commit_hash stringlengths 40 40 | author stringlengths 1 57 | date timestamp[s]date 2010-07-26 04:45:09 2026-04-14 18:21:10 | message stringlengths 8 1.39M | diff stringlengths 68 51.2k | files_changed int64 1 136 | insertions int64 0 2.35k | deletions int64 0 1.9k |
|---|---|---|---|---|---|---|---|
26719d539f833d8d1b2be56d8e86e8f8ab3c7590 | Andrew Gallant | 2025-03-05T23:28:40 | tz/posix: remove `Offset` on `DstInfo`
This became superfluous with a prior refactor and can just be removed.
Previously, we weren't filling in default values eagerly, so this was
set to the default value when the DST offset was absent. But now we fill
in the default values eagerly, so this just is not needed any more... | diff --git a/src/tz/posix.rs b/src/tz/posix.rs
index 2b9dc98..b9c9be6 100644
--- a/src/tz/posix.rs
+++ b/src/tz/posix.rs
@@ -304,7 +304,7 @@ impl ReasonablePosixTimeZone {
let dt = Offset::UTC.to_datetime(timestamp);
self.dst_info_utc(dt.date().year_ranged())
.filter(|dst_info| dst_info.i... | 1 | 19 | 24 |
b336d2ecc4edde29cbd03ba1f76b9a9797919666 | Andrew Gallant | 2025-03-05T23:28:19 | bench: re-arrange and improve `tz` benchmarks
This rounds out some of the different code paths on common
`TimeZone` operations. | diff --git a/bench/Cargo.toml b/bench/Cargo.toml
index 145a07b..6caa0c5 100644
--- a/bench/Cargo.toml
+++ b/bench/Cargo.toml
@@ -21,7 +21,7 @@ path = "src/bench.rs"
[dependencies]
criterion = { version = "0.5.1", default-features = false }
-jiff = { path = ".." }
+jiff = { path = "..", features = ["tzdb-bundle-alwa... | 2 | 176 | 14 |
12ca7aaba77894e3fcdd96e97b61abc4b0266bb5 | daalfox | 2025-02-28T13:05:34 | signed_duration: implement `Sum`
Fixes #274, PR #282 | diff --git a/src/signed_duration.rs b/src/signed_duration.rs
index eb4c426..1273a73 100644
--- a/src/signed_duration.rs
+++ b/src/signed_duration.rs
@@ -2166,6 +2166,18 @@ impl core::ops::Mul<i32> for SignedDuration {
}
}
+impl core::iter::Sum for SignedDuration {
+ fn sum<I: Iterator<Item = Self>>(iter: I) ... | 1 | 36 | 0 |
c966d65a8abd3354ecc48ac815fae0bbd5aaa608 | Andrew Gallant | 2025-02-25T20:27:39 | tests: add proc macro tests!
Woohoo. We can finally use the proc macro. It was very satisfying seeing
these tests run and pass on the first try. :-) | diff --git a/tests/lib.rs b/tests/lib.rs
index 863845f..78ec5fd 100644
--- a/tests/lib.rs
+++ b/tests/lib.rs
@@ -1,4 +1,5 @@
// See: https://github.com/rust-lang/rust/pull/121364
#![allow(unknown_lints, ambiguous_negative_literals)]
+mod procmacro;
mod tc39_262;
diff --git a/tests/procmacro/mod.rs b/tests/procmacr... | 2 | 519 | 0 |
d12ab2cd6d1fdcec3c485d7023cf3535cbf2b993 | Andrew Gallant | 2025-02-24T21:11:18 | jiff-cli: add `jiff-cli generate shared` command
This is responsible for copying `jiff::shared` into the proc
macro. We do some minor transformations, but otherwise the
code remains the same. | diff --git a/crates/jiff-cli/cmd/generate/crc32.rs b/crates/jiff-cli/cmd/generate/crc32.rs
index e914527..891f4d4 100644
--- a/crates/jiff-cli/cmd/generate/crc32.rs
+++ b/crates/jiff-cli/cmd/generate/crc32.rs
@@ -46,10 +46,11 @@ pub fn run(p: &mut Parser) -> anyhow::Result<()> {
args::configure(p, USAGE, &mut [&mu... | 6 | 225 | 1 |
c0be6c625cead998f551aac6cffa820455d81fe7 | Andrew Gallant | 2025-02-24T21:06:25 | bench: fix benchmark name | diff --git a/bench/src/datetime.rs b/bench/src/datetime.rs
index 6698974..c91aa5a 100644
--- a/bench/src/datetime.rs
+++ b/bench/src/datetime.rs
@@ -111,7 +111,7 @@ fn add_years_months_days(c: &mut Criterion) {
/// This is useful when you have a known time zone already and want to get
/// a specific instant for many ... | 1 | 1 | 1 |
89f4946780ca7c5cab048a6c453f6ff3ac418c09 | Andrew Gallant | 2025-02-19T19:22:14 | util: make more utility functions `const`
I believe I'm going to need these in order to maintain
some semblance of encapsulation and convert data handed
to Jiff by the macro into the internal representations
used by Jiff. | diff --git a/src/util/common.rs b/src/util/common.rs
index c64ab61..60693c6 100644
--- a/src/util/common.rs
+++ b/src/util/common.rs
@@ -55,7 +55,7 @@ pub(crate) const fn days_in_month(year: i16, month: i8) -> i8 {
/// Ref: https://github.com/cassioneri/eaf/blob/684d3cc32d14eee371d0abe4f683d6d6a49ed5c1/algorithms/neri... | 1 | 11 | 11 |
e364c9f46c98fa0a772c9ff881282ff39bc12b12 | Andrew Gallant | 2025-02-19T03:53:11 | civil: optimize Date addition with -1 or +1 days
I originally didn't do this because it seemed kinda bush
league, but this turns out to be a pretty common case via
POSIX time zones. So let's just use `yesterday()` and
`tomorrow()`, which will avoid a round-trip through Unix
epoch days. | diff --git a/bench/src/date.rs b/bench/src/date.rs
index 5cb5add..816ecb9 100644
--- a/bench/src/date.rs
+++ b/bench/src/date.rs
@@ -86,6 +86,13 @@ fn add_days(c: &mut Criterion) {
1070,
civil::date(2026, 1, 17),
);
+ benchmark_with(
+ c,
+ "one",
+ civil::date(2023, 2, 12... | 2 | 43 | 18 |
fd3eeecbf58a9507e3c3aa5f3246c83e4c2410dc | Andrew Gallant | 2025-02-19T00:25:22 | rangeint: add const constructor
I believe this will be useful for shunting data between Jiff and proc
macros. | diff --git a/src/util/rangeint.rs b/src/util/rangeint.rs
index 406a114..e43a23d 100644
--- a/src/util/rangeint.rs
+++ b/src/util/rangeint.rs
@@ -138,6 +138,26 @@ macro_rules! define_ranged {
}
}
+ /// Like `new`, but monomorphic and works in a `const` context.
+ #[i... | 1 | 20 | 0 |
1ef1715b355cb26feacb94c2cacc0a3966d583b5 | Andrew Gallant | 2025-02-18T22:28:31 | util: make `ArrayStr::new` be `const`
This is prep for experimenting with a proc macro to create
a static `TimeZone`.
We're going to need to expose a bunch of internal data types
in order to permit a proc macro to create a `TimeZone`. I am
very annoyed by this, but it's probably the least bad option.
In theory, if Ru... | diff --git a/src/util/array_str.rs b/src/util/array_str.rs
index fab7be6..8428ab1 100644
--- a/src/util/array_str.rs
+++ b/src/util/array_str.rs
@@ -28,22 +28,22 @@ impl<const N: usize> ArrayStr<N> {
///
/// If the given string exceeds `N` bytes, then this returns
/// `None`.
- pub(crate) fn new(s: &s... | 1 | 8 | 8 |
6f1e421c7d4ca2245275c46b6f4d41b7808d6e41 | Andrew Gallant | 2025-02-18T22:26:49 | tests: sprinkle `cfg(not(miri))` in various spots
This doesn't even make it possible to run `cargo miri test` for
Jiff, but you can at least now do `cargo miri test tz`. I want
this because I anticipate writing some tricky `unsafe` to do
pointer tagging for `TimeZone`.
The main problem with miri is unfortunately that... | diff --git a/src/civil/date.rs b/src/civil/date.rs
index 7bac0e5..6dfc8de 100644
--- a/src/civil/date.rs
+++ b/src/civil/date.rs
@@ -3796,6 +3796,7 @@ mod tests {
}
#[test]
+ #[cfg(not(miri))]
fn all_days_to_date_roundtrip() {
for rd in -100_000..=100_000 {
let rd = UnixEpochDa... | 9 | 21 | 2 |
23ea9559b7dc8ca669fe249cb4120b665faf2b1b | Andrew Gallant | 2025-02-18T02:53:05 | error: make it possible to build an `Error` in a `const` context
The resulting error is total shit, but at least one can be
constructed. We could improve the messages to arbitrary strings
using pointer tagging, but let's not get ahead of ourselves. | diff --git a/src/error.rs b/src/error.rs
index 88e86ff..cac10ed 100644
--- a/src/error.rs
+++ b/src/error.rs
@@ -54,7 +54,7 @@ pub struct Error {
/// This also makes clones cheap. And it also make the size of error equal
/// to one word (although a `Box` would achieve that last goal). This is
/// why we ... | 1 | 50 | 15 |
475a2ce010d44f5b17268c6eb3b3d3a9ec69ee90 | Andrew Gallant | 2025-02-18T02:52:09 | tz: tighten up `posix` module
When I initially wrote the POSIX time zone parser, I had assumed
that we would use both strict POSIX time zones and IANA v3+ POSIX
time zones. But in practice, we just use the latter.
So, let's bake that into the API and tighten things up a bit. We
remove `IanaTz` and centralize everythi... | diff --git a/src/fmt/temporal/parser.rs b/src/fmt/temporal/parser.rs
index f37c568..d0f33f5 100644
--- a/src/fmt/temporal/parser.rs
+++ b/src/fmt/temporal/parser.rs
@@ -530,9 +530,11 @@ impl DateTimeParser {
}
#[cfg(feature = "alloc")]
{
- match crate::tz::posix::IanaTz::parse_v3pl... | 5 | 86 | 93 |
ffa9ab999fbd093e6da392a811624259b691f827 | Andrew Gallant | 2025-02-17T12:55:35 | jiff-diesel-0.1.2 | diff --git a/crates/jiff-diesel/Cargo.toml b/crates/jiff-diesel/Cargo.toml
index 2634f05..75b3097 100644
--- a/crates/jiff-diesel/Cargo.toml
+++ b/crates/jiff-diesel/Cargo.toml
@@ -1,6 +1,6 @@
[package]
name = "jiff-diesel"
-version = "0.1.1" #:version
+version = "0.1.2" #:version
authors = ["Andrew Gallant <jamsl... | 1 | 1 | 1 |
0054d73b741eaf5f60482e81286db0d8d6345212 | Andrew Gallant | 2025-02-17T12:54:36 | jiff-diesel-0.1.1 | diff --git a/crates/jiff-diesel/Cargo.toml b/crates/jiff-diesel/Cargo.toml
index 79bd40e..2634f05 100644
--- a/crates/jiff-diesel/Cargo.toml
+++ b/crates/jiff-diesel/Cargo.toml
@@ -1,6 +1,6 @@
[package]
name = "jiff-diesel"
-version = "0.1.0" #:version
+version = "0.1.1" #:version
authors = ["Andrew Gallant <jamsl... | 1 | 1 | 1 |
07e010b494372f6c3683e660a085d78e6d29e6cc | Andrew Gallant | 2025-02-17T12:53:39 | jiff-diesel: do some line wrapping
Some of these lines were obscenely long. So wrap them. I'm not sure
why rustfmt isn't doing this for me, but maybe the types are so long
that it breaks rustfmt? Lol. | diff --git a/crates/jiff-diesel/src/nullable.rs b/crates/jiff-diesel/src/nullable.rs
index 881ae5b..03a5c9a 100644
--- a/crates/jiff-diesel/src/nullable.rs
+++ b/crates/jiff-diesel/src/nullable.rs
@@ -15,9 +15,18 @@ use crate::ToDiesel;
any(feature = "mysql", feature = "postgres", feature = "sqlite"),
derive(... | 2 | 24 | 6 |
1756ba03a47de6c98e8be9bb16cdfa2ef69ce59d | Gary Coady | 2025-02-17T12:51:21 | jiff-diesel: update tests to use Insertable and Selectable derivations
This moves the majority of tests to use the `Insertable` and `Selectable`
derivations from the query struct. This follows idiomatic Diesel usage, and
also tests actual database round-tripping via database stores.
The downside of this is that t... | diff --git a/examples/diesel-mysql/main.rs b/examples/diesel-mysql/main.rs
index a4ff7dd..3a6534c 100644
--- a/examples/diesel-mysql/main.rs
+++ b/examples/diesel-mysql/main.rs
@@ -1,10 +1,32 @@
use diesel::{
- connection::Connection, mysql::MysqlConnection, query_dsl::RunQueryDsl,
- sql_query, sql_types, Querya... | 4 | 566 | 86 |
fd884445e1446826f7aface9ea4f49b2c6cced64 | Josh Triplett | 2025-02-14T19:58:34 | doc: fix typo
PR #264 | diff --git a/src/timestamp.rs b/src/timestamp.rs
index 7b601d1..cc3be4a 100644
--- a/src/timestamp.rs
+++ b/src/timestamp.rs
@@ -364,7 +364,7 @@ impl Timestamp {
/// As a timestamp, it corresponds to `0` nanoseconds.
///
/// A timestamp is positive if and only if it is greater than the Unix
- /// epoc... | 1 | 1 | 1 |
96ec82d6f0d97a03690a287806950b40449decc6 | Gary Coady | 2025-02-14T19:56:18 | jiff-diesel: implement AsExpression for nullable wrappers
PR #253 | diff --git a/crates/jiff-diesel/src/nullable.rs b/crates/jiff-diesel/src/nullable.rs
index c4ce2c9..881ae5b 100644
--- a/crates/jiff-diesel/src/nullable.rs
+++ b/crates/jiff-diesel/src/nullable.rs
@@ -11,6 +11,13 @@ use crate::ToDiesel;
/// This can be used when deriving [`diesel::deserialize::Queryable`]
/// or [`di... | 1 | 88 | 23 |
e094b01adfc0faaeb78f2cf0b4888b028c686bee | Andrew Gallant | 2025-02-11T02:39:16 | crates: re-add jiff-{icu,sqlx,diesel} to workspace
Now that `jiff 0.2` is published, I guess this should work now. | diff --git a/Cargo.toml b/Cargo.toml
index a77ba34..352279a 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -31,17 +31,14 @@ include = [
[workspace]
members = [
"crates/jiff-cli",
- # "crates/jiff-diesel",
- # "crates/jiff-icu",
- # "crates/jiff-sqlx",
+ "crates/jiff-diesel",
+ "crates/jiff-icu",
+ "crates/jiff-... | 1 | 3 | 6 |
d0d1b2ec274ece09916acd985da60bf9a01be40e | Andrew Gallant | 2025-02-11T02:30:53 | crates: temporarily disconnect jiff-{icu,sqlx,diesel} from workspace
... I can't figure out how to publish `jiff 0.2` with these crates
trying to select `jiff 0.2` from crates.io... which doesn't exist yet.
WTF. I guess I'm probably holding Cargo workspaces wrong. ¯\_(ツ)_/¯ | diff --git a/Cargo.toml b/Cargo.toml
index 38720ec..201e948 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -31,14 +31,17 @@ include = [
[workspace]
members = [
"crates/jiff-cli",
- "crates/jiff-diesel",
- "crates/jiff-icu",
- "crates/jiff-sqlx",
+ # "crates/jiff-diesel",
+ # "crates/jiff-icu",
+ # "crates/jiff-... | 4 | 15 | 12 |
c587fddbb72ed8bdc3c58ed859dbc2efe4799d79 | Andrew Gallant | 2025-02-11T02:11:50 | fmt: add table of different formats supported by Jiff
I think someone asked for this when `jiff 0.1` was released, and I
thought it was a good idea then. But I hadn't gotten around to it. With
the number of formats Jiff supports having expanded a bit since then, I
think now is a good time to get this documented. | diff --git a/src/fmt/mod.rs b/src/fmt/mod.rs
index 442ef60..fc7a7f4 100644
--- a/src/fmt/mod.rs
+++ b/src/fmt/mod.rs
@@ -5,6 +5,164 @@ Note that for most use cases, you should be using the corresponding
[`Display`](std::fmt::Display) or [`FromStr`](std::str::FromStr) trait
implementations for printing and parsing res... | 2 | 195 | 14 |
8b61b3260227e5db0e0b9e08bb6741fa11eeca9f | Gary Coady | 2025-02-10T22:00:00 | jiff-diesel: add Nullable wrappers for Diesel types
PR #248 | diff --git a/crates/jiff-diesel/src/lib.rs b/crates/jiff-diesel/src/lib.rs
index 2869955..822ed00 100644
--- a/crates/jiff-diesel/src/lib.rs
+++ b/crates/jiff-diesel/src/lib.rs
@@ -27,6 +27,11 @@ from the wrapper type to the corresponding Jiff type.
* There are `From` trait implementations from the wrapper type to the... | 3 | 410 | 0 |
65d48d1450d8c4c545a28cbb62323c65b134cb35 | Gary Coady | 2025-02-10T21:20:42 | examples/diesel-*: update examples to use QueryableByName and sql_query
PR #247 | diff --git a/examples/diesel-mysql/main.rs b/examples/diesel-mysql/main.rs
index d9d32f9..a4ff7dd 100644
--- a/examples/diesel-mysql/main.rs
+++ b/examples/diesel-mysql/main.rs
@@ -1,6 +1,6 @@
use diesel::{
- connection::Connection, dsl::sql, mysql::MysqlConnection,
- query_dsl::RunQueryDsl, select, sql_types, E... | 3 | 138 | 108 |
451d681511172a8f2761acf9f148a6bfa4c1098f | Gary Coady | 2025-02-09T14:25:05 | jiff-diesel: depend on Diesel's `postgres_backend` and `mysql_backend` features
Diesel splits the mysql implementation between the `mysql_backend` feature
which implements the types/traits/etc., and the `mysql` feature which also
interfaces with the mysql client library. By depending on the
backend feature, we can ... | diff --git a/crates/jiff-diesel/Cargo.toml b/crates/jiff-diesel/Cargo.toml
index 7027415..4af22fd 100644
--- a/crates/jiff-diesel/Cargo.toml
+++ b/crates/jiff-diesel/Cargo.toml
@@ -21,8 +21,8 @@ path = "src/lib.rs"
[features]
default = []
-mysql = ["diesel/mysql"]
-postgres = ["diesel/postgres"]
+mysql = ["diesel/m... | 1 | 2 | 2 |
a280ac6d46360d8af09da3c688f3bb6ec392a7a4 | Gary Coady | 2025-02-09T14:13:09 | jiff-diesel: Diesel error needs to implement `std::error::Error`
We fix this by making `jiff-diesel` always enable Jiff's
`std` feature.
PR #244 | diff --git a/crates/jiff-diesel/Cargo.toml b/crates/jiff-diesel/Cargo.toml
index 41eec6b..7027415 100644
--- a/crates/jiff-diesel/Cargo.toml
+++ b/crates/jiff-diesel/Cargo.toml
@@ -26,7 +26,7 @@ postgres = ["diesel/postgres"]
sqlite = ["diesel/sqlite"]
[dependencies]
-jiff = { path = "../..", default-features = fal... | 1 | 1 | 1 |
bc2b1a9eef296807484420b91ec26eaa56648b40 | Gary Coady | 2025-02-09T13:48:09 | jiff-diesel: derive the `Queryable` trait for diesel wrapper types
PR #243 | diff --git a/crates/jiff-diesel/src/mysql.rs b/crates/jiff-diesel/src/mysql.rs
index 401d911..c43cf46 100644
--- a/crates/jiff-diesel/src/mysql.rs
+++ b/crates/jiff-diesel/src/mysql.rs
@@ -1,5 +1,5 @@
use diesel::{
- deserialize::{self, FromSql, Queryable},
+ deserialize::{self, FromSql},
mysql::{
... | 4 | 18 | 109 |
ef211be259a80375cd2737b48d999636e4d219ac | Andrew Gallant | 2025-02-08T20:22:15 | jiff-sqlx: fix link to examples | diff --git a/crates/jiff-sqlx/src/lib.rs b/crates/jiff-sqlx/src/lib.rs
index 9145041..443202a 100644
--- a/crates/jiff-sqlx/src/lib.rs
+++ b/crates/jiff-sqlx/src/lib.rs
@@ -56,7 +56,7 @@ In the future, it may be prudent for this crate to be upstreamed into SQLx
itself.
[sqlx]: https://docs.rs/sqlx/0.8
-[examples]: ... | 1 | 1 | 1 |
80255febda9ec0978d849350fecca67cfbda0318 | Andrew Gallant | 2025-02-02T19:32:04 | crates: fixup homepage github URLs
I forgot this in #236. | diff --git a/crates/jiff-cli/Cargo.toml b/crates/jiff-cli/Cargo.toml
index 555b8f8..5d9d84a 100644
--- a/crates/jiff-cli/Cargo.toml
+++ b/crates/jiff-cli/Cargo.toml
@@ -8,7 +8,7 @@ A command line tool for supporting the development of the Jiff datetime
library. Currently, this tool is not published and only used for d... | 3 | 3 | 3 |
20e29e66c1e73170287465eb6265f4f6ddc5d950 | Andrew Gallant | 2025-01-30T12:31:08 | tz/db/bundled: remove unused imports | diff --git a/src/tz/db/bundled/enabled.rs b/src/tz/db/bundled/enabled.rs
index 163fcd1..06a7179 100644
--- a/src/tz/db/bundled/enabled.rs
+++ b/src/tz/db/bundled/enabled.rs
@@ -1,5 +1,3 @@
-use alloc::{string::String, vec::Vec};
-
use crate::tz::{TimeZone, TimeZoneNameIter};
pub(crate) struct Database; | 1 | 0 | 2 |
b0c3a38f919038d107ef3b207f587beb839c858e | Andrew Gallant | 2025-02-01T18:25:47 | tz: add one bug fix
I meant to include this in #234, but it slipped my notice.
Basically, if two offsets compare equal as-is, then it should always
return true. Before this, if our parsed offset was `-00:44:30` and the
time zone offset was `-00:44:30`, we were returning them as not equal
because we skipped to roundin... | diff --git a/src/tz/offset.rs b/src/tz/offset.rs
index cb0a34a..8aeaaaa 100644
--- a/src/tz/offset.rs
+++ b/src/tz/offset.rs
@@ -1836,7 +1836,7 @@ impl OffsetConflict {
/// for the given datetime unambiguously has offset -00:44:30",
/// );
/// let is_equal = |parsed: Offset, candidate: Offset| {
- ... | 1 | 1 | 1 |
94e8270bf06f0e760e80bcb19a6e3f41033e14a9 | Andrew Gallant | 2025-01-23T16:47:00 | civil: to_date is deprecated, don't remove it yet!
Oof. | diff --git a/src/civil/iso_week_date.rs b/src/civil/iso_week_date.rs
index ff78036..d5dea30 100644
--- a/src/civil/iso_week_date.rs
+++ b/src/civil/iso_week_date.rs
@@ -312,7 +312,6 @@ impl ISOWeekDate {
/// Deprecated APIs.
impl ISOWeekDate {
- /*
/// A deprecated equivalent to [`ISOWeekDate::date`].
... | 1 | 0 | 2 |
a762e7164b13f3cf82a80187471fcf30da86e37d | Andrew Gallant | 2025-01-23T16:46:10 | civil: remove uses of deprecated ISOWeekDate::to_date
I should get better at dog-fooding. | diff --git a/src/civil/date.rs b/src/civil/date.rs
index a73e6f0..bc1a18a 100644
--- a/src/civil/date.rs
+++ b/src/civil/date.rs
@@ -3766,7 +3766,7 @@ mod tests {
for day in 20..=days_in_month(year, month).get() {
let date = date(year.get(), month.get(), day);
... | 2 | 9 | 7 |
717b3f01ea6ab7bfaf24fc173101616d48e3674e | Andrew Gallant | 2025-01-21T20:04:26 | span: unexport unused crate-public API | diff --git a/src/span.rs b/src/span.rs
index db6bc0e..981d05f 100644
--- a/src/span.rs
+++ b/src/span.rs
@@ -4331,7 +4331,7 @@ impl Unit {
/// can vary in time depending on its reference point. A 'day' can as well,
/// but we sorta special case 'day' to mean '24 hours' for cases where
/// the user is dea... | 1 | 1 | 1 |
9dd0b5ea3c64a9838f14e47ccdab908dcbc3ee90 | Andrew Gallant | 2025-01-16T18:34:37 | lint: squash dead code warnings
In most cases we just squash the warning instead of cfg'ing out the
code. It is just super painful to correctly annotate everything piece of
code otherwise. | diff --git a/src/lib.rs b/src/lib.rs
index f4d73c2..4869581 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -669,7 +669,15 @@ For more, see the [`fmt::serde`] sub-module. (This requires enabling Jiff's
deny(rustdoc::broken_intra_doc_links)
)]
// These are just too annoying to squash otherwise.
-#![cfg_attr(not(featu... | 2 | 10 | 2 |
684839fae407c94b5ef0d7150426b44126086a93 | potatotech | 2025-01-18T13:01:29 | doc: suggest `intz` in `to_zoned` docs
PR #207 | diff --git a/src/civil/date.rs b/src/civil/date.rs
index b810b30..116554a 100644
--- a/src/civil/date.rs
+++ b/src/civil/date.rs
@@ -1236,6 +1236,10 @@ impl Date {
/// for more details. Note that ambiguous datetimes are handled in the same
/// way as `DateTime::to_zoned`.
///
+ /// In the common case ... | 3 | 9 | 1 |
6ca47d2afc34fa561c43603ad2bfeccf0534d30b | Andrew Gallant | 2025-01-16T18:34:03 | deps: bump jiff-tzdb-platform | diff --git a/Cargo.toml b/Cargo.toml
index b10842e..01b8655 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -109,7 +109,7 @@ serde = { version = "1.0.203", optional = true, default-features = false }
# is enabled by default, but that the `tzdb-bundle-platform` crate is only
# actually used on platforms without a system t... | 1 | 1 | 1 |
3867b6dd910d3641fc7ca7a4ef9b04017ed51771 | Andrew Gallant | 2025-01-16T18:33:46 | jiff-tzdb-platform-0.1.2 | diff --git a/jiff-tzdb-platform/Cargo.toml b/jiff-tzdb-platform/Cargo.toml
index ea96054..ec2a2cb 100644
--- a/jiff-tzdb-platform/Cargo.toml
+++ b/jiff-tzdb-platform/Cargo.toml
@@ -1,6 +1,6 @@
[package]
name = "jiff-tzdb-platform"
-version = "0.1.1" #:version
+version = "0.1.2" #:version
authors = ["Andrew Gallant... | 1 | 1 | 1 |
d1364880d7488ebcb0d3c42e2a20e2aae025f6eb | Andrew Gallant | 2025-01-16T18:33:36 | deps: bump jiff-tzdb | diff --git a/Cargo.toml b/Cargo.toml
index 3b7090a..b10842e 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -100,7 +100,7 @@ tzdb-concatenated = ["std"]
js = ["dep:wasm-bindgen", "dep:js-sys"]
[dependencies]
-jiff-tzdb = { version = "0.1.1", path = "jiff-tzdb", optional = true }
+jiff-tzdb = { version = "0.1.2", path =... | 2 | 2 | 2 |
837b438f83d0f559ea80d95f153347964efcb680 | Andrew Gallant | 2025-01-16T18:33:01 | jiff-tzdb-0.1.2 | diff --git a/jiff-tzdb/Cargo.toml b/jiff-tzdb/Cargo.toml
index 34ba169..3468ea7 100644
--- a/jiff-tzdb/Cargo.toml
+++ b/jiff-tzdb/Cargo.toml
@@ -1,6 +1,6 @@
[package]
name = "jiff-tzdb"
-version = "0.1.1" #:version
+version = "0.1.2" #:version
authors = ["Andrew Gallant <jamslam@gmail.com>"]
license = "Unlicense ... | 1 | 1 | 1 |
c9c3ec5f2da5788174247466413240ab86e527ca | philomathic_life | 2025-01-12T18:49:58 | perf: allocate exact capacity for RFC 9110 strings
PR #198 | diff --git a/src/fmt/rfc2822.rs b/src/fmt/rfc2822.rs
index 132a3c6..1da745a 100644
--- a/src/fmt/rfc2822.rs
+++ b/src/fmt/rfc2822.rs
@@ -1238,7 +1238,7 @@ impl DateTimePrinter {
&self,
timestamp: &Timestamp,
) -> Result<alloc::string::String, Error> {
- let mut buf = alloc::string::String:... | 1 | 1 | 1 |
2f57df3d31634c993269c589e6db721fd3ab3d73 | Andrew Gallant | 2025-01-01T17:38:32 | span: remove commented out incorrect optimization
I commented this out a while ago and meant to remove it since it turned
out to be incorrect. Unfortunately, I don't quite remember the details
of why it is wrong. | diff --git a/src/span.rs b/src/span.rs
index cb96331..a901557 100644
--- a/src/span.rs
+++ b/src/span.rs
@@ -4986,14 +4986,6 @@ impl<'a> SpanRound<'a> {
smallest = smallest.singular(),
));
}
- // Now that we've got our configuration, we can actually short circuit
- /... | 1 | 0 | 8 |
3b930e41b89844d6ea9a6cfe74b3c9ceba1d3364 | Andrew Gallant | 2024-12-31T19:09:04 | util: centralize definition of time zone abbreviation
I might want to use it in more places, so I think it makes
sense to give a more utilitarian home. | diff --git a/src/fmt/strtime/mod.rs b/src/fmt/strtime/mod.rs
index c56fbc9..f824fab 100644
--- a/src/fmt/strtime/mod.rs
+++ b/src/fmt/strtime/mod.rs
@@ -249,7 +249,9 @@ use crate::{
},
tz::{Offset, OffsetConflict, TimeZone, TimeZoneDatabase},
util::{
- self, escape,
+ self,
+ array_s... | 3 | 64 | 60 |
7a41bf39facccfe5393edccb8359426122bf77a2 | Andrew Gallant | 2024-12-31T18:01:48 | error: make Debug match Display
I'm kind of annoyed that an `Error`'s `Debug` impl shows the full
internal representation. In particular, this is something that *users*
of Jiff will see due to `unwrap()` and the like. So it makes sense to
tighten it up a bit. Jiff users don't need to care about the internal
error kind... | diff --git a/src/error.rs b/src/error.rs
index e8beb1e..8549d6e 100644
--- a/src/error.rs
+++ b/src/error.rs
@@ -43,7 +43,7 @@ pub(crate) use err;
/// [about error types].
///
/// [about error types]: https://github.com/BurntSushi/jiff/issues/8
-#[derive(Clone, Debug)]
+#[derive(Clone)]
pub struct Error {
/// ... | 1 | 23 | 1 |
1834cb510c4be27afde04152608ef3c158baadef | Andrew Gallant | 2024-12-31T17:56:26 | timestamp: improve docs surrounding arithmetic panics
While this doesn't remove the arithmetic impls (I think that would be a
bridge too far), we try to make the docs a little clearer and slightly
improve the error message. The error message still doesn't say "why,"
but does hint at it a bit.
Closes #145 | diff --git a/src/span.rs b/src/span.rs
index 9a00d04..cb96331 100644
--- a/src/span.rs
+++ b/src/span.rs
@@ -3047,7 +3047,9 @@ impl Span {
let non_time_unit = self.largest_calendar_unit()?;
Some(err!(
"operation can only be performed with units of hours \
- or smaller, but fou... | 2 | 26 | 15 |
396fdcfa61fb954e4c420645d0eb6777627cdd7f | Andrew Gallant | 2024-12-28T21:21:42 | tz: remove `Box<str>` for fixed time zone
This removes yet another needless but harmless allocation in pursuit of
progress on #168.
In this case, returning a `&str` is part of the public API when a time
zone is a fixed offset, so we do need to store the offset as a string
somewhere. Thankfully, it's pretty small (sma... | diff --git a/src/tz/mod.rs b/src/tz/mod.rs
index 55776a5..1931ef4 100644
--- a/src/tz/mod.rs
+++ b/src/tz/mod.rs
@@ -84,7 +84,6 @@ TO get the system's default time zone, use [`TimeZone::system`].
*/
use alloc::{
- boxed::Box,
string::{String, ToString},
sync::Arc,
};
@@ -92,6 +91,7 @@ use alloc::{
us... | 3 | 113 | 7 |
95dbd44a927c169e8409eb515d609773943cdcf7 | Andrew Gallant | 2024-12-28T19:29:49 | fmt: rejigger some `alloc` imports
I did this to get a sense for how much `jiff::fmt` depends on `alloc`.
It actually looks pretty good. Most uses of `alloc` are just for
convenience `String` APIs.
The one hiccup is `strtime` where it uses `String` for IANA time zone
names and time zone abbreviations. I think I can u... | diff --git a/src/fmt/mod.rs b/src/fmt/mod.rs
index 86959f4..987f4a8 100644
--- a/src/fmt/mod.rs
+++ b/src/fmt/mod.rs
@@ -7,8 +7,6 @@ implementations for printing and parsing respectively. The APIs in this module
provide more configurable support for printing and parsing.
*/
-use alloc::{string::String, vec::Vec};
-... | 5 | 53 | 47 |
30dcb95447c21cdf95a969fb9a6da9a6a92d43cb | Andrew Gallant | 2024-12-28T19:06:21 | tz/posix: remove last bits of reliance on `alloc`
We no longer keep around the "original" string we parsed for POSIX time
zones. Instead, we implement `Display` for our POSIX TZ types, and
that's how we'll show it in places that need it. This does mean that we
show the parsed form instead of what the user actually pro... | diff --git a/src/tz/db/bundled/enabled.rs b/src/tz/db/bundled/enabled.rs
index d0f393d..ff63c6a 100644
--- a/src/tz/db/bundled/enabled.rs
+++ b/src/tz/db/bundled/enabled.rs
@@ -33,7 +33,7 @@ impl BundledZoneInfo {
}
};
#[cfg(feature = "std")]
- self::global::add(&tz);
+ self... | 7 | 271 | 68 |
a8795c35e0a8dcc46527008f6a086c8cb7251f7d | Andrew Gallant | 2024-12-28T03:51:01 | cargo: fix feature config
It turned out that our feature config testing was somewhat FUBAR.
Namely, we were accidentally and unconditionally always enabling the
`std` feature (among other default features) when running tests. This
was a result of this *dev* dependency:
jiff = { path = "./", features = ["serde"] }
... | diff --git a/Cargo.toml b/Cargo.toml
index cfe9d70..285a6c6 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -108,7 +108,7 @@ icu = { version = "1.5.0", features = ["std"] }
insta = "1.39.0"
# We force `serde` to be enabled in dev mode so that the docs render and test
# correctly. This is highly suspicious.
-jiff = { pat... | 13 | 96 | 89 |
4e82a43e23972b7753811a2a7cace07ef45dcb25 | Andrew Gallant | 2024-12-27T19:55:32 | tz/posix: remove `Box<str>` used for POSIX time zone abbreviations
As part of working toward #168, this commit puts POSIX time zone
abbreviations inline instead of on the heap. We can generally rely on
time zone abbreviations to be short, so this is "fine."
This is obviously more annoying than just using `Box<str>`, ... | diff --git a/src/tz/posix.rs b/src/tz/posix.rs
index 054afff..532ca7f 100644
--- a/src/tz/posix.rs
+++ b/src/tz/posix.rs
@@ -61,6 +61,7 @@ use crate::{
timestamp::Timestamp,
tz::{AmbiguousOffset, Dst, Offset},
util::{
+ array_str::ArrayStr,
escape::{Byte, Bytes},
parse,
... | 3 | 154 | 36 |
b8dad89b7dcd6b4e7605102a269880b4405dfb3c | Andrew Gallant | 2024-12-27T18:37:49 | lint: remove some `allow(..)` annotations
This is mostly just removing some dead code, and in one case, finishing
a refactoring to `SignedDuration` in `src/tz/posix.rs`. | diff --git a/src/span.rs b/src/span.rs
index ebe96e5..15fa32f 100644
--- a/src/span.rs
+++ b/src/span.rs
@@ -2907,7 +2907,7 @@ impl Span {
/// if this span has maximal values for all units, then rebalancing is
/// not possible because the number of days after balancing would exceed
/// the limit.
- #[... | 4 | 18 | 90 |
b482926123daf2a4d744a598c96f0d2ecafab4dd | Ben Beasley | 2024-12-09T14:48:27 | deps: update `chrono-tz` dev-dependency
I confirmed that `cargo test --workspace` still passes.
From the
[release notes](https://github.com/chronotope/chrono-tz/releases/tag/v0.10.0):
```
### Changes
- Make `OffsetName::abbreviation` return an `Option`.
This reflects that numeric values such as `+11` are n... | diff --git a/Cargo.toml b/Cargo.toml
index 6ac317e..90c4b4d 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -99,7 +99,7 @@ wasm-bindgen = { version = "0.2.70", optional = true }
[dev-dependencies]
anyhow = "1.0.81"
chrono = { version = "0.4.38", features = ["serde"] }
-chrono-tz = "0.9.0"
+chrono-tz = "0.10.0"
# This a... | 1 | 1 | 1 |
d538cd95aface3e93c9fcd116b48b47734a0339a | Min Kim | 2024-10-17T21:16:23 | doc: correct function reference in Timestamp
The documentation for the deprecated `as_duration` function in
`Timestamp` referenced a non-existent function `as_signed_duration`.
It should point to `as_jiff_duration` instead.
Closes #142 | diff --git a/src/timestamp.rs b/src/timestamp.rs
index 008be0c..79230b4 100644
--- a/src/timestamp.rs
+++ b/src/timestamp.rs
@@ -2132,7 +2132,7 @@ impl Timestamp {
/// ```
#[deprecated(
since = "0.1.5",
- note = "use Timestamp::as_signed_duration instead"
+ note = "use Timestamp::as_jif... | 1 | 1 | 1 |
74af3856f1b953f567f2a2e08b137fefeb04d52a | LeoniePhiline | 2024-10-30T16:18:32 | docs: fix a typo
Closes #150 | diff --git a/src/civil/date.rs b/src/civil/date.rs
index 7585e1d..1b413fe 100644
--- a/src/civil/date.rs
+++ b/src/civil/date.rs
@@ -1236,7 +1236,7 @@ impl Date {
///
/// # Example
///
- /// This example shows how to created a zoned value with a fixed time zone
+ /// This example shows how to creat... | 3 | 3 | 3 |
a937e453ae2b073cf3ec8502c054a2cf2a752475 | LeoniePhiline | 2024-10-31T15:54:42 | test: verify serde YAML compatibility
Deserializing `jiff` types from YAML is not
currently possible.
Neither `serde_yaml` nor its maintained fork
`serde_yml` support deserializing from bytes.
These tests serve in both
- proving the current incompatibility
- verifying the fix in the following commit,
which migrat... | diff --git a/Cargo.toml b/Cargo.toml
index d9fd996..44d2008 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -110,6 +110,7 @@ jiff = { path = "./", features = ["serde"] }
quickcheck = { version = "1.0.3", default-features = false }
serde = { version = "1.0.203", features = ["derive"] }
serde_json = "1.0.117"
+serde_yml =... | 8 | 230 | 2 |
2f57962ad14a99cab2dbb66014b5b221727180f1 | Andrew Gallant | 2024-11-01T12:11:48 | fmt/rfc2822: add routines explicitly for RFC 9110
It seems that RFC 9110 does not support offsets like `-0000` in the
date format, and also requires that days be padded to two digits in all
cases. This meant that Jiff's RFC 2822 output was, strictly speaking,
incompatible with a strict implementation of RFC 9110. For ... | diff --git a/src/fmt/rfc2822.rs b/src/fmt/rfc2822.rs
index 1fc9de9..1851705 100644
--- a/src/fmt/rfc2822.rs
+++ b/src/fmt/rfc2822.rs
@@ -579,7 +579,7 @@ impl DateTimeParser {
}
let (day, input) = input.split_at(digits);
let day = parse::i64(day).with_context(|| {
- err!("failed to ... | 1 | 174 | 1 |
d3e0a16495df21ba86b908f6152932178753e5f0 | Plecra | 2024-09-22T13:02:20 | doc: tweak definition of Timestamp
The previous definition was slightly incorrect, or
at best, imprecise. It is actually pretty important
to mention that `Timestamp` is in the Unix time
scale (the same as the POSIX time scale), as this
clarifies that it behaves as if leap seconds don't
exist.
PR #134 | diff --git a/src/timestamp.rs b/src/timestamp.rs
index 5df3d15..5510114 100644
--- a/src/timestamp.rs
+++ b/src/timestamp.rs
@@ -23,7 +23,7 @@ use crate::{
/// An instant in time represented as the number of nanoseconds since the Unix
/// epoch.
///
-/// A timestamp is always in UTC.
+/// A timestamp is always in th... | 1 | 1 | 1 |
9d2b8a6963a975cf7ab2f95a310fde82ad833b7f | Andrew Gallant | 2024-09-08T17:41:41 | fmt: rejigger internal precision setting
Previously, we accepted `precision: u8` and inferred "not set" as
"automatic precision." But this actually ends up making some call sites
more complicated than they need to be. It's better to just require an
`Option<u8>` since that tends to be what callers have. | diff --git a/src/fmt/strtime/format.rs b/src/fmt/strtime/format.rs
index 45bb4f9..c185139 100644
--- a/src/fmt/strtime/format.rs
+++ b/src/fmt/strtime/format.rs
@@ -569,10 +569,7 @@ impl Extension {
) -> Result<(), Error> {
let number = number.into();
- let mut formatter = FractionalFormatter::ne... | 3 | 14 | 20 |
b32942779875d8c13b111d88567771592b954245 | Andrew Gallant | 2024-09-08T17:12:48 | util: move UTF-8 decoding to its own module
This moves UTF-8 decoding from an internal helper in
`jiff::util::escape` to a crate-wide helper in `jiff::util::utf8`.
This makes it easier for other code to use this. I may end up using it
for #111, for example. | diff --git a/src/util/escape.rs b/src/util/escape.rs
index a2ebf94..4528f4b 100644
--- a/src/util/escape.rs
+++ b/src/util/escape.rs
@@ -4,6 +4,8 @@ Provides convenience routines for escaping raw bytes.
This was copied from `regex-automata` with a few light edits.
*/
+use crate::util::utf8;
+
/// Provides a conven... | 3 | 52 | 50 |
60aef20dc262745bbbed164d67bdff18ddef71ce | Andrew Gallant | 2024-09-07T14:12:34 | fmt: simplify integer formatting slightly
It turns out that "no padding set" and "padding is zero" result in the
same behavior. I think this might have been different back when I was
(ab)using the integer formatting to also do fractional formatting. But
either way, eliminating a redundant internal state is a win.
I d... | diff --git a/src/fmt/util.rs b/src/fmt/util.rs
index 568abf3..97c5d0e 100644
--- a/src/fmt/util.rs
+++ b/src/fmt/util.rs
@@ -18,7 +18,7 @@ use crate::{
#[derive(Clone, Copy, Debug)]
pub(crate) struct DecimalFormatter {
force_sign: Option<bool>,
- minimum_digits: Option<u8>,
+ minimum_digits: u8,
paddi... | 1 | 6 | 8 |
87054e807d2e2e88a3ba5969c8e1f4a40355b21c | Andrew Gallant | 2024-09-07T13:57:05 | cargo: bump jiff-tzdb and jiff-tzdb-platform to 0.1.1 | diff --git a/Cargo.toml b/Cargo.toml
index 303a1c3..5d673bf 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -75,7 +75,7 @@ tzdb-zoneinfo = ["std"]
js = ["dep:wasm-bindgen", "dep:js-sys"]
[dependencies]
-jiff-tzdb = { version = "0.1.0", path = "jiff-tzdb", optional = true }
+jiff-tzdb = { version = "0.1.1", path = "jiff... | 1 | 2 | 2 |
d6dc8841a405c1d6fbb670a8f028f0cbb901fcf7 | Andrew Gallant | 2024-09-07T13:56:10 | jiff-tzdb-platform-0.1.1 | diff --git a/jiff-tzdb-platform/Cargo.toml b/jiff-tzdb-platform/Cargo.toml
index 38d145c..bc0bbb3 100644
--- a/jiff-tzdb-platform/Cargo.toml
+++ b/jiff-tzdb-platform/Cargo.toml
@@ -1,6 +1,6 @@
[package]
name = "jiff-tzdb-platform"
-version = "0.1.0" #:version
+version = "0.1.1" #:version
authors = ["Andrew Gallant... | 1 | 1 | 1 |
7d866259a2e02dc117adedde17e9d1d896d21544 | Andrew Gallant | 2024-09-07T13:55:55 | tzdb-platform: bump jiff-tzdb to 0.1.1 | diff --git a/jiff-tzdb-platform/Cargo.toml b/jiff-tzdb-platform/Cargo.toml
index 99073eb..38d145c 100644
--- a/jiff-tzdb-platform/Cargo.toml
+++ b/jiff-tzdb-platform/Cargo.toml
@@ -21,4 +21,4 @@ bench = false
path = "lib.rs"
[dependencies]
-jiff-tzdb = { version = "0.1.0", path = "../jiff-tzdb" }
+jiff-tzdb = { ver... | 1 | 1 | 1 |
2d05b8d0a56e51b41267f512048ec60f9e17ba9a | Andrew Gallant | 2024-09-07T13:55:20 | jiff-tzdb-0.1.1 | diff --git a/jiff-tzdb/Cargo.toml b/jiff-tzdb/Cargo.toml
index c962c6e..ee2fa0b 100644
--- a/jiff-tzdb/Cargo.toml
+++ b/jiff-tzdb/Cargo.toml
@@ -1,6 +1,6 @@
[package]
name = "jiff-tzdb"
-version = "0.1.0" #:version
+version = "0.1.1" #:version
authors = ["Andrew Gallant <jamslam@gmail.com>"]
license = "Unlicense ... | 1 | 1 | 1 |
74c47270fb39ea614b37cad43324dde96924df29 | Andrew Gallant | 2024-08-31T18:08:09 | timestamp: add Timestamp::display_with_offset
This provides a more convenient and more easily discoverable interface
to `jiff::fmt::temporal::DateTimePrinter::print_timestamp_with_offset`. | diff --git a/src/lib.rs b/src/lib.rs
index a1d19b0..6493a4f 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -688,8 +688,8 @@ pub use crate::{
SpanTotal, ToSpan, Unit,
},
timestamp::{
- Timestamp, TimestampArithmetic, TimestampDifference, TimestampRound,
- TimestampSeries,
+ Timestamp... | 2 | 109 | 5 |
aa0ca78bbd3741b96bc6ca9a17e962e979b7c2e4 | Andrew Gallant | 2024-08-31T17:31:47 | fmt/rfc2822: add 'to_string' convenience routines
This mimics the previous commit, but for the RFC 2822 datetime printer. | diff --git a/src/fmt/rfc2822.rs b/src/fmt/rfc2822.rs
index b38535d..1fc9de9 100644
--- a/src/fmt/rfc2822.rs
+++ b/src/fmt/rfc2822.rs
@@ -1099,6 +1099,101 @@ impl DateTimePrinter {
DateTimePrinter { _private: () }
}
+ /// Format a `Zoned` datetime into a string.
+ ///
+ /// This never emits `-00... | 1 | 95 | 0 |
13cc178cb5b85c56a991f64cfd9c04582a75930e | Andrew Gallant | 2024-08-31T14:38:37 | fmt/temporal: add 'to_string' convenience routines to Temporal printers
I think this is a big ergonomic win for a very common case. It makes
using the datetime and span printers much nicer, and they are necessary
to use in some cases when the output format needs more configuration. | diff --git a/src/fmt/temporal/mod.rs b/src/fmt/temporal/mod.rs
index aca4c1d..c50a428 100644
--- a/src/fmt/temporal/mod.rs
+++ b/src/fmt/temporal/mod.rs
@@ -163,6 +163,8 @@ There is some more [background on Temporal's format] available.
[background on Temporal's format]: https://github.com/tc39/proposal-temporal/issue... | 1 | 252 | 6 |
b174e257217cee47086278cffe1e88f328b0d6fd | Andrew Gallant | 2024-08-31T14:25:14 | fmt: add some #[inline] annotations
These are just on the config settings. In newer versions of Rust, I'd
expect these to get cross-crate inlined since they are trivial. And of
course, when used in a const context, it doesn't matter if they get
inlined. But these really should always be inlined, so add the
annotation ... | diff --git a/src/fmt/rfc2822.rs b/src/fmt/rfc2822.rs
index 43402d8..b38535d 100644
--- a/src/fmt/rfc2822.rs
+++ b/src/fmt/rfc2822.rs
@@ -105,6 +105,7 @@ pub(crate) static DEFAULT_DATETIME_PRINTER: DateTimePrinter =
///
/// # Ok::<(), Box<dyn std::error::Error>>(())
/// ```
+#[inline]
pub fn to_string(zdt: &Zoned) -... | 2 | 13 | 0 |
8173cdbccdb48b7dc3711b9163dfd594c66e51ab | Andrew Gallant | 2024-08-31T14:16:23 | fmt/temporal: add DateTimePrinter::print_timestamp_with_offset
This makes it possible to format a `Timestamp` with a specific offset.
This was already technically possible with the `strftime` API, but
adding explicit support for it to the `jiff::fmt::temporal` module seems
like good sense.
Note that no dedicated pars... | diff --git a/src/fmt/temporal/mod.rs b/src/fmt/temporal/mod.rs
index 59552dd..3877eb5 100644
--- a/src/fmt/temporal/mod.rs
+++ b/src/fmt/temporal/mod.rs
@@ -168,7 +168,7 @@ use crate::{
error::Error,
fmt::Write,
span::Span,
- tz::{Disambiguation, OffsetConflict, TimeZoneDatabase},
+ tz::{Disambigua... | 2 | 100 | 6 |
3b4bd069d6d011bc927fccfd3d1de421c993bf37 | Andrew Gallant | 2024-08-29T15:57:01 | cargo: update keywords and categories
Jiff is no-std.
I think the `temporal` and `iana` keywords probably don't carry their
weight. So replace the with hopefully better keywords, `calendar` and
`duration`. We keep `zone` to hit "time zone" queries. | diff --git a/Cargo.toml b/Cargo.toml
index ad22303..ffb2094 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -10,8 +10,8 @@ A date-time library that encourages you to jump into the pit of success.
This library is heavily inspired by the Temporal project.
'''
-categories = ["date-and-time"]
-keywords = ["date", "time", "... | 1 | 2 | 2 |
b63d9e6eeedd1f80e93210bd9a108282f011e56c | Martin von Zweigbergk | 2024-08-28T16:07:12 | util/libm: replace `force_eval!` by `core::hint::black_box`
During an internal review of unsafe code at work, @cramertj noticed
that the the `force_eval!` can be replaced by `core::hint::black_box`,
removing the need for unsafe Rust.
It's not totally clear that this is fully correct for x86 (note: not x86-64)
si... | diff --git a/src/util/libm.rs b/src/util/libm.rs
index 75d3772..abb5fe4 100644
--- a/src/util/libm.rs
+++ b/src/util/libm.rs
@@ -24,12 +24,6 @@ pub(crate) trait Float {
fn fract(self) -> Self;
}
-macro_rules! force_eval {
- ($e:expr) => {
- core::ptr::read_volatile(&$e)
- };
-}
-
const TOINT64: f6... | 1 | 8 | 30 |
9c757139a74849b612e3ef4949f43e6cf7eed739 | Andrew Gallant | 2024-08-24T13:15:03 | zoned: add note about Timestamp::now().to_zoned()
This PR adds a note to the `Zoned::now()` documentation describing the
best way to create a `Zoned` for the current time in a time zone that
isn't the system time zone. Unfortunately, one intuitive route is
`Zoned::now().with_time_zone()`, but this 1) fetches the syste... | diff --git a/src/zoned.rs b/src/zoned.rs
index c907a12..5481fb0 100644
--- a/src/zoned.rs
+++ b/src/zoned.rs
@@ -384,6 +384,12 @@ impl Zoned {
/// compatible with the `log` crate and enable Jiff's `logging` Cargo
/// feature.)
///
+ /// To create a `Zoned` value for the current time in a particular
+ ... | 1 | 6 | 0 |
0e569cba9884a4d1093410c81324991cb3ce0984 | Rui Fan | 2024-08-18T08:24:32 | bench: add zoned benchmark for Asia/Shanghai
This is useful because Asia/Shanghai doesn't have DST, so this helps
measure the flow through the code when DST doesn't need to be handled.
That is, there are some shortcuts we can take.
Closes #103 | diff --git a/bench/src/bench.rs b/bench/src/bench.rs
index 427e40f..4de8fa6 100644
--- a/bench/src/bench.rs
+++ b/bench/src/bench.rs
@@ -116,56 +116,66 @@ fn civil_datetime_to_instant_static(c: &mut Criterion) {
/// specific time zone.
fn instant_to_civil_datetime_static(c: &mut Criterion) {
const NAME: &str = "... | 1 | 43 | 33 |
8839f9c1c7054191f5ea42298a1e6823c7d765c7 | Andrew Gallant | 2024-08-22T21:53:08 | doc: small fixes | diff --git a/src/fmt/temporal/parser.rs b/src/fmt/temporal/parser.rs
index 6120ff6..c03b552 100644
--- a/src/fmt/temporal/parser.rs
+++ b/src/fmt/temporal/parser.rs
@@ -18,7 +18,7 @@ use crate::{
pub(super) struct ParsedDateTime<'i> {
/// The original input that the datetime was parsed from.
input: escape::B... | 2 | 3 | 3 |
ae6000b1f457fbf1c650ef61c4677e033f1e83dc | Andrew Gallant | 2024-08-18T21:41:24 | cargo: permit windows-sys >=0.52.0 and <=0.59.*
This matches a similar change made by @ChrisDenton for winapi-util in:
https://github.com/BurntSushi/winapi-util/pull/19
The benefit is that this should hopefully put less constraints on the
ecosystem and to enable fewer copies of windows-sys to appear in
dependency tre... | diff --git a/Cargo.toml b/Cargo.toml
index 0d9a130..ffcacae 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -87,7 +87,7 @@ serde = { version = "1.0.203", optional = true }
jiff-tzdb-platform = { version = "0.1.0", path = "jiff-tzdb-platform", optional = true }
[target.'cfg(windows)'.dependencies.windows-sys]
-version =... | 1 | 1 | 1 |
ee53cd9f45ae9c85f34cb4553dd5d61c7ced8776 | Andrew Gallant | 2024-08-18T21:28:35 | tz: use SignedDuration for POSIX offset arithmetic
Since this operates on a civil::DateTime, I don't think there's much
benefit, although there is a slight improvement in some benchmarks
(~3%). I suspect the civil::DateTime arithmetic (particularly
civil::Time) could probably be improved. | diff --git a/src/tz/offset.rs b/src/tz/offset.rs
index cc401de..449065b 100644
--- a/src/tz/offset.rs
+++ b/src/tz/offset.rs
@@ -834,19 +834,11 @@ impl Offset {
SignedDuration::offset_until(other, self)
}
- /// Returns this offset as a [`Span`].
- #[inline]
- pub(crate) fn to_span(self) -> Span... | 2 | 26 | 32 |
b06cd131920273856609b2f8ae1cb6b54461b027 | Andrew Gallant | 2024-08-18T19:10:31 | bench: tweak benchmarks and optimize some cases
This makes some cases of adding durations to Timestamp/Zoned faster.
It does seem like there is some room left here to improved. But making
Zoned truly faster is difficult because of its higher level abstraction
facilities. | diff --git a/bench/Cargo.toml b/bench/Cargo.toml
index 50b3e05..f991be3 100644
--- a/bench/Cargo.toml
+++ b/bench/Cargo.toml
@@ -31,3 +31,6 @@ time = { version = "0.3.36", features = ["macros", "parsing"] }
[profile.release]
debug = true
+
+[profile.bench]
+debug = true
diff --git a/bench/src/bench.rs b/bench/src/b... | 6 | 161 | 24 |
0904055a11e95b83c0296d1b4f9d5f7e90f29e8e | Andrew Gallant | 2024-08-18T16:40:01 | signed_duration: add some clarifying examples | diff --git a/src/signed_duration.rs b/src/signed_duration.rs
index b19428d..2b5f157 100644
--- a/src/signed_duration.rs
+++ b/src/signed_duration.rs
@@ -170,6 +170,113 @@ use crate::util::libm::Float;
/// (as long as they're within Jiff's limits), including durations with units
/// of years, months, weeks and days. A... | 1 | 107 | 0 |
6de09490f92319a2947a206fd76b39baeb49dc8e | Andrew Gallant | 2024-08-18T15:56:23 | api: remove 'impl Into' from parameter types
I think this was an oversight. I generally don't like using 'impl Trait'
in public APIs unless there is no avoiding it. In particular, when using
'impl Trait', it inhibits the use of turbofish, which can sometimes be
useful. | diff --git a/src/civil/datetime.rs b/src/civil/datetime.rs
index b5da48c..07931f1 100644
--- a/src/civil/datetime.rs
+++ b/src/civil/datetime.rs
@@ -2140,9 +2140,9 @@ impl DateTime {
/// # Ok::<(), Box<dyn std::error::Error>>(())
/// ```
#[inline]
- pub fn round(
+ pub fn round<R: Into<DateTimeRoun... | 6 | 23 | 23 |
981d5eea004279cfafcbf5dbd135c0f1c7c0f395 | Andrew Gallant | 2024-08-18T15:49:48 | span: add support for absolute durations
This makes it possible to use `Span::checked_add` with `SignedDuration`
and `std::time::Duration`. We don't need to add any new "target" types
because we already had a `SpanArithmetic`. So mostly this was just about
adding new trait impls and support for adding an absolute dura... | diff --git a/src/span.rs b/src/span.rs
index 66fd0ce..5c02f32 100644
--- a/src/span.rs
+++ b/src/span.rs
@@ -1,9 +1,10 @@
-use core::{cmp::Ordering, time::Duration};
+use core::{cmp::Ordering, time::Duration as UnsignedDuration};
use alloc::borrow::Cow;
use crate::{
civil::{Date, DateTime, Time},
+ durati... | 1 | 297 | 126 |
f791a12beb4a4cc621f2131e083c667a25544c0c | Andrew Gallant | 2024-08-14T01:46:09 | tz: re-arrange relative ordering of methods on Offset
This makes it match how 'add' and 'until' methods are ordered on the
datetime types. | diff --git a/src/tz/offset.rs b/src/tz/offset.rs
index d29f1b5..b6c1f24 100644
--- a/src/tz/offset.rs
+++ b/src/tz/offset.rs
@@ -453,67 +453,6 @@ impl Offset {
.checked_sub_seconds(i64::from(self.seconds()))
}
- /// Returns the span of time from this offset until the other given.
- ///
- //... | 1 | 61 | 61 |
c395ac05f3e64471f7122610e6d652bd30781cbf | Andrew Gallant | 2024-08-11T13:34:02 | signed_duration: add methods for computing durations between datetimes
These methods permit completely circumventing the `Span` type. They
aren't exported yet. We'll do that by adding new methods on the various
datetime types in a subsequent commit. | diff --git a/src/signed_duration.rs b/src/signed_duration.rs
index 531a32b..01b6957 100644
--- a/src/signed_duration.rs
+++ b/src/signed_duration.rs
@@ -1,8 +1,10 @@
use core::time::Duration;
use crate::{
+ civil::{Date, DateTime, Time},
error::{err, ErrorContext},
- Error,
+ tz::Offset,
+ Error, T... | 2 | 135 | 59 |
0d1f95a99e75f90723b50439e4c6d47c836b4dcd | Andrew Gallant | 2024-08-07T14:01:53 | util/libm: add f32 libm support
This is basically just vendoring the 'f' variants (e.g., 'truncf') from
the `libm` crate. | diff --git a/src/util/libm.rs b/src/util/libm.rs
index bb2b865..75d3772 100644
--- a/src/util/libm.rs
+++ b/src/util/libm.rs
@@ -140,6 +140,123 @@ impl Float for f64 {
}
}
+impl Float for f32 {
+ fn abs(self) -> f32 {
+ if self.is_sign_negative() {
+ -self
+ } else {
+ sel... | 1 | 125 | 0 |
f77220c8289f4e794911585971aaa6a1e39c5b1d | Andrew Gallant | 2024-08-07T13:47:31 | util/libm: prepare for f32 support
Previously we framed our tiny little libm port as specific to f64. But
in order to support SignedDuration::try_from_secs_f32, I think we'll
want f32 libm routines too. So rename and tweak the trait to be less
specific to f64. | diff --git a/src/now.rs b/src/now.rs
index 68eb21c..d059dd7 100644
--- a/src/now.rs
+++ b/src/now.rs
@@ -44,7 +44,7 @@ mod sys {
use std::time::Duration;
#[cfg(not(feature = "std"))]
- use crate::util::libm::F64;
+ use crate::util::libm::Float;
let millis = js_sys::Date::new... | 4 | 20 | 20 |
318c68d9b2df13d5f3100915cbde58065c0e7bf8 | Andrew Gallant | 2024-08-07T13:43:36 | util/libm: add 'fract' method | diff --git a/src/util/libm.rs b/src/util/libm.rs
index 9e0f9de..d5982d2 100644
--- a/src/util/libm.rs
+++ b/src/util/libm.rs
@@ -6,10 +6,10 @@ wouldn't necessarily mind depending on `libm`, but I don't think there's a
way to express "only depend on a crate if some on-by-default feature is not
enabled."
-We also ver... | 1 | 9 | 4 |
529105edceebd867123244a3f9b68c96088e1537 | Rui Fan | 2024-08-18T15:36:09 | perf: avoid binary search in to_offset in some cases
Specifically, when one knows that the timestamp is
bigger than the biggest transition, we can skip the
binary search and just jump directly to using the POSIX
TZ rule. | diff --git a/src/tz/tzif.rs b/src/tz/tzif.rs
index 8990e72..a21b9fe 100644
--- a/src/tz/tzif.rs
+++ b/src/tz/tzif.rs
@@ -117,17 +117,24 @@ impl Tzif {
// The result of the dummy transition is that the code below is simpler
// with fewer special cases.
assert!(!self.transitions.is_empty(), "tr... | 1 | 18 | 11 |
76c103e50288ee326a9d58e9fff2b1af64ef9596 | Rui Fan | 2024-08-18T11:54:56 | perf: short circuit in POSIX tz when there is no DST
Ref #103
PR #104 | diff --git a/src/tz/posix.rs b/src/tz/posix.rs
index 63f7ebf..02ee604 100644
--- a/src/tz/posix.rs
+++ b/src/tz/posix.rs
@@ -286,6 +286,10 @@ impl ReasonablePosixTimeZone {
&self,
timestamp: Timestamp,
) -> (Offset, Dst, &str) {
+ if self.dst.is_none() {
+ return (self.std_offse... | 1 | 4 | 0 |
5b110898ec6572b04d08949ed05ea466c8ae86ff | Andrew Gallant | 2024-08-15T14:42:35 | lint: remove ambiguous_negative_literal allow's in doc tests
In favor of a global one. But also, while these aren't technically
needed any more[1,2], we leave them be for now while things settle.
[1]: https://github.com/rust-lang/rust/pull/128449
[2]: https://github.com/rust-lang/rust/issues/128287 | diff --git a/src/civil/date.rs b/src/civil/date.rs
index 79d84fc..ee56cbe 100644
--- a/src/civil/date.rs
+++ b/src/civil/date.rs
@@ -1367,8 +1367,6 @@ impl Date {
/// # Example: negative spans are supported
///
/// ```
- /// # // See: https://github.com/rust-lang/rust/pull/121364
- /// # #![allow(u... | 8 | 2 | 120 |
abae98ea88c8d699ef466a49d0318856bc2c25f1 | Andrew Gallant | 2024-08-14T23:48:08 | cargo: use 'dep' for optional dependencies
This was an oversight. The 'js-sys' and 'wasm-bindgen' dependencies
only be referenced as dependencies. They aren't features themselves.
(And they were never documented as such, so this change is justifiable
in a semver compatible release.) | diff --git a/Cargo.toml b/Cargo.toml
index 026f0ba..5c419d7 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -72,7 +72,7 @@ tzdb-zoneinfo = ["std"]
# on Jiff directly and enable the `js` feature themselves.
#
# (This is the same dependency setup that the `getrandom` crate uses.)
-js = ["wasm-bindgen", "js-sys"]
+js = ["d... | 1 | 1 | 1 |
7da2731211b3f51b9f9c251f8d9a91d58321ec6c | Andrew Gallant | 2024-08-14T23:47:55 | util/rangeint: squash new warnings on nightly | diff --git a/src/util/rangeint.rs b/src/util/rangeint.rs
index 39c6492..47ae0a0 100644
--- a/src/util/rangeint.rs
+++ b/src/util/rangeint.rs
@@ -113,6 +113,7 @@ macro_rules! define_ranged {
val: impl Into<i64>,
) -> Result<Self, Error> {
let val = val.into();
+ ... | 1 | 3 | 0 |
c659069a2765a5e28ba503ea37e8a2eee4b10a92 | Andrew Gallant | 2024-08-11T12:24:40 | span: improve invalid units error message
Previously, we would report a range error when trying to use units of
days or greater in a span when adding to a `Timestamp`. But this error
doesn't really do a good job of explaining anything. It was a hold-over
from the days where I was using many different error types and i... | diff --git a/src/span.rs b/src/span.rs
index 62aba66..52d1ce9 100644
--- a/src/span.rs
+++ b/src/span.rs
@@ -2722,17 +2722,22 @@ impl Span {
pub(crate) fn smallest_non_time_non_zero_unit_error(
&self,
) -> Option<Error> {
- if self.days != 0 {
- Some(t::SpanDays::error("days", self.... | 1 | 12 | 7 |
224eee99a04ddaec2b6da98acdaa35733c51701e | Nuutti Kotivuori | 2024-08-08T13:42:59 | util: days in month rangeint should always have range 28-31
PR #85 | diff --git a/src/util/common.rs b/src/util/common.rs
index 55fb207..e6ef64c 100644
--- a/src/util/common.rs
+++ b/src/util/common.rs
@@ -62,7 +62,7 @@ pub(crate) fn saturate_day_in_month(
/// February.
pub(crate) fn days_in_month(year: Year, month: Month) -> Day {
if month == 2 && is_leap_year(year) {
- D... | 1 | 14 | 14 |
60ba273fa1232ef859d253ab3b284ca6a0ddb50f | Andrew Gallant | 2024-08-03T20:27:27 | fmt/strtime: remove note that IANA time zone identifiers aren't supported
Because they are supported now, as of #75. | diff --git a/src/fmt/strtime/mod.rs b/src/fmt/strtime/mod.rs
index a29399f..3bee55f 100644
--- a/src/fmt/strtime/mod.rs
+++ b/src/fmt/strtime/mod.rs
@@ -132,11 +132,6 @@ assert_eq!(date.strftime("%Y%m%d").to_string(), "09990715");
# Ok::<(), Box<dyn std::error::Error>>(())
```
-Additionally, the parsing and formatt... | 1 | 0 | 5 |
38bc73ad476e974de2563b0d898a32ccaf220b30 | Andrew Gallant | 2024-08-03T14:05:01 | fmt: migrate other fractional printing to FractionalFormatter
This migrates everything else over. This also explains why we were able
to previously get away with hacking the DecimalFormatter to emit
fractions: every other use of FractionalFormatter doesn't care about
setting a fixed precision. Although we may want to ... | diff --git a/src/fmt/offset.rs b/src/fmt/offset.rs
index 3b15b76..96da13b 100644
--- a/src/fmt/offset.rs
+++ b/src/fmt/offset.rs
@@ -106,7 +106,7 @@ from [Temporal's hybrid grammar].
use crate::{
error::{err, Error, ErrorContext},
fmt::{
- util::{parse_temporal_fraction, DecimalFormatter},
+ ut... | 3 | 11 | 60 |
a23300af39890340f26128acde6473f783793948 | Andrew Gallant | 2024-08-03T13:45:25 | fmt: add FractionalFormatter
Previously, we were formatting fractional components of
numbers by kinda abusing the DecimalFormatter. But in
order to fix #73, I found it simpler to split the
fractional case out of the "normal" number formatting.
This commit just adds the fractional formatter. Subsequent
commits will mi... | diff --git a/src/fmt/mod.rs b/src/fmt/mod.rs
index c0d610d..a0a185a 100644
--- a/src/fmt/mod.rs
+++ b/src/fmt/mod.rs
@@ -14,7 +14,7 @@ use crate::{
util::escape,
};
-use self::util::{Decimal, DecimalFormatter};
+use self::util::{Decimal, DecimalFormatter, Fractional, FractionalFormatter};
mod offset;
pub mo... | 2 | 177 | 1 |
c1b1ca433ac604883a88f7077dff860ae1094c3e | David Peter | 2024-08-02T20:52:16 | doc: minor correction: "printf"-style API
PR #71 | diff --git a/src/civil/date.rs b/src/civil/date.rs
index ed63536..f1b8590 100644
--- a/src/civil/date.rs
+++ b/src/civil/date.rs
@@ -2075,11 +2075,11 @@ impl Date {
}
}
-/// Parsing and formatting using a "printf" style of API.
+/// Parsing and formatting using a "printf"-style API.
impl Date {
/// Parses... | 6 | 16 | 16 |
019696cf61c608a783838400732a947778eecb8e | Andrew Gallant | 2024-08-02T00:15:11 | fmt/strtime: make parsing aware of padding settings
In order to make #62 work, we need some way of "limiting" the parsing
of numbers. So for example, when parsing a year, we shouldn't try to
parse more than 4 digits. And when parsing a month, we shouldn't try to
parse more than 2 digits.
This isn't a big deal, except... | diff --git a/src/fmt/strtime/mod.rs b/src/fmt/strtime/mod.rs
index ba44572..257566b 100644
--- a/src/fmt/strtime/mod.rs
+++ b/src/fmt/strtime/mod.rs
@@ -75,42 +75,50 @@ assert_eq!(
# Advice
The formatting machinery supported by this module is not especially expressive.
-The pattern language is a simple sequence of ... | 2 | 243 | 110 |
5020f75ff8f1176238d2b1d7d00a631abb6083f7 | Andrew Gallant | 2024-08-01T20:49:31 | fmt/strtime: refactor to enable Extension parsing in strptime
Previously, in strptime, we just skipped over ASCII digits and flags
when they might appear. But in the course of trying to fix #62, it
looks like we're going to want to have parsing respect padding
settings. So this commit refactors parsing to actually ext... | diff --git a/src/fmt/strtime/format.rs b/src/fmt/strtime/format.rs
index 391163d..49f770e 100644
--- a/src/fmt/strtime/format.rs
+++ b/src/fmt/strtime/format.rs
@@ -5,7 +5,7 @@ use crate::{
fmt::{
strtime::{
month_name_abbrev, month_name_full, weekday_name_abbrev,
- weekday_name_fu... | 3 | 113 | 109 |
34ef70c2a3d382d31e2c5f4ccf48109df09605be | Andrew Gallant | 2024-07-27T12:15:12 | span: improve docs for Span getters
I noticed that at least [one person] was confused about what it should
return. In particular, the getters on a `Span` return the individual
units, and not the "total" amount of them in the entire span. This is
different from how "absolute" durations work that don't keep track of
the... | diff --git a/src/span.rs b/src/span.rs
index e990401..62aba66 100644
--- a/src/span.rs
+++ b/src/span.rs
@@ -933,61 +933,223 @@ impl Span {
/// Routines for accessing the individual units in a `Span`.
impl Span {
- /// Returns the number of years in this span.
+ /// Returns the number of year units in this sp... | 1 | 172 | 10 |
a8cf3bea6e0aa923004b3aaaa6d7867557b35978 | Andrew Gallant | 2024-07-27T12:29:59 | span: allow `ambiguous_negative_literals`
This was bad timing. The lang team just stabilized (in nightly) a new
deny-by-default lint, named `ambiguous_negative_literals`, which
triggers an error for things like `-1.hour()`. While such things can be
confusingly ambiguous in some cases, in Jiff, `-1.hour()`, `(-1).hour(... | diff --git a/src/civil/date.rs b/src/civil/date.rs
index 47ef6a7..ed63536 100644
--- a/src/civil/date.rs
+++ b/src/civil/date.rs
@@ -1367,6 +1367,8 @@ impl Date {
/// # Example: negative spans are supported
///
/// ```
+ /// # // See: https://github.com/rust-lang/rust/pull/121364
+ /// # #![allow(u... | 9 | 130 | 0 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.