {
env.match.env = mi->sparam_vals;
env.match.ti = mi->specTypes;
}
else {
// this just calls jl_subtyp... | I have no idea what "small enough" means in this context. I think you're just restating that it seemed profitable? These queries all look to me to have no dependency on being a dispatch tuple, they just become more cumbersome to implement, and thus likely slower and less likely to be profitable. |
return mfunc;
}
JL_DLLEXPORT jl_value_t *jl_apply_generic(jl_value_t *F, jl_value_t **args, uint32_t nargs)
{
size_t world = jl_current_task->world_age;
jl_method_instance_t *mfunc = jl_lookup_generic_(F, args, nargs,
jl_int32hash_fast(jl_return_address... | @@ -4266,10 +4323,13 @@ static jl_value_t *ml_matches(jl_methtable_t *mt,
return env.t;
}
}
- // then check the full cache if it seems profitable
+
+ // then check the full cache if it seems profitable (and the type is small enough) | src/gf.c | 26 | C | 0.429 | performance | 296 | 51 | 51 | false | gf.c: Add some comments to `ml_matches` (NFC) | 58,304 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: C
File: src/gf.c
PR Title: gf.c: Add some comments to `ml_matches` (NFC)
BEFORE CODE: env.match.ti = unw;
}
else if (jl_egal((jl_value_t*)type, mi->specTypes)) {
env.match.env = mi->sparam_vals;
... | I have no idea what "small enough" means in this context. I think you're just restating that it seemed profitable? These queries all look to me to have no dependency on being a dispatch tuple, they just become more cumbersome to implement, and thus likely slower and less likely to be profitable. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
jl_method_instance_t *mi = entry->func.linfo;
jl_method_t *meth = mi->def.method;
if (!jl_is_unionall(meth->sig) && ((jl_datatype_t*)unw)->isdispatchtuple) {
env.match.env = jl_emptysvec;
env.match.ti = unw;
}
... | ```suggestion
// if neither cache worked (or was deemed unprofitable to check) visit all matching methods in the MethodTable
``` | // Deprecated: Use jl_gf_invoke_lookup_worlds for future development
size_t min_valid = 0;
size_t max_valid = ~(size_t)0;
if (mt == jl_nothing)
mt = (jl_value_t*)jl_method_table;
jl_method_match_t *matc = _gf_invoke_lookup(types, (jl_methtable_t*)mt, world, 1, &min_valid, &max_valid);
if... | @@ -4294,16 +4354,24 @@ static jl_value_t *ml_matches(jl_methtable_t *mt,
return env.t;
}
}
+
+ // if neither cache worked (or was eligible / profitable) visit all matching methods in the MethodTable | src/gf.c | 26 | C | 0.714 | performance | 138 | 51 | 51 | false | gf.c: Add some comments to `ml_matches` (NFC) | 58,304 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: C
File: src/gf.c
PR Title: gf.c: Add some comments to `ml_matches` (NFC)
BEFORE CODE: jl_method_instance_t *mi = entry->func.linfo;
jl_method_t *meth = mi->def.method;
if (!jl_is_unionall(meth->sig) && ((jl_datatype_t*)unw)->... | ```suggestion
// if neither cache worked (or was deemed unprofitable to check) visit all matching methods in the MethodTable
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
size_t max_world = jl_atomic_load_relaxed(&entry->max_world);
if (*min_valid < min_world)
*min_valid = min_world;
if (*max_valid > max_world)
*max_valid = max_world;
JL_GC_POP();
return env.t;
... | In v1.12, this will be gone (https://github.com/JuliaLang/julia/pull/58131) | mt = (jl_value_t*)jl_method_table;
jl_method_match_t *matc = _gf_invoke_lookup(types, (jl_methtable_t*)mt, world, 1, min_world, max_world);
if (matc == NULL)
return jl_nothing;
return (jl_value_t*)matc;
}
// invoke()
// this does method dispatch with a set of types to match other than the
/... | @@ -4294,16 +4354,24 @@ static jl_value_t *ml_matches(jl_methtable_t *mt,
return env.t;
}
}
+
+ // if neither cache worked (or was eligible / profitable) visit all matching methods in the MethodTable
+ // via the usual typemap visitor, recording any matches in closur... | src/gf.c | 26 | C | 0.214 | suggestion | 75 | 51 | 51 | false | gf.c: Add some comments to `ml_matches` (NFC) | 58,304 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: C
File: src/gf.c
PR Title: gf.c: Add some comments to `ml_matches` (NFC)
BEFORE CODE: size_t max_world = jl_atomic_load_relaxed(&entry->max_world);
if (*min_valid < min_world)
*min_valid = min_world;
if (*... | In v1.12, this will be gone (https://github.com/JuliaLang/julia/pull/58131) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
// next split may be on Type{T} as LeafTypes then TypeName's parents up to Any
// next split may be on LeafType
// next split may be on TypeName
_Atomic(jl_genericmemory_t*) arg1; // contains LeafType (in a map of non-abstract TypeName)
_Atomic(jl_genericmemory_t*) targ; // contains Type{LeafType} (... | ```suggestion
// cache for querying everything else (anything that didn't seem profitable to put into leafcache)
``` | };
struct _jl_globalref_t {
JL_DATA_TYPE
jl_module_t *mod;
jl_sym_t *name;
jl_binding_t *binding;
};
// one Type-to-Value entry
struct _jl_typemap_entry_t {
JL_DATA_TYPE
_Atomic(struct _jl_typemap_entry_t*) next; // invasive linked list
jl_tupletype_t *sig; // the type signature for this e... | @@ -878,9 +878,17 @@ typedef struct _jl_typemap_level_t {
typedef struct _jl_methtable_t {
JL_DATA_TYPE
jl_sym_t *name; // sometimes used for debug printing
+
+ // full set of entries
_Atomic(jl_typemap_t*) defs;
+
+ // hash map from dispatchtuple type to a linked-list of TypeMapEntry
+ // entr... | src/julia.h | 26 | C/C++ | 0.714 | performance | 122 | 51 | 51 | false | gf.c: Add some comments to `ml_matches` (NFC) | 58,304 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: C/C++
File: src/julia.h
PR Title: gf.c: Add some comments to `ml_matches` (NFC)
BEFORE CODE: // next split may be on Type{T} as LeafTypes then TypeName's parents up to Any
// next split may be on LeafType
// next split may be on TypeName
_Atomic(jl_genericm... | ```suggestion
// cache for querying everything else (anything that didn't seem profitable to put into leafcache)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
```jldoctest
julia> io = IOBuffer();
julia> iob64_decode = Base64DecodePipe(io);
julia> write(io, "SGVsbG8h")
8
julia> seekstart(io);
julia> String(read(iob64_decode))
"Hello!"
```
"""
struct Base64DecodePipe{T <: IO} <: IO
io::T
buffer::Buffer
rest::Vector{UInt8}
function Base64DecodePipe{T}(io::T... | ```suggestion
Base64DecodePipe(io::IO) = Base64DecodePipe{IO}(io)
```
Doesn't seem like this PR makes things faster, so by default it seems better to avoid making the compile slower and more complex? | ```jldoctest
julia> io = IOBuffer();
julia> iob64_decode = Base64DecodePipe(io);
julia> write(io, "SGVsbG8h")
8
julia> seekstart(io);
julia> String(read(iob64_decode))
"Hello!"
```
"""
struct Base64DecodePipe{T <: IO} <: IO
io::T
buffer::Buffer
rest::Vector{UInt8}
function Base64DecodePipe{T}(io::T... | @@ -32,17 +32,19 @@ julia> String(read(iob64_decode))
"Hello!"
```
"""
-struct Base64DecodePipe <: IO
- io::IO
+struct Base64DecodePipe{T <: IO} <: IO
+ io::T
buffer::Buffer
rest::Vector{UInt8}
- function Base64DecodePipe(io::IO)
+ function Base64DecodePipe{T}(io::T) where {T <: IO}
... | stdlib/Base64/src/decode.jl | 26 | Julia | 0.857 | performance | 204 | 51 | 51 | false | Type stabilize Base64 | 58,352 | JuliaLang/julia | 35,573 | Julia | vtjnash | jakobnissen | Review this pull request change.
Language: Julia
File: stdlib/Base64/src/decode.jl
PR Title: Type stabilize Base64
BEFORE CODE: ```jldoctest
julia> io = IOBuffer();
julia> iob64_decode = Base64DecodePipe(io);
julia> write(io, "SGVsbG8h")
8
julia> seekstart(io);
julia> String(read(iob64_decode))
"Hello!"
```
"""... | ```suggestion
Base64DecodePipe(io::IO) = Base64DecodePipe{IO}(io)
```
Doesn't seem like this PR makes things faster, so by default it seems better to avoid making the compile slower and more complex? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# Examples
```jldoctest
julia> io = IOBuffer();
julia> iob64_encode = Base64EncodePipe(io);
julia> write(iob64_encode, "Hello!")
6
julia> close(iob64_encode);
julia> str = String(take!(io))
"SGVsbG8h"
julia> String(base64decode(str))
"Hello!"
```
"""
struct Base64EncodePipe{T <: IO} <: IO
io::T
buffer::Buf... | ```suggestion
pipe = new{T}(io, buffer)
``` | # Examples
```jldoctest
julia> io = IOBuffer();
julia> iob64_encode = Base64EncodePipe(io);
julia> write(iob64_encode, "Hello!")
6
julia> close(iob64_encode);
julia> str = takestring!(io)
"SGVsbG8h"
julia> String(base64decode(str))
"Hello!"
```
"""
struct Base64EncodePipe{T <: IO} <: IO
io::T
buffer::Buffe... | @@ -31,19 +31,21 @@ julia> String(base64decode(str))
"Hello!"
```
"""
-struct Base64EncodePipe <: IO
- io::IO
+struct Base64EncodePipe{T <: IO} <: IO
+ io::T
buffer::Buffer
- function Base64EncodePipe(io::IO)
+ function Base64EncodePipe{T}(io::T) where {T <: IO}
# The buffer size must be ... | stdlib/Base64/src/encode.jl | 26 | Julia | 0.571 | suggestion | 53 | 51 | 51 | false | Type stabilize Base64 | 58,352 | JuliaLang/julia | 35,573 | Julia | vtjnash | jakobnissen | Review this pull request change.
Language: Julia
File: stdlib/Base64/src/encode.jl
PR Title: Type stabilize Base64
BEFORE CODE: # Examples
```jldoctest
julia> io = IOBuffer();
julia> iob64_encode = Base64EncodePipe(io);
julia> write(iob64_encode, "Hello!")
6
julia> close(iob64_encode);
julia> str = String(take!... | ```suggestion
pipe = new{T}(io, buffer)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
julia> write(iob64_encode, "Hello!")
6
julia> close(iob64_encode);
julia> str = String(take!(io))
"SGVsbG8h"
julia> String(base64decode(str))
"Hello!"
```
"""
struct Base64EncodePipe{T <: IO} <: IO
io::T
buffer::Buffer
function Base64EncodePipe{T}(io::T) where {T <: IO}
# The buffer size must be... | ```suggestion
Base64EncodePipe(io::IO) = Base64EncodePipe{IO}(io)
``` | julia> write(iob64_encode, "Hello!")
6
julia> close(iob64_encode);
julia> str = takestring!(io)
"SGVsbG8h"
julia> String(base64decode(str))
"Hello!"
```
"""
struct Base64EncodePipe{T <: IO} <: IO
io::T
buffer::Buffer
function Base64EncodePipe{T}(io::T) where {T <: IO}
# The buffer size must be a... | @@ -31,19 +31,21 @@ julia> String(base64decode(str))
"Hello!"
```
"""
-struct Base64EncodePipe <: IO
- io::IO
+struct Base64EncodePipe{T <: IO} <: IO
+ io::T
buffer::Buffer
- function Base64EncodePipe(io::IO)
+ function Base64EncodePipe{T}(io::T) where {T <: IO}
# The buffer size must be ... | stdlib/Base64/src/encode.jl | 26 | Julia | 0.571 | suggestion | 71 | 51 | 51 | false | Type stabilize Base64 | 58,352 | JuliaLang/julia | 35,573 | Julia | vtjnash | jakobnissen | Review this pull request change.
Language: Julia
File: stdlib/Base64/src/encode.jl
PR Title: Type stabilize Base64
BEFORE CODE: julia> write(iob64_encode, "Hello!")
6
julia> close(iob64_encode);
julia> str = String(take!(io))
"SGVsbG8h"
julia> String(base64decode(str))
"Hello!"
```
"""
struct Base64EncodePipe{T ... | ```suggestion
Base64EncodePipe(io::IO) = Base64EncodePipe{IO}(io)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# This file is a part of Julia. License is MIT: https://julialang.org/license
# Updating this listing is fairly easy, assuming existence of a unix system,
# posix shell, and `awk`. Just call this file with `sh`, and it will be
# updated to the latest terminfo snapshot. This works because the file is a
# bit of a quine... | Isn't this referring to this file?
```suggestion
# updated to the latest terminfo snapshot. This works because this file is a
``` | # This file is a part of Julia. License is MIT: https://julialang.org/license
# Updating this listing is fairly easy, assuming existence of a unix system,
# posix shell, and `awk`. Just call this file with `sh`, and it will be
# updated to the latest terminfo snapshot. This works because this file is a
# bit of a quin... | @@ -1,9 +1,9 @@
# This file is a part of Julia. License is MIT: https://julialang.org/license
# Updating this listing is fairly easy, assuming existence of a unix system,
-# posix shell, and `awk`. Just update the version string in the commented out
-# `NCURSES_VERSION` variable, and run this file. This works becau... | base/terminfo_data.jl | 5 | Julia | 0.786 | question | 132 | 30 | 30 | false | Update terminfo data | 58,406 | JuliaLang/julia | 35,573 | Julia | fingolfin | tecosaur | Review this pull request change.
Language: Julia
File: base/terminfo_data.jl
PR Title: Update terminfo data
BEFORE CODE: # This file is a part of Julia. License is MIT: https://julialang.org/license
# Updating this listing is fairly easy, assuming existence of a unix system,
# posix shell, and `awk`. Just call thi... | Isn't this referring to this file?
```suggestion
# updated to the latest terminfo snapshot. This works because this file is a
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
for (name, shortcode, description) in list
print(out, "\n TermCapability(:", name, ',', ' '^(namepad - textwidth(name)),
':', shortcode, ',', ' '^(codepad - textwidth(shortcode)),
'"', escape_string(description), "\"),")
end
println(out, "\n]")
end
function getcustoma... | Could also do this to make future additions nicer (but obviously that's a matter of taste, too)
```suggestion
"fe" => ":focus_enable",
)
``` | for (name, shortcode, description) in list
print(out, "\n TermCapability(:", name, ',', ' '^(namepad - textwidth(name)),
':', shortcode, ',', ' '^(codepad - textwidth(shortcode)),
'"', escape_string(description), "\"),")
end
println(out, "\n]")
end
function getcustoma... | @@ -132,7 +133,9 @@ function getcustomalias(allterms::Vector{NTuple{3, String}}, type, short, descri
"csl" => ":clear_status_line",
"Ms" => ":set_host_clipboard",
"Tc" => ":truecolor",
- "XF" => ":xterm_focus")
+ "XF" => ":xterm_focus",
+ "fd" => ":focus_... | base/terminfo_data.jl | 26 | Julia | 0.786 | suggestion | 163 | 51 | 51 | false | Update terminfo data | 58,406 | JuliaLang/julia | 35,573 | Julia | fingolfin | tecosaur | Review this pull request change.
Language: Julia
File: base/terminfo_data.jl
PR Title: Update terminfo data
BEFORE CODE: for (name, shortcode, description) in list
print(out, "\n TermCapability(:", name, ',', ' '^(namepad - textwidth(name)),
':', shortcode, ',', ' '^(codepad - textwidth... | Could also do this to make future additions nicer (but obviously that's a matter of taste, too)
```suggestion
"fe" => ":focus_enable",
)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
using .M2
export x
end
@test M57638_3.x === 1
module GlobalBindingMulti
module M
export S
module C
export S
struct A end
S = A() # making S const makes the error go away
end
using .C
end
using .M
using .M.C
end
@test GlobalBin... | ```suggestion
@test Base.datatype_haspadding(Tuple{ByteString58434}) == (length(Base.padding(Tuple{ByteString58434})) > 0)
```
Might be good as a sanity check to ensure that reflection stays consistent with this, since this fluctuated in the past. | using .M2
export x
end
@test M57638_3.x === 1
module GlobalBindingMulti
module M
export S
module C
export S
struct A end
S = A() # making S const makes the error go away
end
using .C
end
using .M
using .M.C
end
@test GlobalBin... | @@ -8534,3 +8534,9 @@ module GlobalBindingMulti
using .M.C
end
@test GlobalBindingMulti.S === GlobalBindingMulti.M.C.S
+
+#58434 bitsegal comparison of oddly sized fields
+primitive type ByteString58434 (18 * 8) end
+
+@test Base.datatype_isbitsegal(Tuple{ByteString58434}) == false
+@test Base.datatype_haspaddin... | test/core.jl | 26 | Julia | 0.929 | suggestion | 253 | 27 | 27 | false | Fix layout flags for types that have oddly sized primitive type fields | 58,435 | JuliaLang/julia | 35,573 | Julia | Seelengrab | gbaraldi | Review this pull request change.
Language: Julia
File: test/core.jl
PR Title: Fix layout flags for types that have oddly sized primitive type fields
BEFORE CODE: using .M2
export x
end
@test M57638_3.x === 1
module GlobalBindingMulti
module M
export S
module C
export S
... | ```suggestion
@test Base.datatype_haspadding(Tuple{ByteString58434}) == (length(Base.padding(Tuple{ByteString58434})) > 0)
```
Might be good as a sanity check to ensure that reflection stays consistent with this, since this fluctuated in the past. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
using .M2
export x
end
@test M57638_3.x === 1
module GlobalBindingMulti
module M
export S
module C
export S
struct A end
S = A() # making S const makes the error go away
end
using .C
end
using .M
using .M.C
end
@test GlobalBin... | ```suggestion
@test Base.datatype_haspadding(Tuple{ByteString58434}) == (length(Base.padding(Tuple{ByteString58434})) > 0)
``` | using .M2
export x
end
@test M57638_3.x === 1
module GlobalBindingMulti
module M
export S
module C
export S
struct A end
S = A() # making S const makes the error go away
end
using .C
end
using .M
using .M.C
end
@test GlobalBin... | @@ -8534,3 +8534,9 @@ module GlobalBindingMulti
using .M.C
end
@test GlobalBindingMulti.S === GlobalBindingMulti.M.C.S
+
+#58434 bitsegal comparison of oddly sized fields
+primitive type ByteString58434 (18 * 8) end
+
+@test Base.datatype_isbitsegal(Tuple{ByteString58434}) == false
+@test Base.datatype_haspaddin... | test/core.jl | 26 | Julia | 0.857 | suggestion | 128 | 27 | 27 | false | Fix layout flags for types that have oddly sized primitive type fields | 58,435 | JuliaLang/julia | 35,573 | Julia | giordano | gbaraldi | Review this pull request change.
Language: Julia
File: test/core.jl
PR Title: Fix layout flags for types that have oddly sized primitive type fields
BEFORE CODE: using .M2
export x
end
@test M57638_3.x === 1
module GlobalBindingMulti
module M
export S
module C
export S
... | ```suggestion
@test Base.datatype_haspadding(Tuple{ByteString58434}) == (length(Base.padding(Tuple{ByteString58434})) > 0)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
len == 0 && return ""
ref = v.ref
if ref.ptr_or_offset == ref.mem.ptr
str = ccall(:jl_genericmemory_to_string, Ref{String}, (Any, Int), ref.mem, len)
else
str = ccall(:jl_pchar_to_string, Ref{String}, (Ptr{UInt8}, Int), ref, len)
end
# optimized empty!(v); sizehint!(v, 0) calls
... | I believe this is invalid, because `String(::AbstractVector)` is documented to copy, unless the vector is a `Vector`. So you'd need to make a stringvector, then copy to it.
```suggestion
function String(v::ReinterpretArray{UInt8, 1, S, Vector{S}, IsReshaped}) where {S, IsReshaped}
len = length(v)
iszero(l... | len == 0 && return ""
ref = v.ref
if ref.ptr_or_offset == ref.mem.ptr
str = ccall(:jl_genericmemory_to_string, Ref{String}, (Any, Int), ref.mem, len)
else
str = ccall(:jl_pchar_to_string, Ref{String}, (Ptr{UInt8}, Int), ref, len)
end
# optimized empty!(v); sizehint!(v, 0) calls
... | @@ -77,6 +77,20 @@ function String(v::Vector{UInt8})
setfield!(v, :ref, memoryref(Memory{UInt8}()))
return str
end
+function String(v::ReinterpretArray{UInt8, 1, S, Vector{S}, IsReshaped}) where {S, IsReshaped}
+ len = length(v)
+ len == 0 && return ""
+ ref = v.parent.ref
+ if ref.ptr_or_offset ... | base/strings/string.jl | 26 | Julia | 1 | refactor | 655 | 51 | 51 | false | Add a more efficient implementation for String(::ReinterpretArray) | 58,438 | JuliaLang/julia | 35,573 | Julia | jakobnissen | lntricate1 | Review this pull request change.
Language: Julia
File: base/strings/string.jl
PR Title: Add a more efficient implementation for String(::ReinterpretArray)
BEFORE CODE: len == 0 && return ""
ref = v.ref
if ref.ptr_or_offset == ref.mem.ptr
str = ccall(:jl_genericmemory_to_string, Ref{String}, (Any... | I believe this is invalid, because `String(::AbstractVector)` is documented to copy, unless the vector is a `Vector`. So you'd need to make a stringvector, then copy to it.
```suggestion
function String(v::ReinterpretArray{UInt8, 1, S, Vector{S}, IsReshaped}) where {S, IsReshaped}
len = length(v)
iszero(l... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
T_ppjlvalue = PointerType::get(T_pjlvalue, 0);
T_pprjlvalue = PointerType::get(T_prjlvalue, 0);
T_jlfunc = JuliaType::get_jlfunc_ty(context);
assert(T_jlfunc != NULL);
T_jlfuncparams = JuliaType::get_jlfuncparams_ty(context);
assert(T_jlfuncparams != NULL);
T_jl... | These comments are organized (and indentation is formatted) to represent the tree of the TBAA hierarchy, so any updates should preserve that | T_ppjlvalue = PointerType::get(T_pjlvalue, 0);
T_pprjlvalue = PointerType::get(T_prjlvalue, 0);
T_jlfunc = JuliaType::get_jlfunc_ty(context);
assert(T_jlfunc != NULL);
T_jlfuncparams = JuliaType::get_jlfuncparams_ty(context);
assert(T_jlfuncparams != NULL);
T_jl... | @@ -331,11 +331,12 @@ struct jl_tbaacache_t {
// LLVM should have enough info for alias analysis of non-gcframe stack slot
// this is mainly a place holder for `jl_cgval_t::tbaa`
MDNode *tbaa_stack; // stack slot
- MDNode *tbaa_unionselbyte; // a selector byte in isbits Union struct fields
+ ... | src/codegen.cpp | 26 | C++ | 0.429 | style | 140 | 51 | 51 | false | Fix tbaa usage when storing into heap allocated immutable structs | 58,483 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C++
File: src/codegen.cpp
PR Title: Fix tbaa usage when storing into heap allocated immutable structs
BEFORE CODE: T_ppjlvalue = PointerType::get(T_pjlvalue, 0);
T_pprjlvalue = PointerType::get(T_prjlvalue, 0);
T_jlfunc = JuliaType::get_jlfunc_ty(c... | These comments are organized (and indentation is formatted) to represent the tree of the TBAA hierarchy, so any updates should preserve that | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
void initialize(llvm::LLVMContext &context) {
if (initialized) {
assert(&tbaa_root->getContext() == &context);
return;
}
initialized = true;
MDBuilder mbuilder(context);
MDNode *jtbaa = mbuilder.createTBAARoot("jtbaa");
tbaa_root = mbuilder.cre... | This expresses that `tbaa_immut_unionselbyte` does not alias `tbaa_stack` - is that what you intended? | if (initialized) {
assert(&tbaa_root->getContext() == &context);
return;
}
initialized = true;
MDBuilder mbuilder(context);
MDNode *jtbaa = mbuilder.createTBAARoot("jtbaa");
tbaa_root = mbuilder.createTBAAScalarTypeNode("jtbaa", jtbaa);
tba... | @@ -386,8 +386,11 @@ struct jl_tbaacache_t {
MDNode *tbaa_mutab_scalar;
std::tie(tbaa_mutab, tbaa_mutab_scalar) =
tbaa_make_child(mbuilder, "jtbaa_mutab", tbaa_value_scalar);
+ MDNode *tbaa_immut_scalar;
tbaa_datatype = tbaa_make_child(mbuilder, "jtbaa_datatype", tbaa_muta... | src/codegen.cpp | 26 | C++ | 0.5 | question | 102 | 51 | 51 | false | Fix tbaa usage when storing into heap allocated immutable structs | 58,483 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C++
File: src/codegen.cpp
PR Title: Fix tbaa usage when storing into heap allocated immutable structs
BEFORE CODE: void initialize(llvm::LLVMContext &context) {
if (initialized) {
assert(&tbaa_root->getContext() == &context);
return;
... | This expresses that `tbaa_immut_unionselbyte` does not alias `tbaa_stack` - is that what you intended? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
inds2string(inds) = join(map(_indsstring,inds), '×')
_indsstring(i) = string(i)
_indsstring(i::Union{IdentityUnitRange, Slice}) = string(i.indices)
# anything array-like gets summarized e.g. 10-element Array{Int64,1}
summary(io::IO, a::AbstractArray) = array_summary(io, a, axes(a))
function array_summary(io::IO, a, i... | [nitpick] This sentence is quite long and dense; consider breaking it into two or more shorter sentences for clarity and easier maintenance.
```suggestion
Show the quasi-type of `x`. A quasi-type is either the type of `x` or an expression (possibly
containing quasi-types) that would generate an object of the same type ... | showarg(io, parent(r), false)
print(io, ')')
end
function showarg(io::IO, r::ReshapedReinterpretArray{T}, toplevel) where {T}
print(io, "reinterpret(reshape, ", T, ", ")
showarg(io, parent(r), false)
print(io, ')')
toplevel && print(io, " with eltype ", eltype(r))
return nothing
end
# prin... | @@ -3222,17 +3222,23 @@ end
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
- showarg(io::IO, x, toplevel)
+ Base.showarg(io::IO, x, toplevel)
-Show `x` as if it were an argument to a function. This function is
-used by [`summary`](@ref) to display type information in terms of sequences o... | base/show.jl | 26 | Julia | 1 | suggestion | 373 | 51 | 51 | false | Update showarg docstring | 58,509 | JuliaLang/julia | 35,573 | Julia | Copilot | LilithHafner | Review this pull request change.
Language: Julia
File: base/show.jl
PR Title: Update showarg docstring
BEFORE CODE:
inds2string(inds) = join(map(_indsstring,inds), '×')
_indsstring(i) = string(i)
_indsstring(i::Union{IdentityUnitRange, Slice}) = string(i.indices)
# anything array-like gets summarized e.g. 10-elem... | [nitpick] This sentence is quite long and dense; consider breaking it into two or more shorter sentences for clarity and easier maintenance.
```suggestion
Show the quasi-type of `x`. A quasi-type is either the type of `x` or an expression (possibly
containing quasi-types) that would generate an object of the same type ... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io, " with indices ", inds2string(inds))
end
## `summary` for Function
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
Base.showarg(io::IO, x, toplevel)
Show the quasi-type... | Might be worth mentioning that this is also useful if the API for the constructor is public but the specifics of the underlying type are not intended to be public facing. |
# printing iterators from Base.Iterators
function show(io::IO, e::Iterators.Enumerate)
print(io, "enumerate(")
show(io, e.itr)
print(io, ')')
end
show(io::IO, z::Iterators.Zip) = show_delim_array(io, z.is, "zip(", ',', ')', false)
# pretty printing for Iterators.Pairs
function Base.showarg(io::IO, r::Ite... | @@ -3222,17 +3222,23 @@ end
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
- showarg(io::IO, x, toplevel)
+ Base.showarg(io::IO, x, toplevel)
-Show `x` as if it were an argument to a function. This function is
-used by [`summary`](@ref) to display type information in terms of sequences o... | base/show.jl | 26 | Julia | 0.571 | suggestion | 170 | 51 | 51 | false | Update showarg docstring | 58,509 | JuliaLang/julia | 35,573 | Julia | Tokazama | LilithHafner | Review this pull request change.
Language: Julia
File: base/show.jl
PR Title: Update showarg docstring
BEFORE CODE: function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io, " with indices ", inds2string(inds))
end
## `summary` for Function
summar... | Might be worth mentioning that this is also useful if the API for the constructor is public but the specifics of the underlying type are not intended to be public facing. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
julia> x = MyStruct(1, (2,3));
julia> dump(x)
MyStruct
x: Int64 1
y: Tuple{Int64, Int64}
1: Int64 2
2: Int64 3
julia> dump(x; maxdepth = 1)
MyStruct
x: Int64 1
y: Tuple{Int64, Int64}
```
"""
function dump(arg; maxdepth=DUMP_DEFAULT_MAXDEPTH)
# this is typically used interactively, so default to b... | Should this be `Base.showarg(io::IO, x, toplevel::Bool)`? | # sizes such as 0-dimensional, 4-dimensional, 2x3
dims2string(d) = isempty(d) ? "0-dimensional" :
length(d) == 1 ? "$(d[1])-element" :
join(map(string,d), '×')
inds2string(inds) = join(map(_indsstring,inds), '×')
_indsstring(i) = string(i)
_indsstring(i::Union{IdentityUnitRange, Slice... | @@ -3222,17 +3222,23 @@ end
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
- showarg(io::IO, x, toplevel)
+ Base.showarg(io::IO, x, toplevel) | base/show.jl | 26 | Julia | 0.429 | question | 57 | 51 | 51 | false | Update showarg docstring | 58,509 | JuliaLang/julia | 35,573 | Julia | Tokazama | LilithHafner | Review this pull request change.
Language: Julia
File: base/show.jl
PR Title: Update showarg docstring
BEFORE CODE:
julia> x = MyStruct(1, (2,3));
julia> dump(x)
MyStruct
x: Int64 1
y: Tuple{Int64, Int64}
1: Int64 2
2: Int64 3
julia> dump(x; maxdepth = 1)
MyStruct
x: Int64 1
y: Tuple{Int64, Int64... | Should this be `Base.showarg(io::IO, x, toplevel::Bool)`? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
_indsstring(i) = string(i)
_indsstring(i::Union{IdentityUnitRange, Slice}) = string(i.indices)
# anything array-like gets summarized e.g. 10-element Array{Int64,1}
summary(io::IO, a::AbstractArray) = array_summary(io, a, axes(a))
function array_summary(io::IO, a, inds::Tuple{Vararg{OneTo}})
print(io, dims2string(l... | ```suggestion
This function is used by `summary` to display type information in terms of sequences of
``` | _indsstring(i) = string(i)
_indsstring(i::Union{IdentityUnitRange, Slice}) = string(i.indices)
# anything array-like gets summarized e.g. 10-element Array{Int64,1}
summary(io::IO, a::AbstractArray) = array_summary(io, a, axes(a))
function array_summary(io::IO, a, inds::Tuple{Vararg{OneTo}})
print(io, dims2string(l... | @@ -3114,17 +3114,23 @@ end
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
- showarg(io::IO, x, toplevel)
+ Base.showarg(io::IO, x, toplevel)
-Show `x` as if it were an argument to a function. This function is
-used by [`summary`](@ref) to display type information in terms of sequences o... | base/show.jl | 26 | Julia | 0.929 | suggestion | 107 | 51 | 51 | false | Update showarg docstring | 58,509 | JuliaLang/julia | 35,573 | Julia | timholy | LilithHafner | Review this pull request change.
Language: Julia
File: base/show.jl
PR Title: Update showarg docstring
BEFORE CODE: _indsstring(i) = string(i)
_indsstring(i::Union{IdentityUnitRange, Slice}) = string(i.indices)
# anything array-like gets summarized e.g. 10-element Array{Int64,1}
summary(io::IO, a::AbstractArray) =... | ```suggestion
This function is used by `summary` to display type information in terms of sequences of
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
end
function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io, " with indices ", inds2string(inds))
end
## `summary` for Function
summary(io::IO, f::Function) = show(io, MIME"text/... | ```suggestion
The fallback definition is to print `x` as "::\\\$(typeof(x))", representing argument `x` in terms
``` | print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
end
function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io, " with indices ", inds2string(inds))
end
## `summary` for Function
summary(io::IO, f::Function) = show(io, MIME"text/... | @@ -3114,17 +3114,23 @@ end
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
- showarg(io::IO, x, toplevel)
+ Base.showarg(io::IO, x, toplevel)
-Show `x` as if it were an argument to a function. This function is
-used by [`summary`](@ref) to display type information in terms of sequences o... | base/show.jl | 26 | Julia | 0.857 | suggestion | 118 | 51 | 51 | false | Update showarg docstring | 58,509 | JuliaLang/julia | 35,573 | Julia | timholy | LilithHafner | Review this pull request change.
Language: Julia
File: base/show.jl
PR Title: Update showarg docstring
BEFORE CODE: print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
end
function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io... | ```suggestion
The fallback definition is to print `x` as "::\\\$(typeof(x))", representing argument `x` in terms
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
showarg(io, a, true)
end
function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io, " with indices ", inds2string(inds))
end
## `summary` for Function
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
Base.showarg(io::IO, x, t... | ```suggestion
of its type. (The double-colon is omitted if `toplevel=true`.) However, you can specialize
```
This is partially redundant with the previous description of `toplevel`. I'm surely biased and thus unreliable (I wrote this initially, and docs always make sense to their author no matter how confusing they ... | showarg(io, a, true)
end
function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io, " with indices ", inds2string(inds))
end
## `summary` for Function
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
Base.showarg(io::IO, x, t... | @@ -3114,17 +3114,23 @@ end
summary(io::IO, f::Function) = show(io, MIME"text/plain"(), f)
"""
- showarg(io::IO, x, toplevel)
+ Base.showarg(io::IO, x, toplevel)
-Show `x` as if it were an argument to a function. This function is
-used by [`summary`](@ref) to display type information in terms of sequences o... | base/show.jl | 26 | Julia | 1 | security | 585 | 51 | 51 | false | Update showarg docstring | 58,509 | JuliaLang/julia | 35,573 | Julia | timholy | LilithHafner | Review this pull request change.
Language: Julia
File: base/show.jl
PR Title: Update showarg docstring
BEFORE CODE: showarg(io, a, true)
end
function array_summary(io::IO, a, inds)
print(io, dims2string(length.(inds)), " ")
showarg(io, a, true)
print(io, " with indices ", inds2string(inds))
end
## ... | ```suggestion
of its type. (The double-colon is omitted if `toplevel=true`.) However, you can specialize
```
This is partially redundant with the previous description of `toplevel`. I'm surely biased and thus unreliable (I wrote this initially, and docs always make sense to their author no matter how confusing they ... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
precompile(f, (Int, Int))
end
""")
ji, ofile = Base.compilecache(Base.PkgId("NoBackEdges"))
@eval using NoBackEdges
invokelatest() do
@test first(methods(NoBackEdges.f)).specializations.cache.max_world === typemax(UInt)
end
end
# Test precompilation of generated fu... | ```suggestion
mi = first(Base.specializations(match.method))
```
to be certain that you're not *generating* that `mi` with this call. | precompile(f, (Int, Int))
end
""")
ji, ofile = Base.compilecache(Base.PkgId("NoBackEdges"))
@eval using NoBackEdges
invokelatest() do
@test first(methods(NoBackEdges.f)).specializations.cache.max_world === typemax(UInt)
end
end
precompile_test_harness("Pre-compile ... | @@ -2145,6 +2145,29 @@ precompile_test_harness("No backedge precompile") do load_path
end
end
+precompile_test_harness("Pre-compile Core methods") do load_path
+ # Core methods should support pre-compilation as external CI's like anything else
+ # https://github.com/JuliaLang/julia/issues/58497
+ write... | test/precompile.jl | 26 | Julia | 0.714 | suggestion | 148 | 51 | 51 | false | Don't filter `Core` methods from newly-inferred list | 58,510 | JuliaLang/julia | 35,573 | Julia | timholy | topolarity | Review this pull request change.
Language: Julia
File: test/precompile.jl
PR Title: Don't filter `Core` methods from newly-inferred list
BEFORE CODE: precompile(f, (Int, Int))
end
""")
ji, ofile = Base.compilecache(Base.PkgId("NoBackEdges"))
@eval using NoBackEdges
invokela... | ```suggestion
mi = first(Base.specializations(match.method))
```
to be certain that you're not *generating* that `mi` with this call. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
}
}
else {
int special = 0;
for (size_t i = 0; i < len; i++) {
uint8_t c = str[i];
if (raw && ((c == '\\' && i == len-1) || c == '"')) {
special = 1;
break;
}
else if (!raw && (c < 32 || c == 0x7f || c == '\\... | You have to output a number of `\\` equal to the number proceeding this in both of these cases | }
}
else {
int special = 0;
for (size_t i = 0; i < len; i++) {
uint8_t c = str[i];
if (raw && ((c == '\\' && i == len-1) || c == '"')) {
special = 1;
break;
}
else if (!raw && (c < 32 || c == 0x7f || c == '\\... | @@ -722,6 +728,17 @@ static size_t jl_static_show_string(JL_STREAM *out, const char *str, size_t len,
jl_uv_puts(out, str, len);
n += len;
}
+ else if (raw) {
+ for (size_t i = 0; i < len; i++) {
+ uint8_t c = str[i];
+ if (c == '"')
+ ... | src/rtutils.c | 26 | C | 0.357 | suggestion | 94 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: }
}
else {
int special = 0;
for (size_t i = 0; i < len; i++) {
uint8_t c = str[i];
if (raw && ((c == '\\' && i == len-1) || c == '"'... | You have to output a number of `\\` equal to the number proceeding this in both of these cases | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
return 0;
} else if (_iter_count > 10000) {
// We are very likely stuck in a cyclic datastructure, so we assume this is
// _not_ a Function.
return 0;
} else if (vt == jl_function_type) {
return 1;
}
vt = vt->super;
_ite... | These seem unlikely correct, since the standard given results are one higher for these:
https://github.com/JuliaLang/julia/pull/57977/files#r2028868985
```suggestion
// Float16 5
// Float32 9
``` | // This function checks if `vt <: Function` without triggering GC.
static int jl_static_is_function_(jl_datatype_t *vt) JL_NOTSAFEPOINT {
if (!jl_function_type) { // Make sure there's a Function type defined.
return 0;
}
int _iter_count = 0; // To prevent infinite loops from corrupt type objects.
... | @@ -777,6 +804,42 @@ static int jl_static_is_function_(jl_datatype_t *vt) JL_NOTSAFEPOINT {
return 0;
}
+static size_t jl_static_show_float(JL_STREAM *out, double v,
+ jl_datatype_t *vt) JL_NOTSAFEPOINT
+{
+ size_t n = 0;
+ // TODO: non-canonical NaNs do not round-trip
+ ... | src/rtutils.c | 26 | C | 0.786 | suggestion | 216 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: return 0;
} else if (_iter_count > 10000) {
// We are very likely stuck in a cyclic datastructure, so we assume this is
// _not_ a Function.
... | These seem unlikely correct, since the standard given results are one higher for these:
https://github.com/JuliaLang/julia/pull/57977/files#r2028868985
```suggestion
// Float16 5
// Float32 9
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
}
static size_t jl_static_show_float(JL_STREAM *out, double v,
jl_datatype_t *vt) JL_NOTSAFEPOINT
{
size_t n = 0;
// TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suffix = vt == jl_float16_type ? "16" :
vt ... | This seems likely prone to mistakes with double-rounding errors and implementation-defined choices of representations. I recommend changing this to `snprintf` to a small stack buffer and then outputting that with '%s' and then checking explicitly if it already contains `.`, rather than trying to guess | return 0;
} else if (vt == jl_function_type) {
return 1;
}
vt = vt->super;
_iter_count += 1;
}
return 0;
}
static size_t jl_static_show_float(JL_STREAM *out, double v,
jl_datatype_t *vt) JL_NOTSAFEPOINT
{
size_t n = ... | @@ -777,6 +804,42 @@ static int jl_static_is_function_(jl_datatype_t *vt) JL_NOTSAFEPOINT {
return 0;
}
+static size_t jl_static_show_float(JL_STREAM *out, double v,
+ jl_datatype_t *vt) JL_NOTSAFEPOINT
+{
+ size_t n = 0;
+ // TODO: non-canonical NaNs do not round-trip
+ ... | src/rtutils.c | 26 | C | 0.714 | suggestion | 302 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: }
static size_t jl_static_show_float(JL_STREAM *out, double v,
jl_datatype_t *vt) JL_NOTSAFEPOINT
{
size_t n = 0;
// TODO: non-canonical NaNs do... | This seems likely prone to mistakes with double-rounding errors and implementation-defined choices of representations. I recommend changing this to `snprintf` to a small stack buffer and then outputting that with '%s' and then checking explicitly if it already contains `.`, rather than trying to guess | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
{
size_t n = 0;
// TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suffix = vt == jl_float16_type ? "16" :
vt == jl_float32_type ? "32" :
"";
// Digits required to print p significand bi... | This can be implemented compactly similar to f64 case, with style '%e' and `e` replaced with `f` in the buffer | vt = vt->super;
_iter_count += 1;
}
return 0;
}
static size_t jl_static_show_float(JL_STREAM *out, double v,
jl_datatype_t *vt) JL_NOTSAFEPOINT
{
size_t n = 0;
// TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suf... | @@ -777,6 +804,42 @@ static int jl_static_is_function_(jl_datatype_t *vt) JL_NOTSAFEPOINT {
return 0;
}
+static size_t jl_static_show_float(JL_STREAM *out, double v,
+ jl_datatype_t *vt) JL_NOTSAFEPOINT
+{
+ size_t n = 0;
+ // TODO: non-canonical NaNs do not round-trip
+ ... | src/rtutils.c | 26 | C | 0.571 | style | 111 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: {
size_t n = 0;
// TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suffix = vt == jl_float16_type ? "16" :
vt ==... | This can be implemented compactly similar to f64 case, with style '%e' and `e` replaced with `f` in the buffer | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
static size_t jl_static_show_float(JL_STREAM *out, double v,
jl_datatype_t *vt) JL_NOTSAFEPOINT
{
size_t n = 0;
// TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suffix = vt == jl_float16_type ? "16" :
vt ==... | ```suggestion
n += jl_printf(out, "%#.17g", v);
``` | } else if (vt == jl_function_type) {
return 1;
}
vt = vt->super;
_iter_count += 1;
}
return 0;
}
static size_t jl_static_show_float(JL_STREAM *out, double v,
jl_datatype_t *vt) JL_NOTSAFEPOINT
{
size_t n = 0;
// TODO: non-ca... | @@ -777,6 +804,42 @@ static int jl_static_is_function_(jl_datatype_t *vt) JL_NOTSAFEPOINT {
return 0;
}
+static size_t jl_static_show_float(JL_STREAM *out, double v,
+ jl_datatype_t *vt) JL_NOTSAFEPOINT
+{
+ size_t n = 0;
+ // TODO: non-canonical NaNs do not round-trip
+ ... | src/rtutils.c | 26 | C | 0.571 | suggestion | 61 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE:
static size_t jl_static_show_float(JL_STREAM *out, double v,
jl_datatype_t *vt) JL_NOTSAFEPOINT
{
size_t n = 0;
// TODO: non-canonical NaNs do n... | ```suggestion
n += jl_printf(out, "%#.17g", v);
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
// TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suffix = vt == jl_float16_type ? "16" :
vt == jl_float32_type ? "32" :
"";
// Requires minimum 1 (sign) + 17 (sig) + 1 (dot) + 5 ("e-123") ... | While clever, would it be better to use the standard flag `#` for this to preserve all trailing zeros, which will end with simply `.` (rather than to imply the trailing digit is `.0` which may not be entirely true given final digit rounding, though I think it would have to round-trip the same since 17 digits is full ac... | // TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suffix = vt == jl_float16_type ? "16" :
vt == jl_float32_type ? "32" :
"";
// Requires minimum 1 (sign) + 17 (sig) + 1 (dot) + 5 ("e-123") ... | @@ -777,6 +812,55 @@ static int jl_static_is_function_(jl_datatype_t *vt) JL_NOTSAFEPOINT {
return 0;
}
+static size_t jl_static_show_float(JL_STREAM *out, double v,
+ jl_datatype_t *vt) JL_NOTSAFEPOINT
+{
+ size_t n = 0;
+ // TODO: non-canonical NaNs do not round-trip
+ ... | src/rtutils.c | 26 | C | 0.857 | suggestion | 402 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: // TODO: non-canonical NaNs do not round-trip
// TOOD: BFloat16
const char *size_suffix = vt == jl_float16_type ? "16" :
vt == jl_float32_type ? "... | While clever, would it be better to use the standard flag `#` for this to preserve all trailing zeros, which will end with simply `.` (rather than to imply the trailing digit is `.0` which may not be entirely true given final digit rounding, though I think it would have to round-trip the same since 17 digits is full ac... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
special = 1;
break;
}
}
if (!special) {
jl_uv_puts(out, str, len);
n += len;
}
else if (raw) {
// REF: Base.escape_raw_string
int escapes = 0;
for (size_t i = 0; i < len; i++) {
... | ```suggestion
n += jl_uv_puts(out, "\\", 1);
escapes = 0;
}
n += jl_uv_puts(out, str + i, 1);
}
for (; escapes > 0; escapes--)
n += jl_uv_puts(out, "\\", 1);
``` | special = 1;
break;
}
}
if (!special) {
jl_uv_puts(out, str, len);
n += len;
}
else if (raw) {
// REF: Base.escape_raw_string
int escapes = 0;
for (size_t i = 0; i < len; i++) {
... | @@ -722,6 +728,25 @@ static size_t jl_static_show_string(JL_STREAM *out, const char *str, size_t len,
jl_uv_puts(out, str, len);
n += len;
}
+ else if (raw) {
+ // REF: Base.escape_raw_string
+ int escapes = 0;
+ for (size_t i = 0; i < len; i++)... | src/rtutils.c | 26 | C | 0.786 | security | 291 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: special = 1;
break;
}
}
if (!special) {
jl_uv_puts(out, str, len);
n += len;
}
else if (... | ```suggestion
n += jl_uv_puts(out, "\\", 1);
escapes = 0;
}
n += jl_uv_puts(out, str + i, 1);
}
for (; escapes > 0; escapes--)
n += jl_uv_puts(out, "\\", 1);
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
n += jl_printf(out, "(");
n += jl_static_show_x(out, lb, depth, ctx);
if (ua)
n += jl_printf(out, ")");
n += jl_printf(out, "<:");
}
n += jl_static_show_symbol(out, var->name);
if (showbounds && (ub != (jl_value_t*)jl_any_type |... | Can you fix-up Pointer types as well?
```julia
julia> ccall(:jl_, Cvoid, (Any,), Val(Ptr{Cvoid}(1)))
Base.Val{0x0000000000000001}()
```
It's a little awkward, because their `repr` doesn't round-trip either:
```julia
julia> Val(Ptr{Cvoid}(1))
Val{Ptr{Nothing} @0x0000000000000001}()
julia> Val{Ptr{Nothing} @0x... | else if (vt == jl_int32_type) {
n += jl_printf(out, "Int32(%" PRId32 ")", *(int32_t*)v);
}
else if (vt == jl_int16_type) {
n += jl_printf(out, "Int16(%" PRId16 ")", *(int16_t*)v);
}
else if (vt == jl_int8_type) {
n += jl_printf(out, "Int8(%" PRId8 ")", *(int8_t*)v);
}
... | @@ -985,11 +1069,14 @@ static size_t jl_static_show_x_(JL_STREAM *out, jl_value_t *v, jl_datatype_t *vt
n += jl_printf(out, "0x%08" PRIx32, *(uint32_t*)v); | src/rtutils.c | 26 | C | 1 | suggestion | 592 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | topolarity | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: n += jl_printf(out, "(");
n += jl_static_show_x(out, lb, depth, ctx);
if (ua)
n += jl_printf(out, ")");
n += jl_printf(o... | Can you fix-up Pointer types as well?
```julia
julia> ccall(:jl_, Cvoid, (Any,), Val(Ptr{Cvoid}(1)))
Base.Val{0x0000000000000001}()
```
It's a little awkward, because their `repr` doesn't round-trip either:
```julia
julia> Val(Ptr{Cvoid}(1))
Val{Ptr{Nothing} @0x0000000000000001}()
julia> Val{Ptr{Nothing} @0x... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
}
else if (vt == jl_unionall_type) {
jl_unionall_t *ua = (jl_unionall_t*)v;
n += jl_static_show_x(out, ua->body, depth, ctx);
n += jl_printf(out, " where ");
n += jl_static_show_x(out, (jl_value_t*)ua->var, depth->prev, ctx);
}
else if (vt == jl_typename_type) {
n... | Looks like negative `Int128` values are still problematic:
```julia
julia> ccall(:jl_, Cvoid, (Any,), typemin(Int128))
Int128(0x80000000000000000000000000000000)
julia> ccall(:jl_, Cvoid, (Any,), Int128(-1))
Int128(0xffffffffffffffffffffffffffffffff)
```
And `UInt128` prints with an explicit conversion it does... | n += jl_printf(out, ")");
}
}
if (dv->parameters && (jl_value_t*)dv != dv->name->wrapper) {
size_t j, tlen = jl_nparams(dv);
if (tlen > 0) {
n += jl_printf(out, "{");
for (j = 0; j < tlen; j++) {
jl_v... | @@ -954,17 +1034,21 @@ static size_t jl_static_show_x_(JL_STREAM *out, jl_value_t *v, jl_datatype_t *vt
int f = *(uint32_t*)jl_data_ptr(v);
n += jl_printf(out, "#<intrinsic #%d %s>", f, jl_intrinsic_name(f));
}
+ else if (vt == jl_long_type) {
+ // Avoid unnecessary Int64(x)/Int32(x)
+ ... | src/rtutils.c | 26 | C | 0.929 | suggestion | 533 | 51 | 51 | false | Make more types jl_static_show readably | 58,512 | JuliaLang/julia | 35,573 | Julia | topolarity | xal-0 | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: Make more types jl_static_show readably
BEFORE CODE: }
else if (vt == jl_unionall_type) {
jl_unionall_t *ua = (jl_unionall_t*)v;
n += jl_static_show_x(out, ua->body, depth, ctx);
n += jl_printf(out, " where ")... | Looks like negative `Int128` values are still problematic:
```julia
julia> ccall(:jl_, Cvoid, (Any,), typemin(Int128))
Int128(0x80000000000000000000000000000000)
julia> ccall(:jl_, Cvoid, (Any,), Int128(-1))
Int128(0xffffffffffffffffffffffffffffffff)
```
And `UInt128` prints with an explicit conversion it does... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
@test_throws ArgumentError resize!(Float64[], -2)
b = rand(32)
a = reshape(b, (2, 2, 2, 2, 2))
@test ndims(a) == 5
@test a[2,1,2,2,1] == b[14]
@test a[2,2,2,2,2] == b[end]
@test_throws DimensionMismatch reshape(1:3, 4)
# issue #23107
a = [1,2,3]
@test typeof(a)(a) !== a
@t... | can this just be `Base.Compiler.is_removable_if_unused`? | @test_throws ArgumentError resize!(Float64[], -2)
b = rand(32)
a = reshape(b, (2, 2, 2, 2, 2))
@test ndims(a) == 5
@test a[2,1,2,2,1] == b[14]
@test a[2,2,2,2,2] == b[end]
@test_throws DimensionMismatch reshape(1:3, 4)
# issue #23107
a = [1,2,3]
@test typeof(a)(a) !== a
@t... | @@ -100,6 +100,17 @@ using Dates
@test Array{eltype(a)}(a) !== a
@test Vector(a) !== a
end
+@testset "effect inference for `reshape` for `Array`" begin
+ for Arr ∈ (Array{<:Any, 0}, Vector, Matrix, Array{<:Any, 3})
+ for Shape ∈ (Tuple{Int}, Tuple{Int, Int})
+ effects = Base.infer_effect... | test/arrayops.jl | 26 | Julia | 0.429 | question | 56 | 51 | 51 | false | improve inferred effects for `reshape` for `Array` | 58,672 | JuliaLang/julia | 35,573 | Julia | oscardssmith | nsajko | Review this pull request change.
Language: Julia
File: test/arrayops.jl
PR Title: improve inferred effects for `reshape` for `Array`
BEFORE CODE: @test_throws ArgumentError resize!(Float64[], -2)
b = rand(32)
a = reshape(b, (2, 2, 2, 2, 2))
@test ndims(a) == 5
@test a[2,1,2,2,1] == b[14]
@t... | can this just be `Base.Compiler.is_removable_if_unused`? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
push!(state.new_cfg_blocks[end].succs, length(state.new_cfg_blocks)+1)
append!(state.new_cfg_blocks, inlinee_cfg.blocks)
else
# Merge the last block that was already there with the first block we're adding
l = length(state.new_cfg_blocks)
bb_rename_range = (l:length(inlinee_c... | ```suggestion
# the meaning of predecessor 0 depends on the block we encounter it:
# - in the first block, it represents the function entry and so needs to be re-mapped
# - elsewhere, it represents external control-flow from a caught exception which is ... | push!(state.new_cfg_blocks[end].succs, length(state.new_cfg_blocks)+1)
append!(state.new_cfg_blocks, inlinee_cfg.blocks)
else
# Merge the last block that was already there with the first block we're adding
l = length(state.new_cfg_blocks)
bb_rename_range = (l:length(inlinee_c... | @@ -166,12 +167,15 @@ function cfg_inline_item!(ir::IRCode, idx::Int, todo::InliningTodo, state::CFGIn
end
new_block_range = (length(state.new_cfg_blocks)-length(inlinee_cfg.blocks)+1):length(state.new_cfg_blocks)
- # Fixup the edges of the newely added blocks
+ # Fixup the edges of the newly added bl... | Compiler/src/ssair/inlining.jl | 26 | Julia | 1 | suggestion | 567 | 51 | 51 | false | Add 0 predecessor to entry basic block and handle it in inlining | 58,683 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: Julia
File: Compiler/src/ssair/inlining.jl
PR Title: Add 0 predecessor to entry basic block and handle it in inlining
BEFORE CODE: push!(state.new_cfg_blocks[end].succs, length(state.new_cfg_blocks)+1)
append!(state.new_cfg_blocks, inlinee_cfg.blocks)
e... | ```suggestion
# the meaning of predecessor 0 depends on the block we encounter it:
# - in the first block, it represents the function entry and so needs to be re-mapped
# - elsewhere, it represents external control-flow from a caught exception which is ... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
buffer[3] = 8; // length
memcpy(&buffer[4], &message, 8); // Copy the pointer into the buffer
_os_signpost_emit_with_name_impl(&__dso_handle, block->event->os_log_event.log, OS_SIGNPOST_EVENT,
block->signpost_id, (block)->event->os_log_event.name, "%s", buffer, sizeof(buffer) );
}
void _jl_os_signp... | Is it not possible to use `os/log.h` and the public `os_signpost_emit_with_type`? | block->signpost_id = os_signpost_id_make_with_pointer((block)->event->os_log_event.log, (void *)(block));
_os_signpost_emit_with_name_impl(&__dso_handle, block->event->os_log_event.log, OS_SIGNPOST_INTERVAL_BEGIN,
block->signpost_id, (block)->event->os_log_event.name, "", buffer, sizeof(buffer));
}
voi... | @@ -255,6 +255,45 @@ static __itt_event _jl_timing_ittapi_event_create(const char *event) {
#endif // USE_ITTAPI
+#ifdef USE_APPLE_OSLOG
+
+void _jl_os_log_signpost_print(jl_timing_block_t* block, const char *message) {
+ if (!os_signpost_enabled(block->event->os_log_event.log)) {
+ return;
+ }
+ u... | src/timing.c | 26 | C | 0.571 | question | 81 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | maleadt | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: buffer[3] = 8; // length
memcpy(&buffer[4], &message, 8); // Copy the pointer into the buffer
_os_signpost_emit_with_name_impl(&__dso_handle, block->event->os_lo... | Is it not possible to use `os/log.h` and the public `os_signpost_emit_with_type`? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
static __itt_event _jl_timing_ittapi_event_create(const char *event) {
JL_LOCK_NOGC(&jl_timing_ittapi_events_lock);
const size_t n = jl_timing_ittapi_events.len;
for (size_t i = 0; i < n; i++) {
cached_ittapi_event_t *other_event = (cached_ittapi_event_t *)jl_timing_ittapi_events.items[i];
i... | This should probably be in `jl_timing_puts` instead? | static __itt_event _jl_timing_ittapi_event_create(const char *event) {
JL_LOCK_NOGC(&jl_timing_ittapi_events_lock);
const size_t n = jl_timing_ittapi_events.len;
for (size_t i = 0; i < n; i++) {
cached_ittapi_event_t *other_event = (cached_ittapi_event_t *)jl_timing_ittapi_events.items[i];
i... | @@ -253,6 +253,60 @@ static __itt_event _jl_timing_ittapi_event_create(const char *event) {
#endif // USE_ITTAPI
+#ifdef USE_APPLE_OSLOG
+
+void _jl_os_log_signpost_print(jl_timing_block_t* block, const char *message) { | src/timing.c | 26 | C | 0.429 | question | 52 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: static __itt_event _jl_timing_ittapi_event_create(const char *event) {
JL_LOCK_NOGC(&jl_timing_ittapi_events_lock);
const size_t n = jl_timing_ittapi_events.len;
... | This should probably be in `jl_timing_puts` instead? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
}
JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
{
#ifdef USE_TRACY
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_static_show((JL_STREAM*)&buf, v);
if (buf.size == buf.maxsize)
memset(&buf.buf[IOS... | Probably better to:
```suggestion
#ifdef USE_APPLE_OSLOG
_jl_os_log_signpost_print(cur_block, buf.buf);
#endif
```
and then re-use the `jl_static_show` and buf handling for both Tracy + OSLog | jl_timing_puts(cur_block, buf->buf);
}
JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
{
#if defined(USE_TRACY) || defined(USE_APPLE_OSLOG)
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_static_show((JL_STR... | @@ -479,6 +539,17 @@ JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
TracyCZoneText(cur_block->tracy_ctx, buf.buf, buf.size);
#endif
+#ifdef USE_APPLE_OSLOG
+ ios_t buf;
+ ios_mem(&buf, IOS_INLSIZE);
+ buf.growable = 0; // Restrict to inline buffer to avoid allocation
+
+ ... | src/timing.c | 26 | C | 0.857 | suggestion | 201 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: }
JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
{
#ifdef USE_TRACY
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // R... | Probably better to:
```suggestion
#ifdef USE_APPLE_OSLOG
_jl_os_log_signpost_print(cur_block, buf.buf);
#endif
```
and then re-use the `jl_static_show` and buf handling for both Tracy + OSLog | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
}
return blk;
}
JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
{
#ifdef USE_TRACY
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_static_show((JL_STREAM*)&buf, v);
if (buf.size == buf.maxsize)
... | this looks potentially out-of-bounds | buf->buf[buf->size] = '\0'; // Ensure null-termination
}
jl_timing_puts(cur_block, buf->buf);
}
JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
{
#if defined(USE_TRACY) || defined(USE_APPLE_OSLOG)
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Rest... | @@ -479,6 +539,17 @@ JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
TracyCZoneText(cur_block->tracy_ctx, buf.buf, buf.size);
#endif
+#ifdef USE_APPLE_OSLOG
+ ios_t buf;
+ ios_mem(&buf, IOS_INLSIZE);
+ buf.growable = 0; // Restrict to inline buffer to avoid allocation
+
+ ... | src/timing.c | 26 | C | 0.071 | suggestion | 36 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: }
return blk;
}
JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
{
#ifdef USE_TRACY
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
... | this looks potentially out-of-bounds | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
jl_static_show((JL_STREAM*)&buf, v);
if (buf.size == buf.maxsize)
memset(&buf.buf[IOS_INLSIZE - 3], '.', 3);
buf.buf[buf.size] = '\0'; // Ensure null-termination
_jl_os_log_signpost_print(cur_block, buf.buf);
#endif
}
JL_DLLEXPORT void jl_timing_show_module(jl_module_t *m, jl_timing_block_t *cu... | Perhaps better to share the logic with Tracy above? (and support `jl_timing_printf` for OSLog) | const char *module_name = jl_symbol_name(m->name);
jl_timing_puts(cur_block, module_name);
} else {
jl_timing_printf(cur_block, "%s.%s", jl_symbol_name(root->name), jl_symbol_name(m->name));
}
#endif
}
JL_DLLEXPORT void jl_timing_show_filename(const char *path, jl_timing_block_t *cur_bl... | @@ -492,6 +563,13 @@ JL_DLLEXPORT void jl_timing_show_module(jl_module_t *m, jl_timing_block_t *cur_b
jl_timing_printf(cur_block, "%s.%s", jl_symbol_name(root->name), jl_symbol_name(m->name));
}
#endif
+#ifdef USE_APPLE_OSLOG
+ jl_module_t *root = jl_module_root(m);
+ if (root == m || root == jl_ma... | src/timing.c | 26 | C | 0.429 | suggestion | 94 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: jl_static_show((JL_STREAM*)&buf, v);
if (buf.size == buf.maxsize)
memset(&buf.buf[IOS_INLSIZE - 3], '.', 3);
buf.buf[buf.size] = '\0'; // Ensure null-ter... | Perhaps better to share the logic with Tracy above? (and support `jl_timing_printf` for OSLog) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
cached_ittapi_event_t *new_event = (cached_ittapi_event_t *)malloc(sizeof(cached_ittapi_event_t));
arraylist_push(&jl_timing_ittapi_events, (void *)new_event);
new_event->name = event;
new_event->event = __itt_event_create(event, strlen(event));
JL_UNLOCK_NOGC(&jl_timing_ittapi_events_lock);
re... | this (and all the other functions) should follow the naming convention of other functions `_jl_timing_os_log_signpost_start` | cached_ittapi_event_t *new_event = (cached_ittapi_event_t *)malloc(sizeof(cached_ittapi_event_t));
arraylist_push(&jl_timing_ittapi_events, (void *)new_event);
new_event->name = event;
new_event->event = __itt_event_create(event, strlen(event));
JL_UNLOCK_NOGC(&jl_timing_ittapi_events_lock);
re... | @@ -253,6 +253,60 @@ static __itt_event _jl_timing_ittapi_event_create(const char *event) {
#endif // USE_ITTAPI
+#ifdef USE_APPLE_OSLOG
+
+void _jl_os_log_signpost_print(jl_timing_block_t* block, const char *message) {
+ // Buffer the metadata to be emitted with the interval end
+ size_t len = strlen(messag... | src/timing.c | 26 | C | 0.571 | style | 124 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: cached_ittapi_event_t *new_event = (cached_ittapi_event_t *)malloc(sizeof(cached_ittapi_event_t));
arraylist_push(&jl_timing_ittapi_events, (void *)new_event);
n... | this (and all the other functions) should follow the naming convention of other functions `_jl_timing_os_log_signpost_start` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
// No matching event found - create a new one
cached_ittapi_event_t *new_event = (cached_ittapi_event_t *)malloc(sizeof(cached_ittapi_event_t));
arraylist_push(&jl_timing_ittapi_events, (void *)new_event);
new_event->name = event;
new_event->event = __itt_event_create(event, strlen(event));
JL_... | This `block` is stack-allocated, so this seems wrong according to the documentation:
> Creates a signpost identifier that’s unique among signposts logging to the specified log, using a pointer value to generate the unique value.
Should this be part of the `event` instead? |
// No matching event found - create a new one
cached_ittapi_event_t *new_event = (cached_ittapi_event_t *)malloc(sizeof(cached_ittapi_event_t));
arraylist_push(&jl_timing_ittapi_events, (void *)new_event);
new_event->name = event;
new_event->event = __itt_event_create(event, strlen(event));
JL_... | @@ -253,6 +253,60 @@ static __itt_event _jl_timing_ittapi_event_create(const char *event) {
#endif // USE_ITTAPI
+#ifdef USE_APPLE_OSLOG
+
+void _jl_os_log_signpost_print(jl_timing_block_t* block, const char *message) {
+ // Buffer the metadata to be emitted with the interval end
+ size_t len = strlen(messag... | src/timing.c | 26 | C | 0.643 | bug | 276 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE:
// No matching event found - create a new one
cached_ittapi_event_t *new_event = (cached_ittapi_event_t *)malloc(sizeof(cached_ittapi_event_t));
arraylist_push(... | This `block` is stack-allocated, so this seems wrong according to the documentation:
> Creates a signpost identifier that’s unique among signposts logging to the specified log, using a pointer value to generate the unique value.
Should this be part of the `event` instead? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
jl_method_t *def = mi->def.method;
jl_timing_show_location(jl_symbol_name(def->file), def->line, def->module, cur_block);
} else {
jl_timing_printf(cur_block, "<top-level thunk> in %s",
jl_symbol_name(mi->def.module->name));
}
}
JL_DLLEXPORT void jl_timing_show_... | this now overwrites part of the ellipsis that may have been added | jl_symbol_name(root->name),
jl_symbol_name(mod->name));
}
#endif
}
JL_DLLEXPORT void jl_timing_show_method_instance(jl_method_instance_t *mi, jl_timing_block_t *cur_block)
{
jl_timing_show_func_sig(mi->specTypes, cur_block);
if (jl_is_method(mi->def.value))... | @@ -551,8 +599,8 @@ JL_DLLEXPORT void jl_timing_show_func_sig(jl_value_t *v, jl_timing_block_t *cur_
jl_static_show_func_sig_((JL_STREAM*)&buf, v, config);
if (buf.size == buf.maxsize)
memset(&buf.buf[IOS_INLSIZE - 3], '.', 3);
-
- TracyCZoneText(cur_block->tracy_ctx, buf.buf, buf.size);
+ buf.... | src/timing.c | 26 | C | 0.214 | suggestion | 65 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: jl_method_t *def = mi->def.method;
jl_timing_show_location(jl_symbol_name(def->file), def->line, def->module, cur_block);
} else {
jl_timing_prin... | this now overwrites part of the ellipsis that may have been added | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_vprintf((JL_STREAM*)&buf, format, args);
if (buf.size == buf.maxsize)
memset(&buf.buf[IOS_INLSIZE - 3], '.', 3);
buf.buf[buf.size] = '\0'; // Ensure null-termination
jl_timing... | ```suggestion
assert(cur_block->oslog_metadata_len < sizeof(cur_block->oslog_metadata));
size_t remaining = sizeof(cur_block->oslog_metadata) - cur_block->oslog_metadata_len - 1;
```
this is correct as-is, but easy to violate so worth the assert |
JL_DLLEXPORT void jl_timing_printf(jl_timing_block_t *cur_block, const char *format, ...)
{
va_list args;
va_start(args, format);
#if defined(USE_APPLE_OSLOG) || defined(USE_TRACY)
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_vpri... | @@ -589,6 +636,17 @@ JL_DLLEXPORT void jl_timing_puts(jl_timing_block_t *cur_block, const char *str)
#ifdef USE_TRACY
TracyCZoneText(cur_block->tracy_ctx, str, strlen(str));
#endif
+#ifdef USE_APPLE_OSLOG
+ // Buffer the metadata to be emitted with the interval end
+ size_t len = strlen(str);
+ size... | src/timing.c | 26 | C | 0.786 | suggestion | 257 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_vprintf((JL_STREAM*)&buf, format, args);
if ... | ```suggestion
assert(cur_block->oslog_metadata_len < sizeof(cur_block->oslog_metadata));
size_t remaining = sizeof(cur_block->oslog_metadata) - cur_block->oslog_metadata_len - 1;
```
this is correct as-is, but easy to violate so worth the assert | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
//TracyCFiberLeave;
#endif
(void)ct;
jl_timing_block_t *blk = ptls->timing_stack;
ptls->timing_stack = NULL;
if (blk != NULL) {
_COUNTS_PAUSE(&blk->counts_ctx, cycleclock());
}
return blk;
}
JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
{
#if defin... | Is this missing a null terminator? |
//TracyCFiberLeave;
#endif
(void)ct;
jl_timing_block_t *blk = ptls->timing_stack;
ptls->timing_stack = NULL;
if (blk != NULL) {
_COUNTS_PAUSE(&blk->counts_ctx, cycleclock());
}
return blk;
}
static void _jl_timing_show_buf(jl_timing_block_t *cur_block, ios_t *buf)
{
if (buf->... | @@ -477,17 +525,17 @@ JL_DLLEXPORT void jl_timing_show(jl_value_t *v, jl_timing_block_t *cur_block)
if (buf.size == buf.maxsize)
memset(&buf.buf[IOS_INLSIZE - 3], '.', 3);
- TracyCZoneText(cur_block->tracy_ctx, buf.buf, buf.size);
+ jl_timing_puts(cur_block, buf.buf); | src/timing.c | 26 | C | 0.214 | question | 34 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE:
//TracyCFiberLeave;
#endif
(void)ct;
jl_timing_block_t *blk = ptls->timing_stack;
ptls->timing_stack = NULL;
if (blk != NULL) {
_COUNTS_PAUSE(... | Is this missing a null terminator? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
jl_timing_puts(cur_block, buf.buf);
#endif
}
JL_DLLEXPORT void jl_timing_show_macro(jl_method_instance_t *macro, jl_value_t* lno, jl_module_t* mod, jl_timing_block_t *cur_block)
{
jl_timing_printf(cur_block, "%s", jl_symbol_name(macro->def.method->name));
assert(jl_typetagis(lno, jl_linenumbernode_type));
... | this also overwrites the ellipsis |
jl_static_show_config_t config = { /* quiet */ 1 };
jl_static_show_func_sig_((JL_STREAM*)&buf, v, config);
_jl_timing_show_buf(cur_block, &buf);
#endif
}
JL_DLLEXPORT void jl_timing_show_macro(jl_method_instance_t *macro, jl_value_t* lno, jl_module_t* mod, jl_timing_block_t *cur_block)
{
jl_timing_pri... | @@ -570,16 +622,15 @@ JL_DLLEXPORT void jl_timing_printf(jl_timing_block_t *cur_block, const char *for
va_list args;
va_start(args, format);
-#ifdef USE_TRACY
+#if defined(USE_APPLE_OSLOG) || defined(USE_TRACY)
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buf... | src/timing.c | 26 | C | 0.071 | suggestion | 33 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: jl_timing_puts(cur_block, buf.buf);
#endif
}
JL_DLLEXPORT void jl_timing_show_macro(jl_method_instance_t *macro, jl_value_t* lno, jl_module_t* mod, jl_timing_block_t *c... | this also overwrites the ellipsis | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_static_show((JL_STREAM*)&buf, v);
_jl_timing_show_buf(cur_block, &buf);
#endif
}
JL_DLLEXPORT void jl_timing_show_module(jl_module_t *m, jl_timing_block_t *cur_block)
{
#if defined(USE_TRAC... | Might be worth moving this to a `JL_TIMING_HAS_PUTS` or `JL_TIMING_HAS_PRINTF` define, but not required | #if defined(USE_TRACY) || defined(USE_APPLE_OSLOG)
ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_static_show((JL_STREAM*)&buf, v);
_jl_timing_show_buf(cur_block, &buf);
#endif
}
JL_DLLEXPORT void jl_timing_show_module(jl_module_t *m, j... | @@ -496,15 +553,15 @@ JL_DLLEXPORT void jl_timing_show_module(jl_module_t *m, jl_timing_block_t *cur_b
JL_DLLEXPORT void jl_timing_show_filename(const char *path, jl_timing_block_t *cur_block)
{
-#ifdef USE_TRACY
+#if defined(USE_TRACY) || defined(USE_APPLE_OSLOG) | src/timing.c | 26 | C | 0.5 | suggestion | 103 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: ios_t buf;
ios_mem(&buf, IOS_INLSIZE);
buf.growable = 0; // Restrict to inline buffer to avoid allocation
jl_static_show((JL_STREAM*)&buf, v);
_jl_timin... | Might be worth moving this to a `JL_TIMING_HAS_PUTS` or `JL_TIMING_HAS_PRINTF` define, but not required | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
//
// Eventually, Tracy will support telling the server which fibers
// are active upon connection, but until then we work around the
// problem by not explicitly leaving the fiber at all.
//
// Later when we enter the new fiber directly, that will cause the
// the active fiber to be left im... | ```suggestion
buf->size -= 1;
buf->buf[buf->size] = '\0'; // Ensure null-termination
```
just to be consistent about whether the null terminator is included in the buffer or not | //
// Eventually, Tracy will support telling the server which fibers
// are active upon connection, but until then we work around the
// problem by not explicitly leaving the fiber at all.
//
// Later when we enter the new fiber directly, that will cause the
// the active fiber to be left im... | @@ -466,28 +514,37 @@ jl_timing_block_t *jl_timing_block_task_exit(jl_task_t *ct, jl_ptls_t ptls)
return blk;
}
+static void _jl_timing_show_buf(jl_timing_block_t *cur_block, ios_t *buf)
+{
+ if (buf->size == buf->maxsize) {
+ memset(&buf->buf[IOS_INLSIZE - 4], '.', 3);
+ buf->buf[buf->size-1] ... | src/timing.c | 26 | C | 0.857 | suggestion | 199 | 51 | 51 | false | Add option to use Apple os_signpost for instrumentation | 58,702 | JuliaLang/julia | 35,573 | Julia | topolarity | gbaraldi | Review this pull request change.
Language: C
File: src/timing.c
PR Title: Add option to use Apple os_signpost for instrumentation
BEFORE CODE: //
// Eventually, Tracy will support telling the server which fibers
// are active upon connection, but until then we work around the
// problem by not expli... | ```suggestion
buf->size -= 1;
buf->buf[buf->size] = '\0'; // Ensure null-termination
```
just to be consistent about whether the null terminator is included in the buffer or not | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
continue
end
for prop_name in names(m)
prop = getproperty(m, prop_name)
if isa(prop, Libdl.LazyLibrary)
lib_path = dlpath(prop)
lazy_lib_deps = strip_soversion.(basename.(dlpath.(prop.dependencies)))
... | I'm not sure Windows has concept of soversion at all, libraries tend to append numbers to the basename to mimic soversions like in other platforms, but that's custom, not very standard. Instead of special-casing `libpcre2` here, wouldn't it be possible to list multiple values of library name when `libpcre2` is used? ... | continue
end
for prop_name in names(m)
prop = getproperty(m, prop_name)
if isa(prop, Libdl.LazyLibrary)
lib_path = dlpath(prop)
lazy_lib_deps = strip_soversion.(basename.(dlpath.(prop.dependencies)))
... | @@ -220,6 +220,19 @@ try
missing_deps = setdiff(real_lib_deps, lazy_lib_deps)
extraneous_deps = setdiff(lazy_lib_deps, real_lib_deps)
+ # The library name is `libpcre2-8`, with a dash in
+ # its name. That works fine on Unix. On Windows, ... | test/stdlib_dependencies.jl | 26 | Julia | 0.643 | suggestion | 457 | 51 | 51 | false | LibGit2: Update to 1.9.1 | 58,731 | JuliaLang/julia | 35,573 | Julia | giordano | eschnett | Review this pull request change.
Language: Julia
File: test/stdlib_dependencies.jl
PR Title: LibGit2: Update to 1.9.1
BEFORE CODE: continue
end
for prop_name in names(m)
prop = getproperty(m, prop_name)
if isa(prop, Libdl.LazyLibrary)
... | I'm not sure Windows has concept of soversion at all, libraries tend to append numbers to the basename to mimic soversions like in other platforms, but that's custom, not very standard. Instead of special-casing `libpcre2` here, wouldn't it be possible to list multiple values of library name when `libpcre2` is used? ... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
)
end
# Check if we are building a PDF
const render_pdf = "pdf" in ARGS
# Generate a suitable markdown file from NEWS.md and put it in src
function generate_markdown(basename)
str = read(joinpath(@__DIR__, "..", "$basename.md"), String)
splitted = split(str, "<!--- generated by $basename-update.jl: -->")
... | This makes the new section section 4 of 64 of the Manual. IMO this is a way too early/prominent of a position for the GC section. | )
end
# Check if we are building a PDF
const render_pdf = "pdf" in ARGS
# Generate a suitable markdown file from NEWS.md and put it in src
function generate_markdown(basename)
str = read(joinpath(@__DIR__, "..", "$basename.md"), String)
splitted = split(str, "<!--- generated by $basename-update.jl: -->")
... | @@ -161,6 +161,7 @@ generate_markdown("NEWS")
Manual = [
"manual/getting-started.md",
"manual/installation.md",
+ "manual/memory-management.md", | doc/make.jl | 26 | Julia | 0.357 | suggestion | 129 | 51 | 51 | false | Docs: add GC user docs | 58,733 | JuliaLang/julia | 35,573 | Julia | nsajko | IanButterworth | Review this pull request change.
Language: Julia
File: doc/make.jl
PR Title: Docs: add GC user docs
BEFORE CODE: )
end
# Check if we are building a PDF
const render_pdf = "pdf" in ARGS
# Generate a suitable markdown file from NEWS.md and put it in src
function generate_markdown(basename)
str = read(joinpa... | This makes the new section section 4 of 64 of the Manual. IMO this is a way too early/prominent of a position for the GC section. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
t = @_gc_preserve_begin a
p = unsafe_convert(Ptr{Cvoid}, a)
T = eltype(a)
memset(p, x isa T ? x : convert(T, x), length(a) % UInt)
@_gc_preserve_end t
return a
end
## Conversions ##
convert(::Type{T}, a::AbstractArray) where {T<:Memory} = a isa T ? a : T(a)::T
promote_rule(a::Type{Memory{T}},... | Should this be
```suggestion
_checkbounds_array(Bool, A, i) ? (@inbounds A[i], i + 1) : nothing
```
like the `iterate(s::CodeUnits, i=1)` method (which would then be basically identical)? | t = @_gc_preserve_begin a
p = unsafe_convert(Ptr{Cvoid}, a)
T = eltype(a)
memset(p, x isa T ? x : convert(T, x), length(a) % UInt)
@_gc_preserve_end t
return a
end
## Conversions ##
convert(::Type{T}, a::AbstractArray) where {T<:Memory} = a isa T ? a : T(a)::T
promote_rule(a::Type{Memory{T}},... | @@ -227,7 +227,7 @@ Memory{T}(x::AbstractArray{S,1}) where {T,S} = copyto_axcheck!(Memory{T}(undef,
function _iterate_array(A::Union{Memory, Array}, i::Int)
@inline
- (i - 1)%UInt < length(A)%UInt ? (A[i], i + 1) : nothing
+ _checkbounds_array(Bool, A, i) ? (A[i], i + 1) : nothing | base/genericmemory.jl | 26 | Julia | 0.786 | question | 195 | 51 | 51 | false | Unify `_checkbounds_array` into `checkbounds` and use it in more places | 58,785 | JuliaLang/julia | 35,573 | Julia | giordano | giordano | Review this pull request change.
Language: Julia
File: base/genericmemory.jl
PR Title: Unify `_checkbounds_array` into `checkbounds` and use it in more places
BEFORE CODE: t = @_gc_preserve_begin a
p = unsafe_convert(Ptr{Cvoid}, a)
T = eltype(a)
memset(p, x isa T ? x : convert(T, x), length(a) % UIn... | Should this be
```suggestion
_checkbounds_array(Bool, A, i) ? (@inbounds A[i], i + 1) : nothing
```
like the `iterate(s::CodeUnits, i=1)` method (which would then be basically identical)? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
(^)(s::Union{AbstractString,AbstractChar}, r::Integer) = repeat(s, r)
# reverse-order iteration for strings and indices thereof
iterate(r::Iterators.Reverse{<:AbstractString}, i=lastindex(r.itr)) = i < firstindex(r.itr) ? nothing : (r.itr[i], prevind(r.itr, i))
iterate(r::Iterators.Reverse{<:EachStringIndex}, i=lastin... | Ah, I guess we would need to extend the signature of `_checkbounds_array` to accept `CodeUnits`, too. | (^)(s::Union{AbstractString,AbstractChar}, r::Integer) = repeat(s, r)
# reverse-order iteration for strings and indices thereof
iterate(r::Iterators.Reverse{<:AbstractString}, i=lastindex(r.itr)) = i < firstindex(r.itr) ? nothing : (r.itr[i], prevind(r.itr, i))
iterate(r::Iterators.Reverse{<:EachStringIndex}, i=lastin... | @@ -797,7 +797,7 @@ size(s::CodeUnits) = (length(s),)
elsize(s::Type{<:CodeUnits{T}}) where {T} = sizeof(T)
@propagate_inbounds getindex(s::CodeUnits, i::Int) = codeunit(s.s, i)
IndexStyle(::Type{<:CodeUnits}) = IndexLinear()
-@inline iterate(s::CodeUnits, i=1) = (i % UInt) - 1 < length(s) ? (@inbounds s[i], i + 1) ... | base/strings/basic.jl | 26 | Julia | 0.5 | suggestion | 101 | 51 | 51 | false | Unify `_checkbounds_array` into `checkbounds` and use it in more places | 58,785 | JuliaLang/julia | 35,573 | Julia | giordano | giordano | Review this pull request change.
Language: Julia
File: base/strings/basic.jl
PR Title: Unify `_checkbounds_array` into `checkbounds` and use it in more places
BEFORE CODE: (^)(s::Union{AbstractString,AbstractChar}, r::Integer) = repeat(s, r)
# reverse-order iteration for strings and indices thereof
iterate(r::Iter... | Ah, I guess we would need to extend the signature of `_checkbounds_array` to accept `CodeUnits`, too. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
(rows, cols) = axes(mat)
(length(rows) == length(cols)) ||
throw(DimensionMismatch("multiplicative identity defined only for square matrices"))
zer = zero(unit)::T
require_one_based_indexing(mat)
I = similar(mat, T)
fill!(I, zer)
for i ∈ rows
I[i, i] = unit
end
I
end
o... | I'm not entirely sure about adding this `@inline`, but it was used in `Array` and `Memory` methods already. | (length(rows) == length(cols)) ||
throw(DimensionMismatch("multiplicative identity defined only for square matrices"))
zer = zero(unit)::T
require_one_based_indexing(mat)
I = similar(mat, T)
fill!(I, zer)
for i ∈ rows
I[i, i] = unit
end
I
end
one(x::AbstractMatrix{T}) wher... | @@ -1237,15 +1238,15 @@ oneunit(x::AbstractMatrix{T}) where {T} = _one(oneunit(T), x)
iterate_starting_state(A) = iterate_starting_state(A, IndexStyle(A))
iterate_starting_state(A, ::IndexLinear) = firstindex(A)
iterate_starting_state(A, ::IndexStyle) = (eachindex(A),)
-iterate(A::AbstractArray, state = iterate_star... | base/abstractarray.jl | 26 | Julia | 0.714 | suggestion | 107 | 51 | 51 | false | Unify `_checkbounds_array` into `checkbounds` and use it in more places | 58,785 | JuliaLang/julia | 35,573 | Julia | giordano | giordano | Review this pull request change.
Language: Julia
File: base/abstractarray.jl
PR Title: Unify `_checkbounds_array` into `checkbounds` and use it in more places
BEFORE CODE: (rows, cols) = axes(mat)
(length(rows) == length(cols)) ||
throw(DimensionMismatch("multiplicative identity defined only for squar... | I'm not entirely sure about adding this `@inline`, but it was used in `Array` and `Memory` methods already. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
I[i, i] = unit
end
I
end
one(x::AbstractMatrix{T}) where {T} = _one(one(T), x)
oneunit(x::AbstractMatrix{T}) where {T} = _one(oneunit(T), x)
## iteration support for arrays by iterating over `eachindex` in the array ##
# Allows fast iteration by default for both IndexLinear and IndexCartesian arrays
... | Does removing the `@inbounds` impact performance here? I had added it originally because iterating over a view of an `Array` with a non-unit stride was slow without it | end
I
end
one(x::AbstractMatrix{T}) where {T} = _one(one(T), x)
oneunit(x::AbstractMatrix{T}) where {T} = _one(oneunit(T), x)
## iteration support for arrays by iterating over `eachindex` in the array ##
# Allows fast iteration by default for both IndexLinear and IndexCartesian arrays
# While the definitions... | @@ -1237,15 +1238,15 @@ oneunit(x::AbstractMatrix{T}) where {T} = _one(oneunit(T), x)
iterate_starting_state(A) = iterate_starting_state(A, IndexStyle(A))
iterate_starting_state(A, ::IndexLinear) = firstindex(A)
iterate_starting_state(A, ::IndexStyle) = (eachindex(A),)
-iterate(A::AbstractArray, state = iterate_star... | base/abstractarray.jl | 26 | Julia | 0.714 | performance | 167 | 51 | 51 | false | Unify `_checkbounds_array` into `checkbounds` and use it in more places | 58,785 | JuliaLang/julia | 35,573 | Julia | jishnub | giordano | Review this pull request change.
Language: Julia
File: base/abstractarray.jl
PR Title: Unify `_checkbounds_array` into `checkbounds` and use it in more places
BEFORE CODE: I[i, i] = unit
end
I
end
one(x::AbstractMatrix{T}) where {T} = _one(one(T), x)
oneunit(x::AbstractMatrix{T}) where {T} = _one(o... | Does removing the `@inbounds` impact performance here? I had added it originally because iterating over a view of an `Array` with a non-unit stride was slow without it | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
checkbounds_indices(::Type{Bool}, inds::Tuple, ::Tuple{}) = (@inline; all(x->length(x)==1, inds))
# check along a single dimension
"""
checkindex(Bool, inds::AbstractUnitRange, index)
Return `true` if the given `index` is within the bounds of
`inds`. Custom types that would like to behave as indices for all
array... | I fear this whole endeavor is a bit of a wild goose chase, driven by my own misunderstanding of `_checkbounds_array`. I think its use of intrinsics is either unnecessary or driven entirely by bootstrap orderings. Because these two definitions generate exactly the same LLVM for `Int`. Way better to use the generic th... |
# check along a single dimension
"""
checkindex(Bool, inds::AbstractUnitRange, index)
Return `true` if the given `index` is within the bounds of
`inds`. Custom types that would like to behave as indices for all
arrays can extend this method in order to provide a specialized bounds
checking implementation.
See al... | @@ -749,7 +750,7 @@ false
checkindex(::Type{Bool}, inds, i) = throw(ArgumentError(LazyString("unable to check bounds for indices of type ", typeof(i))))
checkindex(::Type{Bool}, inds::AbstractUnitRange, i::Real) = (first(inds) <= i) & (i <= last(inds))
checkindex(::Type{Bool}, inds::IdentityUnitRange, i::Real) = che... | base/abstractarray.jl | 26 | Julia | 0.786 | suggestion | 427 | 51 | 51 | false | Unify `_checkbounds_array` into `checkbounds` and use it in more places | 58,785 | JuliaLang/julia | 35,573 | Julia | mbauman | giordano | Review this pull request change.
Language: Julia
File: base/abstractarray.jl
PR Title: Unify `_checkbounds_array` into `checkbounds` and use it in more places
BEFORE CODE: checkbounds_indices(::Type{Bool}, inds::Tuple, ::Tuple{}) = (@inline; all(x->length(x)==1, inds))
# check along a single dimension
"""
chec... | I fear this whole endeavor is a bit of a wild goose chase, driven by my own misunderstanding of `_checkbounds_array`. I think its use of intrinsics is either unnecessary or driven entirely by bootstrap orderings. Because these two definitions generate exactly the same LLVM for `Int`. Way better to use the generic th... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
#ifdef _OS_WINDOWS_
err = GetLastError();
break; // LoadLibrary already tested the rest
#else
// bail out and show the error if file actually exists
if (jl_stat(path, (char*)&stbuf) == 0)
break;
#endif
}
notfound:
if (throw_err) {
#ifdef _OS_WINDOWS_
char rea... | It could be nice from a readability perspective to not have the argument represent a negation, as the double negative in `!no_deps` is (subjectively) less intuitive. That would also make it consistent with the `throw_err` argument, which isn't e.g. `dont_throw_err`. Anyway, just my drive-by $0.02, now back into the sha... | ios_putc(0, &path);
handle = jl_dlopen(path.buf, flags);
if (handle && !(flags & JL_RTLD_NOLOAD))
jl_timing_puts(JL_TIMING_DEFAULT_BLOCK, jl_pathname_for_handle(handle));
if (handle)
goto success;
#ifdef _OS_WINDOWS_
err = GetLastError();
break; //... | @@ -403,7 +404,7 @@ JL_DLLEXPORT void *jl_load_dynamic_library(const char *modname, unsigned flags,
return handle;
}
-JL_DLLEXPORT int jl_dlsym(void *handle, const char *symbol, void ** value, int throw_err) JL_NOTSAFEPOINT
+JL_DLLEXPORT int jl_dlsym(void *handle, const char *symbol, void ** value, int throw_er... | src/dlload.c | 26 | C | 0.643 | suggestion | 332 | 51 | 51 | false | jl_dlfind: do not find symbols in library dependencies | 58,815 | JuliaLang/julia | 35,573 | Julia | ararslan | xal-0 | Review this pull request change.
Language: C
File: src/dlload.c
PR Title: jl_dlfind: do not find symbols in library dependencies
BEFORE CODE: #ifdef _OS_WINDOWS_
err = GetLastError();
break; // LoadLibrary already tested the rest
#else
// bail out and show the error if file actually exists
... | It could be nice from a readability perspective to not have the argument represent a negation, as the double negative in `!no_deps` is (subjectively) less intuitive. That would also make it consistent with the `throw_err` argument, which isn't e.g. `dont_throw_err`. Anyway, just my drive-by $0.02, now back into the sha... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
return -1;
}
return dlclose(handle);
#endif
}
void *jl_find_dynamic_library_by_addr(void *symbol, int throw_err, int close) JL_NOTSAFEPOINT
{
void *handle;
#ifdef _OS_WINDOWS_
if (!GetModuleHandleExW(GET_MODULE_HANDLE_EX_FLAG_FROM_ADDRESS | GET_MODULE_HANDLE_EX_FLAG_UNCHANGED_REFCOUNT,
... | ```suggestion
dlerror();
handle = dlopen(info.dli_fname, RTLD_NOW | RTLD_NOLOAD | RTLD_LOCAL);
``` | return -1;
}
return dlclose(handle);
#endif
}
void *jl_find_dynamic_library_by_addr(void *symbol, int throw_err, int close) JL_NOTSAFEPOINT
{
void *handle;
#ifdef _OS_WINDOWS_
if (!GetModuleHandleExW(GET_MODULE_HANDLE_EX_FLAG_FROM_ADDRESS | GET_MODULE_HANDLE_EX_FLAG_UNCHANGED_REFCOUNT,
... | @@ -256,13 +257,19 @@ void *jl_find_dynamic_library_by_addr(void *symbol, int throw_err) {
return NULL;
}
handle = dlopen(info.dli_fname, RTLD_NOW | RTLD_NOLOAD | RTLD_LOCAL); | src/dlload.c | 26 | C | 0.857 | suggestion | 109 | 51 | 51 | false | jl_dlfind: do not find symbols in library dependencies | 58,815 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/dlload.c
PR Title: jl_dlfind: do not find symbols in library dependencies
BEFORE CODE: return -1;
}
return dlclose(handle);
#endif
}
void *jl_find_dynamic_library_by_addr(void *symbol, int throw_err, int close) JL_NOTSAFEPOINT
{
void *handl... | ```suggestion
dlerror();
handle = dlopen(info.dli_fname, RTLD_NOW | RTLD_NOLOAD | RTLD_LOCAL);
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
// we can't use a normal safepoint because we don't have signal handlers yet.
jl_atomic_fetch_add(&jl_gc_disable_counter, 1);
// pass NULL as a special token to indicate we are running on an unmanaged task
jl_safepoint_wait_gc(NULL);
// this check is coupled with the one in `jl_safepoint_wait_gc`, w... | ```suggestion
void *handle = jl_find_dynamic_library_by_addr(retaddr, 0, 0);
``` | // we can't use a normal safepoint because we don't have signal handlers yet.
jl_atomic_fetch_add(&jl_gc_disable_counter, 1);
// pass NULL as a special token to indicate we are running on an unmanaged task
jl_safepoint_wait_gc(NULL);
// this check is coupled with the one in `jl_safepoint_wait_gc`, w... | @@ -461,7 +461,7 @@ JL_DLLEXPORT jl_gcframe_t **jl_autoinit_and_adopt_thread(void)
{
if (!jl_is_initialized()) {
void *retaddr = __builtin_extract_return_addr(__builtin_return_address(0));
- void *handle = jl_find_dynamic_library_by_addr(retaddr, 0);
+ void *handle = jl_find_dynamic_library... | src/threading.c | 26 | C | 0.714 | suggestion | 90 | 51 | 51 | false | jl_dlfind: do not find symbols in library dependencies | 58,815 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/threading.c
PR Title: jl_dlfind: do not find symbols in library dependencies
BEFORE CODE: // we can't use a normal safepoint because we don't have signal handlers yet.
jl_atomic_fetch_add(&jl_gc_disable_counter, 1);
// pass NULL as a special token t... | ```suggestion
void *handle = jl_find_dynamic_library_by_addr(retaddr, 0, 0);
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
}
#endif
if (handle != NULL && close)
dlclose(handle); // Undo ref count increment from `dlopen`
#endif
return handle;
}
JL_DLLEXPORT void *jl_load_dynamic_library(const char *modname, unsigned flags, int throw_err)
{
ios_t path, relocated;
int i;
#ifdef _OS_WINDOWS_
int err;
#endif
... | ```suggestion
return jl_libjulia_internal_handle;
```
(remove close parameter) | handle = dlopen("", RTLD_NOW | RTLD_NOLOAD | RTLD_LOCAL);
}
#endif
if (handle != NULL && close)
dlclose(handle); // Undo ref count increment from `dlopen`
#endif
return handle;
}
JL_DLLEXPORT void *jl_load_dynamic_library(const char *modname, unsigned flags, int throw_err)
{
ios_t path,... | @@ -285,7 +292,7 @@ JL_DLLEXPORT void *jl_load_dynamic_library(const char *modname, unsigned flags,
// modname == NULL is a sentinel value requesting the handle of libjulia-internal
if (modname == NULL)
- return jl_find_dynamic_library_by_addr(&jl_load_dynamic_library, throw_err);
+ return jl_... | src/dlload.c | 26 | C | 0.786 | suggestion | 89 | 51 | 51 | false | jl_dlfind: do not find symbols in library dependencies | 58,815 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/dlload.c
PR Title: jl_dlfind: do not find symbols in library dependencies
BEFORE CODE: }
#endif
if (handle != NULL && close)
dlclose(handle); // Undo ref count increment from `dlopen`
#endif
return handle;
}
JL_DLLEXPORT void *jl_load_dynam... | ```suggestion
return jl_libjulia_internal_handle;
```
(remove close parameter) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
*value = dlsym(handle, symbol);
#endif
/* Next, check for errors. On Windows, a NULL pointer means the symbol was
* not found. On everything else, we can have NULL symbols, so we check for
* non-NULL returns from dlerror(). Since POSIX doesn't require `dlerror`
* to be implemented safely, FreeBS... | I'd recommend just tagging the pointer with `| 1` before `dlsym` instead of using jl_find_dynamic_library_by_addr, since handle comparison is potentially annoying (you have to mask with &-4 to remove mode bits)
https://github.com/apple-oss-distributions/dyld/blob/9307719dd8dc9b385daa412b03cfceb897b2b398/dyld/DyldAPI... | /* When !search_deps and the handle isn't special, force RTLD_FIRST. */
if (!search_deps && handle != RTLD_NEXT && handle != RTLD_DEFAULT &&
handle != RTLD_SELF && handle != RTLD_MAIN_ONLY)
handle = (void *)((uintptr_t)handle | 1);
*value = dlsym(handle, symbol);
#else
*value = dlsym(han... | @@ -437,14 +452,35 @@ JL_DLLEXPORT int jl_dlsym(void *handle, const char *symbol, void ** value, int t
}
#endif
- if (!symbol_found && throw_err) {
+ int rtld_first_handle = 0;
+#ifdef _OS_DARWIN_
+ rtld_first_handle = (uintptr_t)handle & 1;
+#endif | src/dlload.c | 26 | C | 0.786 | suggestion | 448 | 51 | 51 | false | jl_dlfind: do not find symbols in library dependencies | 58,815 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: C
File: src/dlload.c
PR Title: jl_dlfind: do not find symbols in library dependencies
BEFORE CODE: *value = dlsym(handle, symbol);
#endif
/* Next, check for errors. On Windows, a NULL pointer means the symbol was
* not found. On everything else, we can have N... | I'd recommend just tagging the pointer with `| 1` before `dlsym` instead of using jl_find_dynamic_library_by_addr, since handle comparison is potentially annoying (you have to mask with &-4 to remove mode bits)
https://github.com/apple-oss-distributions/dyld/blob/9307719dd8dc9b385daa412b03cfceb897b2b398/dyld/DyldAPI... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
end
return Integer(a)
end
length(r::OneTo) = Integer(r.stop - zero(r.stop))
length(r::StepRangeLen) = r.len
length(r::LinRange) = r.len
let bigints = Union{Int, UInt, Int64, UInt64, Int128, UInt128},
smallints = (Int === Int64 ?
Union{Int8, UInt8, Int16, UInt16, Int32, UInt32} :
... | ```suggestion
return a + oneunit(a)
```
Oops, this was leftover from me trying to remove the isempty branch, but that doesn't work. | end
return Integer(a)
end
length(r::OneTo) = Integer(r.stop - zero(r.stop))
length(r::StepRangeLen) = r.len
length(r::LinRange) = r.len
let bigints = Union{Int, UInt, Int64, UInt64, Int128, UInt128},
smallints = (Int === Int64 ?
Union{Int8, UInt8, Int16, UInt16, Int32, UInt32} :
... | @@ -805,17 +805,10 @@ let bigints = Union{Int, UInt, Int64, UInt64, Int128, UInt128},
s = step(r)
diff = last(r) - first(r)
isempty(r) && return zero(diff)
- # if |s| > 1, diff might have overflowed, but unsigned(diff)÷s should
- # therefore still be valid (if the result is repr... | base/range.jl | 26 | Julia | 0.786 | suggestion | 144 | 51 | 51 | false | optimize `length(::OrdinalRange)` for large bit-ints | 58,864 | JuliaLang/julia | 35,573 | Julia | mbauman | mbauman | Review this pull request change.
Language: Julia
File: base/range.jl
PR Title: optimize `length(::OrdinalRange)` for large bit-ints
BEFORE CODE: end
return Integer(a)
end
length(r::OneTo) = Integer(r.stop - zero(r.stop))
length(r::StepRangeLen) = r.len
length(r::LinRange) = r.len
let bigints = Union{Int, ... | ```suggestion
return a + oneunit(a)
```
Oops, this was leftover from me trying to remove the isempty branch, but that doesn't work. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
end
return Integer(a)
end
length(r::OneTo) = Integer(r.stop - zero(r.stop))
length(r::StepRangeLen) = r.len
length(r::LinRange) = r.len
let bigints = Union{Int, UInt, Int64, UInt64, Int128, UInt128},
smallints = (Int === Int64 ?
Union{Int8, UInt8, Int16, UInt16, Int32, UInt32} :
... | Just notice this comment, `unsigned` might not be defined for `s`, e.g. `StepRange{Int,BigInt}(1,1,1)`
```julia
a = div(unsigned(flipsign(diff, s)), s) % typeof(diff)
return flipsign(a, s) + oneunit(a)
```
seems better? | end
return Integer(a)
end
length(r::OneTo) = Integer(r.stop - zero(r.stop))
length(r::StepRangeLen) = r.len
length(r::LinRange) = r.len
let bigints = Union{Int, UInt, Int64, UInt64, Int128, UInt128},
smallints = (Int === Int64 ?
Union{Int8, UInt8, Int16, UInt16, Int32, UInt32} :
... | @@ -805,16 +805,9 @@ let bigints = Union{Int, UInt, Int64, UInt64, Int128, UInt128},
s = step(r)
diff = last(r) - first(r)
isempty(r) && return zero(diff)
- # if |s| > 1, diff might have overflowed, but unsigned(diff)÷s should
- # therefore still be valid (if the result is repre... | base/range.jl | 26 | Julia | 1 | question | 240 | 51 | 51 | false | optimize `length(::OrdinalRange)` for large bit-ints | 58,864 | JuliaLang/julia | 35,573 | Julia | N5N3 | mbauman | Review this pull request change.
Language: Julia
File: base/range.jl
PR Title: optimize `length(::OrdinalRange)` for large bit-ints
BEFORE CODE: end
return Integer(a)
end
length(r::OneTo) = Integer(r.stop - zero(r.stop))
length(r::StepRangeLen) = r.len
length(r::LinRange) = r.len
let bigints = Union{Int, ... | Just notice this comment, `unsigned` might not be defined for `s`, e.g. `StepRange{Int,BigInt}(1,1,1)`
```julia
a = div(unsigned(flipsign(diff, s)), s) % typeof(diff)
return flipsign(a, s) + oneunit(a)
```
seems better? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
function refine_partial_type(@nospecialize t)
t′ = unwrap_unionall(t)
if isa(t′, DataType) && t′.name === _NAMEDTUPLE_NAME && length(t′.parameters) == 2 &&
(t′.parameters[1] === () || t′.parameters[2] === Tuple{})
# if the first/second parameter of `NamedTuple` is known to be empty,
# th... | I think things can break if we hit this case, so possibly we should bail out if this happens. Need some advice on this. | function refine_partial_type(@nospecialize t)
t′ = unwrap_unionall(t)
if isa(t′, DataType) && t′.name === _NAMEDTUPLE_NAME && length(t′.parameters) == 2 &&
(t′.parameters[1] === () || t′.parameters[2] === Tuple{})
# if the first/second parameter of `NamedTuple` is known to be empty,
# th... | @@ -3485,7 +3485,39 @@ function refine_partial_type(@nospecialize t)
return t
end
+function abstract_eval_nonlinearized_foreigncall_name(interp::AbstractInterpreter, e, sstate::StatementState, sv::AbsIntState)
+ if isexpr(e, :call)
+ n = length(e.args)
+ argtypes = Vector{Any}(undef, n)
+ ... | Compiler/src/abstractinterpretation.jl | 26 | Julia | 0.429 | suggestion | 119 | 51 | 51 | false | fix #57749, model ccall binding access in inference | 58,872 | JuliaLang/julia | 35,573 | Julia | JeffBezanson | JeffBezanson | Review this pull request change.
Language: Julia
File: Compiler/src/abstractinterpretation.jl
PR Title: fix #57749, model ccall binding access in inference
BEFORE CODE: function refine_partial_type(@nospecialize t)
t′ = unwrap_unionall(t)
if isa(t′, DataType) && t′.name === _NAMEDTUPLE_NAME && length(t′.par... | I think things can break if we hit this case, so possibly we should bail out if this happens. Need some advice on this. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
if isexpr(e, :call)
n = length(e.args)
argtypes = Vector{Any}(undef, n)
callresult = Future{CallMeta}()
i::Int = 1
nextstate::UInt8 = 0x0
local ai, res
function evalargs(interp, sv)
if nextstate === 0x1
@goto state1
else... | IIRC, this call will occasionally corrupt the InferenceState structure for the current statement, since this is not actually the current statement (there might be some support code in return_type_tfunc for working around it though) |
function abstract_eval_nonlinearized_foreigncall_name(interp::AbstractInterpreter, e, sstate::StatementState, sv::AbsIntState)
if isexpr(e, :call)
n = length(e.args)
argtypes = Vector{Any}(undef, n)
callresult = Future{CallMeta}()
i::Int = 1
nextstate::UInt8 = 0x0
lo... | @@ -3485,7 +3485,57 @@ function refine_partial_type(@nospecialize t)
return t
end
+function abstract_eval_nonlinearized_foreigncall_name(interp::AbstractInterpreter, e, sstate::StatementState, sv::AbsIntState)
+ if isexpr(e, :call)
+ n = length(e.args)
+ argtypes = Vector{Any}(undef, n)
+ ... | Compiler/src/abstractinterpretation.jl | 26 | Julia | 0.571 | suggestion | 231 | 51 | 51 | false | fix #57749, model ccall binding access in inference | 58,872 | JuliaLang/julia | 35,573 | Julia | vtjnash | JeffBezanson | Review this pull request change.
Language: Julia
File: Compiler/src/abstractinterpretation.jl
PR Title: fix #57749, model ccall binding access in inference
BEFORE CODE: if isexpr(e, :call)
n = length(e.args)
argtypes = Vector{Any}(undef, n)
callresult = Future{CallMeta}()
i::Int ... | IIRC, this call will occasionally corrupt the InferenceState structure for the current statement, since this is not actually the current statement (there might be some support code in return_type_tfunc for working around it though) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
sz = 2 * sizeof(Int)
else
sz = Core.sizeof(obj)
end
if sz == 0
# 0-field mutable structs are not unique
return gc_alignment(0)
end
return count ? 1 : sz
end
(::SummarySize)(obj::Symbol) = 0
(::SummarySize)(obj::SummarySize) = 0
function (ss::SummarySize)(obj::String... | I don't know where these magic numbers were pulled from, so I'm not sure if this is 1 object or 2 (or 3?)
I'll need to reverse engineer before merging. | # black-list of items that don't have a Core.sizeof
sz = 2 * sizeof(Int)
else
sz = Core.sizeof(obj)
end
if sz == 0
# 0-field mutable structs are not unique
return gc_alignment(0)
end
return count ? 1 : sz
end
(::SummarySize)(obj::Symbol) = 0
(::SummarySize)(o... | @@ -121,14 +128,14 @@ end
function (ss::SummarySize)(obj::String)
key = ccall(:jl_value_ptr, Ptr{Cvoid}, (Any,), obj)
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
- return Core.sizeof(Int) + Core.sizeof(obj)
+ return (ss.count ? 1 : (Core.sizeof(Int) + Core.sizeof(obj)))
end
function (... | base/summarysize.jl | 26 | Julia | 0.357 | suggestion | 154 | 51 | 51 | false | Base: Add `count` argument to `Base.summarysize` | 58,874 | JuliaLang/julia | 35,573 | Julia | topolarity | topolarity | Review this pull request change.
Language: Julia
File: base/summarysize.jl
PR Title: Base: Add `count` argument to `Base.summarysize`
BEFORE CODE: sz = 2 * sizeof(Int)
else
sz = Core.sizeof(obj)
end
if sz == 0
# 0-field mutable structs are not unique
return gc_alignment(0... | I don't know where these magic numbers were pulled from, so I'm not sure if this is 1 object or 2 (or 3?)
I'll need to reverse engineer before merging. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
return (ss.count ? 1 : (Core.sizeof(Int) + Core.sizeof(obj)))
end
function (ss::SummarySize)(obj::DataType)
key = pointer_from_objref(obj)
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
size::Int = (ss.count ? 1 : (7 * Core.sizeof... | @oscardssmith Is the header here separate from the `obj` itself that we count below? I don't quite grok the accounting | key = ccall(:jl_value_ptr, Ptr{Cvoid}, (Any,), obj)
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
return (ss.count ? 1 : (Core.sizeof(Int) + Core.sizeof(obj)))
end
function (ss::SummarySize)(obj::DataType)
key = pointer_from_objref(obj)
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] ... | @@ -139,17 +146,17 @@ end
function (ss::SummarySize)(obj::Core.TypeName)
key = pointer_from_objref(obj)
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
- return Core.sizeof(obj)
+ return (ss.count ? 1 : Core.sizeof(obj))
end
function (ss::SummarySize)(obj::GenericMemory)
haskey(ss.se... | base/summarysize.jl | 26 | Julia | 0.5 | suggestion | 118 | 51 | 51 | false | Base: Add `count` argument to `Base.summarysize` | 58,874 | JuliaLang/julia | 35,573 | Julia | topolarity | topolarity | Review this pull request change.
Language: Julia
File: base/summarysize.jl
PR Title: Base: Add `count` argument to `Base.summarysize`
BEFORE CODE: haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
return (ss.count ? 1 : (Core.sizeof(Int) + Core.sizeof(obj)))
end
function (ss::SummarySize)(obj::Data... | @oscardssmith Is the header here separate from the `obj` itself that we count below? I don't quite grok the accounting | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
size::Int = (ss.count ? 1 : (7 * Core.sizeof(Int) + 6 * Core.sizeof(Int32)))
size += (ss.count ? 1 : (4 * nfields(obj) + ifelse(Sys.WORD_SIZE == 64, 4, 0)))
size += ss(obj.parameters)::Int
if isdefined(obj, :types)
size += ss(obj.type... | ```suggestion
elseif pointer_from_objref(obj) + 16 != datakey
``` | haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
size::Int = ss.count ? 1 : sizeof(DataType)
size += ss(obj.parameters)::Int
if isdefined(obj, :types)
size += ss(obj.types)::Int
end
return size
end
function (ss::SummarySize)(obj::Core.TypeName)
key = pointer_from_objref(obj... | @@ -139,17 +147,21 @@ end
function (ss::SummarySize)(obj::Core.TypeName)
key = pointer_from_objref(obj)
haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
- return Core.sizeof(obj)
+ return (ss.count ? 1 : Core.sizeof(obj))
end
function (ss::SummarySize)(obj::GenericMemory)
haskey(ss.se... | base/summarysize.jl | 26 | Julia | 0.571 | suggestion | 75 | 51 | 51 | false | Base: Add `count` argument to `Base.summarysize` | 58,874 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: Julia
File: base/summarysize.jl
PR Title: Base: Add `count` argument to `Base.summarysize`
BEFORE CODE: haskey(ss.seen, key) ? (return 0) : (ss.seen[key] = true)
size::Int = (ss.count ? 1 : (7 * Core.sizeof(Int) + 6 * Core.sizeof(Int32)))
size += (ss.count ? 1 ... | ```suggestion
elseif pointer_from_objref(obj) + 16 != datakey
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
value = unwrap_unionall(value)
end
if isa(value, DataType) && parentmodule(value) === obj && nameof(value) === binding
# charge a TypeName to its module (but not to the type)
size += ss(value.name)::Int
end
... | ```suggestion
(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj, ss.count) + (ss.count ? 1 : obj.alloc * sizeof(Base.GMP.Limb))
``` | end
if isa(value, DataType) && parentmodule(value) === obj && nameof(value) === binding
# charge a TypeName to its module (but not to the type)
size += ss(value.name)::Int
end
end
end
end
return size
end
... | @@ -204,4 +216,4 @@ function (ss::SummarySize)(obj::Task)
return size
end
-(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj) + obj.alloc*sizeof(Base.GMP.Limb)
+(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj, ss.count) + obj.alloc * (ss.count ? 1 : sizeof(Base.GMP.Limb)) | base/summarysize.jl | 26 | Julia | 0.714 | suggestion | 137 | 27 | 26 | false | Base: Add `count` argument to `Base.summarysize` | 58,874 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: Julia
File: base/summarysize.jl
PR Title: Base: Add `count` argument to `Base.summarysize`
BEFORE CODE: value = unwrap_unionall(value)
end
if isa(value, DataType) && parentmodule(value) === obj && nameof(value) === bindin... | ```suggestion
(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj, ss.count) + (ss.count ? 1 : obj.alloc * sizeof(Base.GMP.Limb))
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
value = unwrap_unionall(value)
end
if isa(value, DataType) && parentmodule(value) === obj && nameof(value) === binding
# charge a TypeName to its module (but not to the type)
size += ss(value.name)::Int
end
... | ```suggestion
(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj, ss.count) + (ss.count ? 1 : obj.alloc * sizeof(GMP.Limb))
``` | end
if isa(value, DataType) && parentmodule(value) === obj && nameof(value) === binding
# charge a TypeName to its module (but not to the type)
size += ss(value.name)::Int
end
end
end
end
return size
end
... | @@ -204,4 +216,4 @@ function (ss::SummarySize)(obj::Task)
return size
end
-(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj) + obj.alloc*sizeof(Base.GMP.Limb)
+(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj, ss.count) + obj.alloc * (ss.count ? 1 : sizeof(GMP.Limb)) | base/summarysize.jl | 26 | Julia | 0.714 | suggestion | 132 | 27 | 26 | false | Base: Add `count` argument to `Base.summarysize` | 58,874 | JuliaLang/julia | 35,573 | Julia | vtjnash | topolarity | Review this pull request change.
Language: Julia
File: base/summarysize.jl
PR Title: Base: Add `count` argument to `Base.summarysize`
BEFORE CODE: value = unwrap_unionall(value)
end
if isa(value, DataType) && parentmodule(value) === obj && nameof(value) === bindin... | ```suggestion
(ss::SummarySize)(obj::BigInt) = _summarysize(ss, obj, ss.count) + (ss.count ? 1 : obj.alloc * sizeof(GMP.Limb))
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
wrapper(t::Tuple) = map(wrapper, t)
fcviews(t::Tuple) = map(fcviews, t)
offset(t::Tuple) = map(offset, t)
test_many_wrappers(testf, A, wrappers) = foreach(w -> testf(w(A)), wrappers)
test_many_wrappers(testf, A) = test_many_wrappers(testf, A, (identity, tslow, wrapper, fcviews, offset))
A = Int64[1, 2, 3, 4]
Ars = In... | just folding this into the same at L163 | tslow(t::Tuple) = map(tslow, t)
wrapper(t::Tuple) = map(wrapper, t)
fcviews(t::Tuple) = map(fcviews, t)
offset_nominal(t::Tuple) = map(offset_nominal, t)
offset_maybe(t::Tuple) = map(offset_maybe, t)
test_many_wrappers(testf, A, wrappers) = foreach(w -> testf(w(A)), wrappers)
test_many_wrappers(testf, A) = test_many_w... | @@ -37,10 +39,6 @@ test_many_wrappers(B, (identity, tslow)) do _B
@test @inferred(size(reinterpret(reshape, Int128, _B))) == (3,)
end
-test_many_wrappers(C) do Cr | test/reinterpretarray.jl | 26 | Julia | 0.071 | suggestion | 39 | 51 | 51 | false | make ReinterpretArray more Offset-safe | 58,898 | JuliaLang/julia | 35,573 | Julia | adienes | adienes | Review this pull request change.
Language: Julia
File: test/reinterpretarray.jl
PR Title: make ReinterpretArray more Offset-safe
BEFORE CODE: wrapper(t::Tuple) = map(wrapper, t)
fcviews(t::Tuple) = map(fcviews, t)
offset(t::Tuple) = map(offset, t)
test_many_wrappers(testf, A, wrappers) = foreach(w -> testf(w(A)), ... | just folding this into the same at L163 | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
A macro to execute an expression and produce a report of any time spent importing packages and their
dependencies. Any compilation time will be reported as a percentage, and how much of which was recompilation, if any.
One line is printed per package or package extension. The duration shown is the time to import that ... | Looks like another set of functions to add functor support to!
P.S. Do we want to enable this for `@which`, `@less`, etc. above? | return Expr(:call, :error, "@$fcn expects only one non-keyword argument")
end
end
return gen_call_with_extracted_types(__module__, fcn, arg, kws; is_source_reflection, supports_binding_reflection, use_signature_tuple)
end
for fname in [:which, :less, :edit, :functionloc]
@eval begin
... | @@ -398,13 +401,13 @@ end
for fname in [:code_warntype, :code_llvm, :code_native,
:infer_return_type, :infer_effects, :infer_exception_type]
@eval macro ($fname)(ex0...)
- gen_call_with_extracted_types_and_kwargs(__module__, $(QuoteNode(fname)), ex0; is_source_reflection = false)
+ ge... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.5 | question | 131 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | topolarity | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: A macro to execute an expression and produce a report of any time spent importing packages and their
dependencies. Any c... | Looks like another set of functions to add functor support to!
P.S. Do we want to enable this for `@which`, `@less`, etc. above? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
if isexpr(ex, :...)
newarg = Expr(:..., newarg)
push!(args, only(ex.args))
elseif isexpr(ex, :kw)
newarg = Expr(:kw, ex.args[1], newarg)
push!(args, ex.args[end])
else
push!(args, ex)
end
return newarg, i+1
end
function extract_farg(@nospecialize arg)
!is... | ```suggestion
To remove this restriction, the reflection macro may set `use_signature_tuple = true` if the reflection function accepts "functor" arguments, specified as a `Tuple{...}` type."))
``` | elseif isexpr(ex, :.)
(start, branches) = (1, ex.args[2].args)
else
(start, branches) = (2, ex.args)
end
for j in start:length(branches)::Int
branch, i = recursive_dotcalls!(branches[j], args, i)
branches[j] = branch
end
ret... | @@ -105,7 +106,8 @@ function construct_callable(@nospecialize(func::Type))
# Don't support type annotations otherwise, we don't want to give wrong answers
# for callables such as `(::Returns{Int})(args...)` where using `Returns{Int}`
# would give us code for the constructor, not for the callable object.
... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 1 | refactor | 223 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | topolarity | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: if isexpr(ex, :...)
newarg = Expr(:..., newarg)
push!(args, only(ex.args))
elseif isexpr(ex, :kw... | ```suggestion
To remove this restriction, the reflection macro may set `use_signature_tuple = true` if the reflection function accepts "functor" arguments, specified as a `Tuple{...}` type."))
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
args = [#=__source__::=#LineNumberNode, #=__module__::=#Module, Core.Typeof.(ex0.args[3:end])...]
return gen_call(fcn, Any[ex0.args[1], Ref.(args)...], where_params, kws; use_signature_tuple)
end
ex = Meta.lower(__module__, ex0)
if !isa(ex, Expr)
return Expr(:call, :error, "expressi... | This should probably document how `use_signature_tuple` affects the generated code / called function, but the edit to the error above will also help folks understand that this places interface requirements on `fcn` | then carry on using this function. For example, for the input expression `a .+ b`, we emit the anonymous function
`(a, b) -> a .+ b` then call `fcn` just as if the user had issued a call to this anonymous function. That should be the
desired behavior for most macros that want to map an expression to the corresponding g... | @@ -190,12 +192,30 @@ function merge_namedtuple_types(nt::Type{<:NamedTuple}, nts::Type{<:NamedTuple}.
end
end
end
- NamedTuple{Tuple(names), Tuple{types...}}
+ return NamedTuple{Tuple(names), Tuple{types...}}
+end
+
+function gen_call(fcn, args, where_params, kws; use_signature_tuple::... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.786 | suggestion | 214 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | topolarity | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: args = [#=__source__::=#LineNumberNode, #=__module__::=#Module, Core.Typeof.(ex0.args[3:end])...]
return... | This should probably document how `use_signature_tuple` affects the generated code / called function, but the edit to the error above will also help folks understand that this places interface requirements on `fcn` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
using Base: insert!, replace_ref_begin_end_!,
infer_return_type, infer_exception_type, infer_effects, code_ircode, isexpr
# defined in Base so it's possible to time all imports, including InteractiveUtils and its deps
# via. `Base.@time_imports` etc.
import Base: @time_imports, @trace_compile, @trace_dispatch
typ... | Why are you vendoring a buggy copy of Meta.reescape now? | using Base: insert!, replace_ref_begin_end!,
infer_return_type, infer_exception_type, infer_effects, code_ircode, isexpr
# defined in Base so it's possible to time all imports, including InteractiveUtils and its deps
# via. `Base.@time_imports` etc.
import Base: @time_imports, @trace_compile, @trace_dispatch
type... | @@ -21,25 +22,33 @@ function rewrap_where(ex::Expr, where_params::Union{Nothing, Vector{Any}})
Expr(:where, ex, esc.(where_params)...)
end
-function get_typeof(@nospecialize ex)
- # Always unescape to get the core expression, then reescape and esc
- original_ex = ex
- ex = Meta.unescape(ex)
+function r... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.357 | bug | 56 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: using Base: insert!, replace_ref_begin_end_!,
infer_return_type, infer_exception_type, infer_effects, code_ircode, i... | Why are you vendoring a buggy copy of Meta.reescape now? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
for ex in exs
if isexpr(ex, :kw)
push!(kwargs, ex)
elseif isexpr(ex, :parameters)
for kw in ex.args
push!(kwargs, kw)
end
else
push!(args, ex)
end
end
args, kwargs
end
function are_kwargs_valid(kwargs::Vector{An... | These are invalid kwargs, unless unescaping them contains a valid kwarg |
function separate_kwargs(exs::Vector{Any})
args = []
kwargs = []
for ex in exs
if isexpr(ex, :kw)
push!(kwargs, ex)
elseif isexpr(ex, :parameters)
for kw in ex.args
push!(kwargs, kw)
end
else
push!(args, ex)
end... | @@ -147,6 +151,8 @@ function are_kwargs_valid(kwargs::Vector{Any})
isexpr(kwarg, :kw, 2) && isa(kwarg.args[1], Symbol) && continue
isexpr(kwarg, :(::), 2) && continue
isa(kwarg, Symbol) && continue
+ isexpr(kwarg, :escape) && continue
+ isexpr(kwarg, :var"hygienic-scope") && con... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.214 | suggestion | 71 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: for ex in exs
if isexpr(ex, :kw)
push!(kwargs, ex)
elseif isexpr(ex, :parameters)
... | These are invalid kwargs, unless unescaping them contains a valid kwarg | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
elseif isexpr(ex0, :.) && is_source_reflection
# If `ex0` has the form A.B (or some chain A.B.C.D) and `fcn` reflects into the source,
# `A` (or `A.B.C`) may be a module, in which case `fcn` is probably more interested in
# the binding rather than the `getproperty` call.
... | This seems like a bad idea to find some random variable by name and then apply random mutation to it (mainly it seems it would result in incorrect code being executed if someone has a statement with multiple `begin` or `end` like `a[end][end]`) | new = replace_ref_begin_end!(__module__, ex)
modified = ex.args .≠ args
if any(modified) && (isexpr(arr, :(::), 1) || isexpr(arr, :(::), 2) || isexpr(arr, :..., 1))
return Expr(:call, :error, "`begin` or `end` cannot be used with a type-annotated left-hand side argument for an indexing syntax")
... | @@ -206,15 +211,167 @@ function merge_namedtuple_types(nt::Type{<:NamedTuple}, nts::Type{<:NamedTuple}.
end
end
end
- NamedTuple{Tuple(names), Tuple{types...}}
+ return NamedTuple{Tuple(names), Tuple{types...}}
+end
+
+function gen_call(fcn, args, where_params, kws; use_signature_tuple:... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.571 | bug | 244 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: elseif isexpr(ex0, :.) && is_source_reflection
# If `ex0` has the form A.B (or some chain A.B.C.D) a... | This seems like a bad idea to find some random variable by name and then apply random mutation to it (mainly it seems it would result in incorrect code being executed if someone has a statement with multiple `begin` or `end` like `a[end][end]`) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
end
if ex.head === :thunk || exret.head === :none
exret = Expr(:call, :error, "expression is not a function call, \
or is too complex for @$fcn to analyze; \
break it down to simpler parts if possible. \
... | Style (and performance) nit: we tend to use `...` instead of `;` for explicit concat, and an explicit `Any` type:
```
call_reflection = gen_call(fcn, Any[getproperty, ex0.args...], where_params, kws; use_signature_tuple)
```
This seems to have been detected as a issue elsewhere already, but then i... | dotfuncname = gensym("dotfunction")
call = gen_call(fcn, Any[dotfuncname, args...], where_params, kws; use_signature_tuple)
return quote
let $(esc(:($dotfuncname($(xargs...)) = $ex)))
$call
end
end
elseif isexpr(... | @@ -262,10 +420,10 @@ function gen_call_with_extracted_types(__module__, fcn, ex0, kws = Expr[]; is_so
end
fully_qualified_symbol &= ex1 isa Symbol
if fully_qualified_symbol || isexpr(ex1, :(::), 1)
- call_reflection = :($(fcn)(Base.getproperty, $(typesof_expr(ex0.a... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 1 | bug | 570 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: end
if ex.head === :thunk || exret.head === :none
exret = Expr(:call, :error, "expression is not a funct... | Style (and performance) nit: we tend to use `...` instead of `;` for explicit concat, and an explicit `Any` type:
```
call_reflection = gen_call(fcn, Any[getproperty, ex0.args...], where_params, kws; use_signature_tuple)
```
This seems to have been detected as a issue elsewhere already, but then i... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
end
for fname in [:which, :less, :edit, :functionloc]
@eval begin
macro ($fname)(ex0)
gen_call_with_extracted_types(__module__, $(Expr(:quote, fname)), ex0, Expr[];
is_source_reflection = true,
supports_binding_... | This seems very sketchy, since `Ref(nt)` is a perfectly fine, and somewhat common, thing to normally encounter in an AST already (after this function called macroexpand). There should not be any unique behavior of calling `typeof` based on the value of a object found in the code. | binding_reflection = :(error("expression is not a function call"))
end
return quote
local arg1 = $(esc(ex0.args[1]))
if isa(arg1, Module)
$binding_reflection
else
... | @@ -292,41 +450,24 @@ function gen_call_with_extracted_types(__module__, fcn, ex0, kws = Expr[]; is_so
$(esc(ex0)) # trigger syntax errors if any
end
nt = generate_merged_namedtuple_type(kwargs)
- tt = rewrap_where(:(Tuple{$nt, $(get_typeof.(args)...)}), where_param... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.786 | suggestion | 280 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: end
for fname in [:which, :less, :edit, :functionloc]
@eval begin
macro ($fname)(ex0)
gen_call_... | This seems very sketchy, since `Ref(nt)` is a perfectly fine, and somewhat common, thing to normally encounter in an AST already (after this function called macroexpand). There should not be any unique behavior of calling `typeof` based on the value of a object found in the code. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
rows = Any[ (isa(x,Expr) && x.head === :row ? x.args : Any[x]) for x in args ]
lens = map(length, rows)
args = Any[Expr(:tuple, lens...); vcat(rows...)]
ex0.head === :typed_vcat && pushfirst!(args, ex0.args[1])
return Expr(:call, fcn, hf, t... | Please reapply this change to add PairSymAny, so that this part of the code can be type-stable | to a source location, or `@which` to get the method matching the input). In this case, we don't have a clear
semantic source to give (shall it be `broadcasted`, or `materialize`, or something else?), so we return a throwing
expression.
However, if provenance is not of interest, we define an intermediate function on th... | @@ -342,75 +483,46 @@ function gen_call_with_extracted_types(__module__, fcn, ex0, kws = Expr[]; is_so
lens = map(length, rows)
args = Any[Expr(:tuple, lens...); vcat(rows...)]
ex0.head === :typed_vcat && pushfirst!(args, ex0.args[1])
- return Expr(:call... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.429 | suggestion | 94 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: rows = Any[ (isa(x,Expr) && x.head === :row ? x.args : Any[x]) for x in args ]
lens = ma... | Please reapply this change to add PairSymAny, so that this part of the code can be type-stable | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
"""
@less
Evaluates the arguments to the function or macro call, determines their types, and calls the [`less`](@ref)
function on the resulting expression.
See also: [`@edit`](@ref), [`@which`](@ref), [`@code_lowered`](@ref).
"""
:@less
"""
@edit
Evaluates the arguments to the function or macro call, determ... | same comment about `Ref.()` being problematic here (because it shouldn't affect behavior and because is is the wrong way to map to an Any array) | lens = map(length, rows)
args = Any[Expr(:tuple, lens...); vcat(rows...)]
ex0.head === :typed_vcat && pushfirst!(args, ex0.args[1])
return gen_call(fcn, Any[hf, args...], where_params, kws; use_signature_tuple)
else
return gen_c... | @@ -342,75 +483,46 @@ function gen_call_with_extracted_types(__module__, fcn, ex0, kws = Expr[]; is_so
lens = map(length, rows)
args = Any[Expr(:tuple, lens...); vcat(rows...)]
ex0.head === :typed_vcat && pushfirst!(args, ex0.args[1])
- return Expr(:call... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.571 | bug | 144 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: """
@less
Evaluates the arguments to the function or macro call, determines their types, and calls the [`less`](@re... | same comment about `Ref.()` being problematic here (because it shouldn't affect behavior and because is is the wrong way to map to an Any array) | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
PairSymAny(:(.), Base.getproperty),
PairSymAny(:vect, Base.vect),
PairSymAny(Symbol("'"), Base.adjoint),
PairSymAny(:typed_hcat, Base.typed_hcat),
PairSymAny(:string, str... | Did this get deleted by accident? I don't see why you are still calling `Meta.lower` if you unconditionally ignore the result now (and I'm also not entirely certain what this part of the code did before either). | to the type constructor.
To remove the ambiguity, `code_typed` must support an implementation that directly accepts a function type. This implementation
is assumed to be the method for `fcn(sigt::Type{<:Tuple})`.
"""
function gen_call_with_extracted_types(__module__, fcn, ex0, kws = Expr[]; is_source_reflection = !is_c... | @@ -342,75 +483,46 @@ function gen_call_with_extracted_types(__module__, fcn, ex0, kws = Expr[]; is_so
lens = map(length, rows)
args = Any[Expr(:tuple, lens...); vcat(rows...)]
ex0.head === :typed_vcat && pushfirst!(args, ex0.args[1])
- return Expr(:call... | stdlib/InteractiveUtils/src/macros.jl | 26 | Julia | 0.571 | suggestion | 211 | 51 | 51 | false | InteractiveUtils: Support callable objects as functions in introspection macros | 58,905 | JuliaLang/julia | 35,573 | Julia | vtjnash | serenity4 | Review this pull request change.
Language: Julia
File: stdlib/InteractiveUtils/src/macros.jl
PR Title: InteractiveUtils: Support callable objects as functions in introspection macros
BEFORE CODE: PairSymAny(:(.), Base.getproperty),
PairSymAny(:vect, Base.v... | Did this get deleted by accident? I don't see why you are still calling `Meta.lower` if you unconditionally ignore the result now (and I'm also not entirely certain what this part of the code did before either). | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
$(Expr(:thisfunction))
end
inner_func
end
let cs = CallableStruct2()
@test cs()() === cs()
@test cs()() !== cs
end
# Keywords
let
@eval f2(; n=1) = n <= 1 ? n : n * $(Expr(:thisfunction))(; n=n-1)
result = f2(n=5)
@test result == ... | ```suggestion
@test occursin("Cols", result)
``` |
@testset "Closures and nested functions" begin
# Prevents boxed closures
function make_closure()
fib(n) = n <= 1 ? 1 : (@__FUNCTION__)(n - 1) + (@__FUNCTION__)(n - 2)
return fib
end
Test.@inferred make_closure()
closure = make_closure()
@test ... | @@ -4351,3 +4351,59 @@ let f = NoSpecClosure.K(1)
@test f(2) == 1
@test typeof(f).parameters == Core.svec()
end
+
+@testset "Expr(:thisfunction)" begin
+ # regular functions can use Expr(:thisfunction) to refer to the function itself
+ @eval regular_func() = $(Expr(:thisfunction))
+ @test regular_fu... | test/syntax.jl | 26 | Julia | 0.571 | suggestion | 58 | 29 | 51 | false | Create `@__FUNCTION__` and `Expr(:thisfunction)` as generic function self-reference | 58,940 | JuliaLang/julia | 35,573 | Julia | MilesCranmer | MilesCranmer | Review this pull request change.
Language: Julia
File: test/syntax.jl
PR Title: Create `@__FUNCTION__` and `Expr(:thisfunction)` as generic function self-reference
BEFORE CODE: $(Expr(:thisfunction))
end
inner_func
end
let cs = CallableStruct2()
@test cs()() === cs()
... | ```suggestion
@test occursin("Cols", result)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
for f in Base.project_names
if isfile(joinpath(path, f))
return nothing
end
end
# Move up a directory
path = dirname(path)
end
# We never found anything, just return `nothing`
return nothing
end
# We do this to avoid doing the `joinp... | Should we re-use `Base._require_world_age`? | for f in Base.project_names
if isfile(joinpath(path, f))
return nothing
end
end
# Move up a directory
path = dirname(path)
end
# We never found anything, just return `nothing`
return nothing
end
# We do this to avoid doing the `joinp... | @@ -542,7 +542,21 @@ function jointail(dir, tail)
end
end
+const _artifact_str_world_age = Ref{UInt}(typemax(UInt)) | stdlib/Artifacts/src/Artifacts.jl | 26 | Julia | 0.286 | question | 43 | 51 | 51 | false | shield `_artifact_str` function behind a world age barrier | 58,957 | JuliaLang/julia | 35,573 | Julia | topolarity | KristofferC | Review this pull request change.
Language: Julia
File: stdlib/Artifacts/src/Artifacts.jl
PR Title: shield `_artifact_str` function behind a world age barrier
BEFORE CODE: for f in Base.project_names
if isfile(joinpath(path, f))
return nothing
end
end
... | Should we re-use `Base._require_world_age`? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# Move up a directory
path = dirname(path)
end
# We never found anything, just return `nothing`
return nothing
end
# We do this to avoid doing the `joinpath()` work if we don't have to, and also to
# avoid a trailing slash due to `joinpath()`'s habit of including one when the last
# argum... | ```suggestion
return Base.invoke_in_world(world, __artifact_str, __module__, artifacts_toml, name, path_tail, artifact_dict, hash, platform, Val(LazyArtifacts))::String
``` |
# Move up a directory
path = dirname(path)
end
# We never found anything, just return `nothing`
return nothing
end
# We do this to avoid doing the `joinpath()` work if we don't have to, and also to
# avoid a trailing slash due to `joinpath()`'s habit of including one when the last
# argum... | @@ -543,6 +543,14 @@ function jointail(dir, tail)
end
function _artifact_str(__module__, artifacts_toml, name, path_tail, artifact_dict, hash, platform, ::Val{LazyArtifacts}) where LazyArtifacts
+ world = Base._require_world_age[]
+ if world == typemax(UInt)
+ world = Base.get_world_counter()
+ end
... | stdlib/Artifacts/src/Artifacts.jl | 26 | Julia | 0.714 | suggestion | 178 | 51 | 51 | false | shield `_artifact_str` function behind a world age barrier | 58,957 | JuliaLang/julia | 35,573 | Julia | topolarity | KristofferC | Review this pull request change.
Language: Julia
File: stdlib/Artifacts/src/Artifacts.jl
PR Title: shield `_artifact_str` function behind a world age barrier
BEFORE CODE:
# Move up a directory
path = dirname(path)
end
# We never found anything, just return `nothing`
return nothing
end
... | ```suggestion
return Base.invoke_in_world(world, __artifact_str, __module__, artifacts_toml, name, path_tail, artifact_dict, hash, platform, Val(LazyArtifacts))::String
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
end
function __artifact_str(__module__, artifacts_toml, name, path_tail, artifact_dict, hash, platform, ::Val{LazyArtifacts}) where LazyArtifacts
pkg = Base.PkgId(__module__)
if pkg.uuid !== nothing
# Process overrides for this UUID, if we know what it is
process_overrides(artifact_dict, pkg.uu... | ```suggestion
path_base = (@invokelatest LazyArtifacts.ensure_artifact_installed(string(name), meta, artifacts_toml; platform))::String
``` | end
function __artifact_str(__module__, artifacts_toml, name, path_tail, artifact_dict, hash, platform, ::Val{LazyArtifacts}) where LazyArtifacts
pkg = Base.PkgId(__module__)
if pkg.uuid !== nothing
# Process overrides for this UUID, if we know what it is
process_overrides(artifact_dict, pkg.uu... | @@ -565,7 +573,8 @@ function _artifact_str(__module__, artifacts_toml, name, path_tail, artifact_dic
if nameof(LazyArtifacts) in (:Pkg, :Artifacts, :PkgArtifacts)
Base.depwarn("using Pkg instead of using LazyArtifacts is deprecated", :var"@artifact_str", force=true)
end
- ... | stdlib/Artifacts/src/Artifacts.jl | 26 | Julia | 0.714 | suggestion | 153 | 51 | 51 | false | shield `_artifact_str` function behind a world age barrier | 58,957 | JuliaLang/julia | 35,573 | Julia | topolarity | KristofferC | Review this pull request change.
Language: Julia
File: stdlib/Artifacts/src/Artifacts.jl
PR Title: shield `_artifact_str` function behind a world age barrier
BEFORE CODE: end
function __artifact_str(__module__, artifacts_toml, name, path_tail, artifact_dict, hash, platform, ::Val{LazyArtifacts}) where LazyArtifact... | ```suggestion
path_base = (@invokelatest LazyArtifacts.ensure_artifact_installed(string(name), meta, artifacts_toml; platform))::String
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
UVcpus = Ref{Ptr{UV_cpu_info_t}}()
count = Ref{Int32}()
err = ccall(:uv_cpu_info, Int32, (Ptr{Ptr{UV_cpu_info_t}}, Ptr{Int32}), UVcpus, count)
Base.uv_error("uv_cpu_info", err)
cpus = Vector{CPUinfo}(undef, count[])
for i = 1:length(cpus)
cpus[i] = CPUinfo(unsafe_load(UVcpus[], i))
e... | Can we just call this `sysimage_target`? | UVcpus = Ref{Ptr{UV_cpu_info_t}}()
count = Ref{Int32}()
err = ccall(:uv_cpu_info, Int32, (Ptr{Ptr{UV_cpu_info_t}}, Ptr{Int32}), UVcpus, count)
Base.uv_error("uv_cpu_info", err)
cpus = Vector{CPUinfo}(undef, count[])
for i = 1:length(cpus)
cpus[i] = CPUinfo(unsafe_load(UVcpus[], i))
e... | @@ -312,6 +313,23 @@ function cpu_info()
return cpus
end
+"""
+ Sys.sysimage_cpu_target_str()
+
+Return the CPU target string that was used to build the current system image.
+
+This function returns the original CPU target specification that was passed to Julia
+when the system image was compiled. This can ... | base/sysinfo.jl | 26 | Julia | 0.214 | question | 40 | 51 | 51 | false | Add 'sysimage' keyword for `JULIA_CPU_TARGET` to match (or extend) the sysimage target | 58,970 | JuliaLang/julia | 35,573 | Julia | vchuravy | IanButterworth | Review this pull request change.
Language: Julia
File: base/sysinfo.jl
PR Title: Add 'sysimage' keyword for `JULIA_CPU_TARGET` to match (or extend) the sysimage target
BEFORE CODE: UVcpus = Ref{Ptr{UV_cpu_info_t}}()
count = Ref{Int32}()
err = ccall(:uv_cpu_info, Int32, (Ptr{Ptr{UV_cpu_info_t}}, Ptr{Int3... | Can we just call this `sysimage_target`? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
jl_value_t *p = *ptr;
if (p != NULL && (uintptr_t)p >= 4096U) {
jl_value_t *p_ty = jl_typeof(p);
if ((uintptr_t)p_ty >= 4096U) {
if (!jl_isbits(p_ty)) {
nlsep = 1;
brea... | Wait a second, boxed elements should never be locked, right? | jl_value_t *p = *ptr;
if (p != NULL && (uintptr_t)p >= 4096U) {
jl_value_t *p_ty = jl_typeof(p);
if ((uintptr_t)p_ty >= 4096U) {
if (!jl_isbits(p_ty)) {
nlsep = 1;
brea... | @@ -1290,10 +1290,20 @@ static size_t jl_static_show_x_(JL_STREAM *out, jl_value_t *v, jl_datatype_t *vt
for (size_t j = 0; j < tlen; j++) {
if (layout->flags.arrayelem_isboxed) {
jl_value_t **ptr = ((jl_value_t**)m->ptr) + j;
+ if (layout->flags.arr... | src/rtutils.c | 26 | C | 0.286 | question | 60 | 51 | 51 | false | add array element mutex offset in print and gc | 58,997 | JuliaLang/julia | 35,573 | Julia | bbrehm | vtjnash | Review this pull request change.
Language: C
File: src/rtutils.c
PR Title: add array element mutex offset in print and gc
BEFORE CODE: jl_value_t *p = *ptr;
if (p != NULL && (uintptr_t)p >= 4096U) {
jl_value_t *p_ty = jl_typeof(p);
if ((uintptr... | Wait a second, boxed elements should never be locked, right? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
A = mmap(file, Vector{Float64}, 20)
@test Mmap.madvise!(A, Mmap.MADV_WILLNEED) === nothing # checking for no error
finalize(A); A = nothing; GC.gc()
write(file, BitArray(rand(Bool, 20)))
b = mmap(file, BitArray, 20)
@test Mmap.madvise!(b, Mmap.MADV_WILLNEED) === nothing
finalize(b); b = not... | Should this be put into a `@testset` to make failures here more localized? | A = mmap(file, Vector{Float64}, 20)
@test Mmap.madvise!(A, Mmap.MADV_WILLNEED) === nothing # checking for no error
finalize(A); A = nothing; GC.gc()
write(file, BitArray(rand(Bool, 20)))
b = mmap(file, BitArray, 20)
@test Mmap.madvise!(b, Mmap.MADV_WILLNEED) === nothing
finalize(b); b = not... | @@ -343,6 +343,18 @@ end
GC.gc()
rm(file)
+# test for #58982 | stdlib/Mmap/test/runtests.jl | 26 | Julia | 0.429 | bug | 74 | 41 | 42 | false | Bugfix: Use Base.aligned_sizeof instead of sizeof in Mmap.mmap | 58,998 | JuliaLang/julia | 35,573 | Julia | Seelengrab | JonasIsensee | Review this pull request change.
Language: Julia
File: stdlib/Mmap/test/runtests.jl
PR Title: Bugfix: Use Base.aligned_sizeof instead of sizeof in Mmap.mmap
BEFORE CODE: A = mmap(file, Vector{Float64}, 20)
@test Mmap.madvise!(A, Mmap.MADV_WILLNEED) === nothing # checking for no error
finalize(A); A = no... | Should this be put into a `@testset` to make failures here more localized? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
JL_DLLEXPORT jl_value_t *jl_fptr_args(jl_value_t *f, jl_value_t **args, uint32_t nargs, jl_code_instance_t *m)
{
jl_fptr_args_t invoke = jl_atomic_load_relaxed(&m->specptr.fptr1);
assert(invoke && "Forgot to set specptr for jl_fptr_args!");
return invoke(f, args, nargs);
}
JL_DLLEXPORT jl_value_t *jl_fptr... | I'm not sure what this is about, but without this, i was seeing several lines like this:
```
precompile(Tuple)
precompile(Tuple)
precompile(Tuple)
``` | jl_atomic_fetch_add(&jl_force_trace_dispatch_enabled, -1);
}
static void record_dispatch_statement(jl_method_instance_t *mi)
{
static ios_t f_dispatch;
static JL_STREAM* s_dispatch = NULL;
jl_method_t *def = mi->def.method;
if (!jl_is_method(def))
return;
uint8_t force_trace_dispatch =... | @@ -2795,7 +2795,8 @@ static void record_dispatch_statement(jl_method_instance_t *mi)
s_dispatch = (JL_STREAM*) &f_dispatch;
}
}
- if (!jl_has_free_typevars(mi->specTypes)) {
+ // NOTE: Sometimes the specType is just `Tuple`, which is not useful to print.
+ if (!jl_has_free_typevars(... | src/gf.c | 26 | C | 0.714 | suggestion | 155 | 51 | 51 | false | Improve --trace-dispatch coverage: emit in "full-cache" fast path as well. | 59,012 | JuliaLang/julia | 35,573 | Julia | NHDaly | NHDaly | Review this pull request change.
Language: C
File: src/gf.c
PR Title: Improve --trace-dispatch coverage: emit in "full-cache" fast path as well.
BEFORE CODE:
JL_DLLEXPORT jl_value_t *jl_fptr_args(jl_value_t *f, jl_value_t **args, uint32_t nargs, jl_code_instance_t *m)
{
jl_fptr_args_t invoke = jl_atomic_load_r... | I'm not sure what this is about, but without this, i was seeing several lines like this:
```
precompile(Tuple)
precompile(Tuple)
precompile(Tuple)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
// Increment the flag to allow reentrant callers to `@trace_dispatch`.
jl_atomic_fetch_add(&jl_force_trace_dispatch_enabled, -1);
}
static void record_dispatch_statement(jl_method_instance_t *mi)
{
static ios_t f_dispatch;
static JL_STREAM* s_dispatch = NULL;
jl_method_t *def = mi->def.method;
... | ```suggestion
// NOTE: For builtin functions, the specType is just `Tuple`, which is not useful to print.
``` | // Increment the flag to allow reentrant callers to `@trace_dispatch`.
jl_atomic_fetch_add(&jl_force_trace_dispatch_enabled, -1);
}
static void record_dispatch_statement(jl_method_instance_t *mi)
{
static ios_t f_dispatch;
static JL_STREAM* s_dispatch = NULL;
jl_method_t *def = mi->def.method;
... | @@ -3096,7 +3096,8 @@ static void record_dispatch_statement(jl_method_instance_t *mi)
s_dispatch = (JL_STREAM*) &f_dispatch;
}
}
- if (!jl_has_free_typevars(mi->specTypes)) {
+ // NOTE: Sometimes the specType is just `Tuple`, which is not useful to print. | src/gf.c | 26 | C | 0.929 | suggestion | 115 | 51 | 51 | false | Improve --trace-dispatch coverage: emit in "full-cache" fast path as well. | 59,012 | JuliaLang/julia | 35,573 | Julia | NHDaly | NHDaly | Review this pull request change.
Language: C
File: src/gf.c
PR Title: Improve --trace-dispatch coverage: emit in "full-cache" fast path as well.
BEFORE CODE: // Increment the flag to allow reentrant callers to `@trace_dispatch`.
jl_atomic_fetch_add(&jl_force_trace_dispatch_enabled, -1);
}
static void recor... | ```suggestion
// NOTE: For builtin functions, the specType is just `Tuple`, which is not useful to print.
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
function handle_invoke_expr!(todo::Vector{Pair{Int,Any}}, ir::IRCode,
idx::Int, stmt::Expr, @nospecialize(info::CallInfo), flag::UInt32, sig::Signature, state::InliningState)
edge = stmt.args[1]
mi = isa(edge, MethodInstance) ? edge : get_ci_mi(edge::CodeInstance)
case = resolve_todo(mi, info, flag, sta... | looks like `process_simple!` can call `inline_apply!` which contains `ir.stmts[idx][:info] = new_info`, so maybe the variable assignment `info = inst[:info]` needs to still happen on L1634 ? | function handle_invoke_expr!(todo::Vector{Pair{Int,Any}}, ir::IRCode,
idx::Int, stmt::Expr, @nospecialize(info::CallInfo), flag::UInt32, sig::Signature, state::InliningState)
edge = stmt.args[1]
mi = isa(edge, MethodInstance) ? edge : get_ci_mi(edge::CodeInstance)
case = resolve_todo(mi, info, flag, sta... | @@ -1625,14 +1623,14 @@ function assemble_inline_todo!(ir::IRCode, state::InliningState)
todo = Pair{Int, Any}[]
for idx in 1:length(ir.stmts)
- flag = ir.stmts[idx][:flag]
+ inst = ir.stmts[idx]
+ flag = inst[:flag]
+ info = inst[:info] | Compiler/src/ssair/inlining.jl | 26 | Julia | 0.5 | question | 190 | 51 | 51 | false | Don't inline generated functions if we can't invoke their generator | 59,018 | JuliaLang/julia | 35,573 | Julia | adienes | serenity4 | Review this pull request change.
Language: Julia
File: Compiler/src/ssair/inlining.jl
PR Title: Don't inline generated functions if we can't invoke their generator
BEFORE CODE: function handle_invoke_expr!(todo::Vector{Pair{Int,Any}}, ir::IRCode,
idx::Int, stmt::Expr, @nospecialize(info::CallInfo), flag::UInt32... | looks like `process_simple!` can call `inline_apply!` which contains `ir.stmts[idx][:info] = new_info`, so maybe the variable assignment `info = inst[:info]` needs to still happen on L1634 ? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
return evalpoly(x*x, (1.0f0, -4.934802f0, 4.058712f0, -1.3352624f0, 0.23531426f0, -0.0255071f0))
end
struct CosPiEvaluationScheme{T <: AbstractFloat, Rest <: Tuple{T, Vararg{T}}}
"""
Coefficient 0 of the polynomial.
"""
c₀::T
"""
Coefficient 2 of the polynomial, in double-word format (uneva... | I don't love these structs/I'm not sure they pull their weight as an abstraction (compared to just a function to evaluate the kernel like this replaces). | return evalpoly(x*x, (1.0f0, -4.934802f0, 4.058712f0, -1.3352624f0, 0.23531426f0, -0.0255071f0))
end
function _cospi_kernel(x::AbstractFloat, sch)
x² = x * x
c₀ = sch.c₀
(c₂, c₂_lo) = sch.c₂
rest = sch.rest
c₂ = -c₂
c₂_lo = -c₂_lo
r = evalpoly(x², rest) * x²
a_x² = c₂ * x²
a_x²_... | @@ -784,6 +772,194 @@ end
return evalpoly(x*x, (1.0f0, -4.934802f0, 4.058712f0, -1.3352624f0, 0.23531426f0, -0.0255071f0))
end
+struct CosPiEvaluationScheme{T <: AbstractFloat, Rest <: Tuple{T, Vararg{T}}}
+ """
+ Coefficient 0 of the polynomial.
+ """
+ c₀::T
+ """
+ Coefficient 2 of the poly... | base/special/trig.jl | 26 | Julia | 0.429 | refactor | 153 | 51 | 51 | false | redo `sinpi` and `cospi` kernel polynomial approximation | 59,031 | JuliaLang/julia | 35,573 | Julia | oscardssmith | nsajko | Review this pull request change.
Language: Julia
File: base/special/trig.jl
PR Title: redo `sinpi` and `cospi` kernel polynomial approximation
BEFORE CODE: return evalpoly(x*x, (1.0f0, -4.934802f0, 4.058712f0, -1.3352624f0, 0.23531426f0, -0.0255071f0))
end
struct CosPiEvaluationScheme{T <: AbstractFloat, Rest ... | I don't love these structs/I'm not sure they pull their weight as an abstraction (compared to just a function to evaluate the kernel like this replaces). | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
elseif opts.malloc_log == 3
push!(addflags, "--track-allocation=@$(unsafe_string(opts.tracked_path))")
end
if opts.color == 1
push!(addflags, "--color=yes")
elseif opts.color == 2
push!(addflags, "--color=no")
end
if opts.startupfile == 2
push!(addflags, "--startu... | I don't believe we propagate any flags related to `--output` (why would we?)
```suggestion
``` | elseif opts.malloc_log == 3
push!(addflags, "--track-allocation=@$(unsafe_string(opts.tracked_path))")
end
if opts.color == 1
push!(addflags, "--color=yes")
elseif opts.color == 2
push!(addflags, "--color=no")
end
if opts.startupfile == 2
push!(addflags, "--startu... | @@ -248,6 +248,15 @@ function julia_cmd(julia=joinpath(Sys.BINDIR, julia_exename()); cpu_target::Unio
if opts.compress_sysimage == 1
push!(addflags, "--compress-sysimage=yes")
end
+ if opts.target_sanitize_memory == 1
+ push!(addflags, "--target-sanitize=memory")
+ end
+ if opts.targe... | base/util.jl | 26 | Julia | 0.571 | suggestion | 94 | 51 | 51 | false | Add --target-sanitize=(memory|thread|address) | 59,035 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: Julia
File: base/util.jl
PR Title: Add --target-sanitize=(memory|thread|address)
BEFORE CODE: elseif opts.malloc_log == 3
push!(addflags, "--track-allocation=@$(unsafe_string(opts.tracked_path))")
end
if opts.color == 1
push!(addflags, "--color=... | I don't believe we propagate any flags related to `--output` (why would we?)
```suggestion
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
empty!(EXT_DORMITORY_FAILED)
filter!(failed) do extid
return !run_extension_callbacks(extid)
end
prepend!(EXT_DORMITORY_FAILED, failed)
end
return
end
"""
get_extension(parent::Module, extension::Symbol)
Return the module for `extension` of `parent` or return `nothing` if the exten... | Revert remaining changes to this file also?
```suggestion
# OOICCDDP - see jl_cache_flags
``` | end
false
finally
global loading_extension = false
end
return succeeded
end
function run_extension_callbacks(pkgid::PkgId)
assert_havelock(require_lock)
# take ownership of extids that depend on this pkgid
extids = pop!(EXT_DORMITORY, pkgid, nothing)
extids === nothi... | @@ -1656,7 +1656,7 @@ end
struct CacheFlags
- # OOICCDDP - see jl_cache_flags
+ # OOICCDDP ?????ATM - see jl_cache_flags | base/loading.jl | 26 | Julia | 0.643 | performance | 97 | 51 | 51 | false | Add --target-sanitize=(memory|thread|address) | 59,035 | JuliaLang/julia | 35,573 | Julia | vtjnash | xal-0 | Review this pull request change.
Language: Julia
File: base/loading.jl
PR Title: Add --target-sanitize=(memory|thread|address)
BEFORE CODE: empty!(EXT_DORMITORY_FAILED)
filter!(failed) do extid
return !run_extension_callbacks(extid)
end
prepend!(EXT_DORMITORY_FAILED, failed)
end
retu... | Revert remaining changes to this file also?
```suggestion
# OOICCDDP - see jl_cache_flags
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
@inbounds setindex!(zd, val % UInt32, nlimbs-1)
@inbounds setindex!(zd, (val >> 32) % UInt32, nlimbs)
end
z
end
function BigFloat(x::BigInt, r::MPFRRoundingMode=rounding_raw(BigFloat); precision::Integer=_precision_with_base_2(BigFloat))
z = BigFloat(;precision=precision)
ccall((:mpfr_s... | why does the default rounding mode change? | @inbounds setindex!(zd, val % UInt32, nlimbs-1)
@inbounds setindex!(zd, (val >> 32) % UInt32, nlimbs)
end
z
end
function BigFloat(x::BigInt, r::MPFRRoundingMode=rounding_raw(BigFloat); precision::Integer=_precision_with_base_2(BigFloat))
z = BigFloat(;precision=precision)
ccall((:mpfr_s... | @@ -391,13 +391,19 @@ BigFloat(x::Union{UInt8,UInt16,UInt32}, r::MPFRRoundingMode=rounding_raw(BigFloa
BigFloat(x::Union{Float16,Float32}, r::MPFRRoundingMode=rounding_raw(BigFloat); precision::Integer=_precision_with_base_2(BigFloat)) =
BigFloat(Float64(x), r; precision=precision)
-function BigFloat(x::Rationa... | base/mpfr.jl | 26 | Julia | 0.143 | question | 42 | 51 | 51 | false | Fix rounding when converting Rational to BigFloat | 59,063 | JuliaLang/julia | 35,573 | Julia | adienes | OlivierHnt | Review this pull request change.
Language: Julia
File: base/mpfr.jl
PR Title: Fix rounding when converting Rational to BigFloat
BEFORE CODE: @inbounds setindex!(zd, val % UInt32, nlimbs-1)
@inbounds setindex!(zd, (val >> 32) % UInt32, nlimbs)
end
z
end
function BigFloat(x::BigInt, r::MPFRRo... | why does the default rounding mode change? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
# This file is a part of Julia. License is MIT: https://julialang.org/license
module ULPError
export ulp_error, ulp_error_maximum
@noinline function throw_invalid()
throw(ArgumentError("invalid"))
end
function ulp_error(accurate::AbstractFloat, approximate::AbstractFloat)
if isnan(accur... | this should return `Inf` if one is NaN and the other isn't, or 0 if both are NaN. | # This file is a part of Julia. License is MIT: https://julialang.org/license
module ULPError
export ulp_error, ulp_error_maximum
function ulp_error(accurate::AbstractFloat, approximate::AbstractFloat)
# the ULP error is usually not required to great accuracy, so `Float32` should be precise enough
... | @@ -0,0 +1,57 @@
+# This file is a part of Julia. License is MIT: https://julialang.org/license
+
+module ULPError
+ export ulp_error, ulp_error_maximum
+ @noinline function throw_invalid()
+ throw(ArgumentError("invalid"))
+ end
+ function ulp_error(accurate::AbstractFloat, approximate::AbstractFloa... | test/testhelpers/ULPError.jl | 14 | Julia | 0.429 | suggestion | 81 | 39 | 39 | false | improve `cosc(::Float32)` and `cosc(::Float64)` accuracy | 59,087 | JuliaLang/julia | 35,573 | Julia | oscardssmith | nsajko | Review this pull request change.
Language: Julia
File: test/testhelpers/ULPError.jl
PR Title: improve `cosc(::Float32)` and `cosc(::Float64)` accuracy
BEFORE CODE: # This file is a part of Julia. License is MIT: https://julialang.org/license
module ULPError
export ulp_error, ulp_error_maximum
@noinline fun... | this should return `Inf` if one is NaN and the other isn't, or 0 if both are NaN. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
module ULPError
export ulp_error, ulp_error_maximum
@noinline function throw_invalid()
throw(ArgumentError("invalid"))
end
function ulp_error(accurate::AbstractFloat, approximate::AbstractFloat)
# the ULP error is usually not required to great accuracy, so `Float32` should be precise en... | maybe put all of this inside `!(isfinite(approx) && isfinite(accurate))`? Also, I don't think these variables are used enough to be worthwhile. |
module ULPError
export ulp_error, ulp_error_maximum
function ulp_error(accurate::AbstractFloat, approximate::AbstractFloat)
# the ULP error is usually not required to great accuracy, so `Float32` should be precise enough
zero_return = 0f0
inf_return = Inf32
# handle floating-poi... | @@ -0,0 +1,59 @@
+# This file is a part of Julia. License is MIT: https://julialang.org/license
+
+module ULPError
+ export ulp_error, ulp_error_maximum
+ @noinline function throw_invalid()
+ throw(ArgumentError("invalid"))
+ end
+ function ulp_error(accurate::AbstractFloat, approximate::AbstractFloa... | test/testhelpers/ULPError.jl | 26 | Julia | 0.571 | suggestion | 143 | 51 | 47 | false | improve `cosc(::Float32)` and `cosc(::Float64)` accuracy | 59,087 | JuliaLang/julia | 35,573 | Julia | oscardssmith | nsajko | Review this pull request change.
Language: Julia
File: test/testhelpers/ULPError.jl
PR Title: improve `cosc(::Float32)` and `cosc(::Float64)` accuracy
BEFORE CODE:
module ULPError
export ulp_error, ulp_error_maximum
@noinline function throw_invalid()
throw(ArgumentError("invalid"))
end
func... | maybe put all of this inside `!(isfinite(approx) && isfinite(accurate))`? Also, I don't think these variables are used enough to be worthwhile. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
accur_is_zero = iszero(accurate)
accur_is_nan = isnan(accurate)
accur_is_inf = isinf(accurate)
approx_is_inf = isinf(approximate)
approx_is_nan = isnan(approximate)
approx_is_zero = iszero(approximate)
if accur_is_nan || approx_is_nan
return if accur_i... | why is this necessary? I would trust the precision of `accurate` It's often nice to be able to use lower precision than BigFloat to avoid the allocation and speed penalty. | if accur_is_nan || approx_is_nan
return if accur_is_nan === approx_is_nan
zero_return
else
inf_return
end
end
if isinf(approximate)
return if isinf(accurate) && (signbit(accurate) ... | @@ -0,0 +1,59 @@
+# This file is a part of Julia. License is MIT: https://julialang.org/license
+
+module ULPError
+ export ulp_error, ulp_error_maximum
+ @noinline function throw_invalid()
+ throw(ArgumentError("invalid"))
+ end
+ function ulp_error(accurate::AbstractFloat, approximate::AbstractFloa... | test/testhelpers/ULPError.jl | 26 | Julia | 0.571 | question | 172 | 48 | 36 | false | improve `cosc(::Float32)` and `cosc(::Float64)` accuracy | 59,087 | JuliaLang/julia | 35,573 | Julia | oscardssmith | nsajko | Review this pull request change.
Language: Julia
File: test/testhelpers/ULPError.jl
PR Title: improve `cosc(::Float32)` and `cosc(::Float64)` accuracy
BEFORE CODE: accur_is_zero = iszero(accurate)
accur_is_nan = isnan(accurate)
accur_is_inf = isinf(accurate)
approx_is_inf = isinf(app... | why is this necessary? I would trust the precision of `accurate` It's often nice to be able to use lower precision than BigFloat to avoid the allocation and speed penalty. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
i == 2 || i == 1 || error("unexpected index")
pi = p[i]
prod, err = two_mul(hi,x)
hi = pi+prod
lo = fma(lo, x, prod - (hi - pi) + err)
return hi, lo
end
# generic, but involves double rounding
function _180_over_pi(z::AbstractFloat)
180 / oftype(z, pi)
end
function _pi_over_180(z::AbstractF... | Should just be `z::Float16` (and no if branch below), or is missing a `Float32`? | i == 2 || i == 1 || error("unexpected index")
pi = p[i]
prod, err = two_mul(hi,x)
hi = pi+prod
lo = fma(lo, x, prod - (hi - pi) + err)
return hi, lo
end
# generic, but involves double rounding
function _180_over_pi(z::AbstractFloat)
180 / oftype(z, pi)
end
function _pi_over_180(z::AbstractF... | @@ -228,6 +228,30 @@ end
return hi, lo
end
+# generic, but involves double rounding
+function _180_over_pi(z::AbstractFloat)
+ 180 / oftype(z, pi)
+end
+function _pi_over_180(z::AbstractFloat)
+ oftype(z, pi) / 180
+end
+
+# rounded to closest representable number where necessary
+function _180_over_pi(z:... | base/math.jl | 26 | Julia | 0.429 | question | 80 | 51 | 51 | false | more accurate `rad2deg` and `deg2rad` for `Float16` and `Float32` | 59,097 | JuliaLang/julia | 35,573 | Julia | thchr | nsajko | Review this pull request change.
Language: Julia
File: base/math.jl
PR Title: more accurate `rad2deg` and `deg2rad` for `Float16` and `Float32`
BEFORE CODE: i == 2 || i == 1 || error("unexpected index")
pi = p[i]
prod, err = two_mul(hi,x)
hi = pi+prod
lo = fma(lo, x, prod - (hi - pi) + err)
... | Should just be `z::Float16` (and no if branch below), or is missing a `Float32`? | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
side::Float64
end
struct Circle <: Shape
radius::Float64
end
area(s::Square) = s.side^2
area(c::Circle) = pi*c.radius^2
sum_areas(v::Vector{Shape}) = sum(area, v)
function @main(args::Vector{String})::Cint
println(Core.stdout, str())
foreach(x->println(Core.stdout, x), args)
# test map/mapreduce;... | This one now works for me locally!
```suggestion
d = mapreduce(x -> x^2, +, sorted_arr)
``` | side::Float64
end
struct Circle <: Shape
radius::Float64
end
area(s::Square) = s.side^2
area(c::Circle) = pi*c.radius^2
sum_areas(v::Vector{Shape}) = sum(area, v)
function @main(args::Vector{String})::Cint
println(Core.stdout, str())
foreach(x->println(Core.stdout, x), args)
# test map/mapreduce;... | @@ -3,8 +3,34 @@ const str = OncePerProcess{String}() do
return "Hello, " * world
end
+abstract type Shape end
+struct Square <: Shape
+ side::Float64
+end
+struct Circle <: Shape
+ radius::Float64
+end
+area(s::Square) = s.side^2
+area(c::Circle) = pi*c.radius^2
+
+sum_areas(v::Vector{Shape}) = sum(area,... | test/trimming/hello.jl | 26 | Julia | 0.714 | suggestion | 100 | 30 | 30 | false | make more cases of mapreduce trimmable | 59,104 | JuliaLang/julia | 35,573 | Julia | mbauman | JeffBezanson | Review this pull request change.
Language: Julia
File: test/trimming/hello.jl
PR Title: make more cases of mapreduce trimmable
BEFORE CODE: side::Float64
end
struct Circle <: Shape
radius::Float64
end
area(s::Square) = s.side^2
area(c::Circle) = pi*c.radius^2
sum_areas(v::Vector{Shape}) = sum(area, v)
fun... | This one now works for me locally!
```suggestion
d = mapreduce(x -> x^2, +, sorted_arr)
``` | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
end
struct Circle <: Shape
radius::Float64
end
area(s::Square) = s.side^2
area(c::Circle) = pi*c.radius^2
sum_areas(v::Vector{Shape}) = sum(area, v)
function @main(args::Vector{String})::Cint
println(Core.stdout, str())
foreach(x->println(Core.stdout, x), args)
# test map/mapreduce; should work but r... | And this one works if we pass an init. I don't understand why it's not working without it, though. The verifier is still routing through all the reduce empty methods:
```julia
Verifier error #1: unresolved call from statement Base.reduce_empty(op::Function, T::Type)::Any
Stacktrace:
[1] mapreduce_empty(::type... | end
struct Circle <: Shape
radius::Float64
end
area(s::Square) = s.side^2
area(c::Circle) = pi*c.radius^2
sum_areas(v::Vector{Shape}) = sum(area, v)
function @main(args::Vector{String})::Cint
println(Core.stdout, str())
foreach(x->println(Core.stdout, x), args)
# test map/mapreduce; should work but r... | @@ -3,8 +3,34 @@ const str = OncePerProcess{String}() do
return "Hello, " * world
end
+abstract type Shape end
+struct Square <: Shape
+ side::Float64
+end
+struct Circle <: Shape
+ radius::Float64
+end
+area(s::Square) = s.side^2
+area(c::Circle) = pi*c.radius^2
+
+sum_areas(v::Vector{Shape}) = sum(area,... | test/trimming/hello.jl | 26 | Julia | 1 | nitpick | 3,959 | 29 | 29 | false | make more cases of mapreduce trimmable | 59,104 | JuliaLang/julia | 35,573 | Julia | mbauman | JeffBezanson | Review this pull request change.
Language: Julia
File: test/trimming/hello.jl
PR Title: make more cases of mapreduce trimmable
BEFORE CODE: end
struct Circle <: Shape
radius::Float64
end
area(s::Square) = s.side^2
area(c::Circle) = pi*c.radius^2
sum_areas(v::Vector{Shape}) = sum(area, v)
function @main(args::... | And this one works if we pass an init. I don't understand why it's not working without it, though. The verifier is still routing through all the reduce empty methods:
```julia
Verifier error #1: unresolved call from statement Base.reduce_empty(op::Function, T::Type)::Any
Stacktrace:
[1] mapreduce_empty(::type... | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
others...)
end
_alloftype(ofdesiredtype, accumulated) = ofdesiredtype, Base.front(accumulated)
let
d = MixedKeyDict((Dict(1 => 3), Dict(4. => 2)))
e = MixedKeyDict((Dict(1 => 7), Dict(5. => 9)))
@test merge(+, d, e).dicts == (Dict(1 => 10), Dict(4.0 => 2, 5.0 => 9))
f = MixedKeyDic... | This looks like it has a decent chance to be inferred as `WrapperWithUnionall33472{Int}`. | others...)
end
_alloftype(ofdesiredtype, accumulated) = ofdesiredtype, Base.front(accumulated)
let
d = MixedKeyDict((Dict(1 => 3), Dict(4. => 2)))
e = MixedKeyDict((Dict(1 => 7), Dict(5. => 9)))
@test merge(+, d, e).dicts == (Dict(1 => 10), Dict(4.0 => 2, 5.0 => 9))
f = MixedKeyDic... | @@ -3549,8 +3549,14 @@ f31974(n::Int) = f31974(1:n)
# call cycles.
@test code_typed(f31974, Tuple{Int}) !== nothing
-f_overly_abstract_complex() = Complex(Ref{Number}(1)[])
-@test Base.return_types(f_overly_abstract_complex, Tuple{}) == [Complex]
+# Issue #33472
+struct WrapperWithUnionall33472{T<:Real}
+ x::T
+... | Compiler/test/inference.jl | 26 | Julia | 0.357 | suggestion | 89 | 51 | 51 | false | inference: Make test indepdent of the `Complex` method table | 59,105 | JuliaLang/julia | 35,573 | Julia | JeffBezanson | Keno | Review this pull request change.
Language: Julia
File: Compiler/test/inference.jl
PR Title: inference: Make test indepdent of the `Complex` method table
BEFORE CODE: others...)
end
_alloftype(ofdesiredtype, accumulated) = ofdesiredtype, Base.front(accumulated)
let
d = MixedKeyDict((Dict(1 ... | This looks like it has a decent chance to be inferred as `WrapperWithUnionall33472{Int}`. | You are an expert software engineer and code reviewer. Review pull requests carefully for correctness, bugs, security issues, performance problems, maintainability, and unnecessary complexity. Only raise actionable issues supported by the code. Explain why an issue matters and suggest a concrete fix when appropriate. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.