Merge branch 'main' into uts46bis

CHANGELOG.md

@@ -10,6 +10,7 @@
   - `icu_calendar`
     - New `DateTime::local_unix_epoch()` convenience constructor (https://github.com/unicode-org/icu4x/pull/4479)
     - Improved approximation for Persian calendrical calculations (https://github.com/unicode-org/icu4x/issues/4713)
+    - Fix weekday calculations in negative ISO years (https://github.com/unicode-org/icu4x/pull/4894)
   - `icu_datetime`
     - `FormattedDateTime` and `FormattedZonedDateTime` now implement `Clone` and `Copy` (https://github.com/unicode-org/icu4x/pull/4476)
   - `icu_locid`
@@ -20,6 +21,7 @@
     - Add `Aran` script code (https://github.com/unicode-org/icu4x/pull/4426)
     - Mark additional constructors as `const` (https://github.com/unicode-org/icu4x/pull/4584, https://github.com/unicode-org/icu4x/pull/4574)
     - Implement Joining_Type property (https://github.com/unicode-org/icu4x/pull/4599)
+    - Implement Hangul_Syllable_Type property (https://github.com/unicode-org/icu4x/pull/4885)
   - `icu_segmenter`
     - Fix Unicode 15.0 line breaking (https://github.com/unicode-org/icu4x/pull/4389)
 - Data model and providers
@@ -38,6 +40,7 @@
     - All languages
       - Correctly handle invalid UTF8 (https://github.com/unicode-org/icu4x/pull/4353)
       - Correctly handle chars (https://github.com/unicode-org/icu4x/pull/4349)
+      - Add `day_of_year` getter (https://github.com/unicode-org/icu4x/issues/4891)
     - JS
       - Fixed a bug where slice length is computed incorrectly (https://github.com/rust-diplomat/diplomat/pull/372)
       - Changed file extensions for better compatibility with nodejs modules (https://github.com/rust-diplomat/diplomat/pull/387)
@@ -61,6 +64,7 @@
     - `zerotrie`
         - Add `as_borrowed_slice` and `AsRef` impl (https://github.com/unicode-org/icu4x/pull/4381)
         - Add `ZeroTrieSimpleAsciiCursor` for manual iteration (https://github.com/unicode-org/icu4x/pull/4383)
+        - Increase bound of `p` to solve more perfect hash functions; _might_ break serialized ZeroTriePerfectHash from previous versions (https://github.com/unicode-org/icu4x/pull/4888)
     - `zerovec`
         - Change `ZeroHashMap` to use `twox-hash` (https://github.com/unicode-org/icu4x/pull/4592)
     - `writeable`

Cargo.toml

@@ -209,65 +209,87 @@ diplomat-runtime = { git = "https://github.com/rust-diplomat/diplomat.git", rev
 diplomat_core = { git = "https://github.com/rust-diplomat/diplomat.git", rev = "3167e39e89ae612233706f1ade3d9e8a5c6ad358" }
 diplomat-tool = { git = "https://github.com/rust-diplomat/diplomat.git", rev = "3167e39e89ae612233706f1ade3d9e8a5c6ad358" }
 
-# External deps from crates.io
-bincode = { version = "1.3.1", default-features = false }
-clap = { version = "4.2.0", default-features = false }
-combine = { version = "4.3.1", default-features = false }
+# EXTERNAL DEPENDENCIES
+#
+# Due to semver-related constraints, we generally want to consume crates.io dependencies with
+# default features, with some notable exceptions:
+#
+# 1. Dependencies of core library crates need to be `no_std`
+# 2. Some crates have a lot of default features that pull in undesirable transitive deps
+#
+# We therefore have three groups for our crates.io dependencies.
+
+## External Deps Group 1: Library deps except datagen. No default features.
+# Keep in sync with allowlist.rs:
+# BASIC_RUNTIME_DEPS
+# BASIC_BUILD_DEPS
+# EXTRA_SERDE_DEPS
+# EXTRA_EXPERIMENTAL_DEPS
+# EXTRA_LSTM_DEPS
+# EXTRA_RYU_DEPS
+# EXTRA_CAPI_DEPS
+# EXTRA_BLOB_DEPS
+# EXTRA_FS_DEPS
 core_maths = { version = "0.1.0", default-features = false }
 displaydoc = { version = "0.2.3", default-features = false }
 either = { version = "1.9.0", default-features = false }
-elsa = { version = "1.10.0", default-features = false }
-erased-serde = { version = "0.3.11", default-features = false }
-eyre = { version = "0.6.0", default-features = false }
-indexmap = { version = "1.0.0", default-features = false }
-itertools = { version = "0.10.3", default-features = false}
 libc_alloc = { version = "1.0.6", default-features = false }
 log = { version = "0.4.17", default-features = false }
 memchr = { version = "2.6.0", default-features = false }
-ndarray = { version = "0.15.0", default-features = false }
-nom = { version = "7.0.0", default-features = false }
 num-bigint = { version = "0.4.3", default-features = false }
 num-rational = { version = "0.4.0", default-features = false }
 num-traits = { version = "0.2.0", default-features = false }
-once_cell = {version = "1.15.0", default-features = false}
 postcard = { version = "1.0.1", default-features = false }
-proc-macro2 = { version = "1.0.61", default-features = false }
-quote = { version = "1.0.28", default-features = false }
-rayon = { version = "1.3.0", default-features = false }
 regex-automata = { version = "0.2.0", default-features = false }
 ryu = { version = "1.0.5", default-features = false }
 serde = { version = "1.0.110", default-features = false }
-serde_json = { version = "1.0.45", default-features = false, features = ["alloc"] } # requires either alloc or std to build
-serde-aux = { version = "4.0.0", default-features = false }
 serde-json-core = { version = "0.4.0", default-features = false }
-simple_logger = { version = "4.0.0", default-features = false }
 smallvec = { version = "1.10.0", default-features = false }
 stable_deref_trait = { version = "1.2.0", default-features = false }
-syn = { version = "2.0.21", default-features = false }
-synstructure = { version = "0.13.0", default-features = false }
-toml = { version = "0.5.8", default-features = false }
-twox-hash = { version = "1.4.2", default-features = false }
 unicode-bidi = { version = "0.3.11", default-features = false }
-ureq = { version = "2.3.0", default-features = false }
 utf16_iter = { version = "1.0.2", default-features = false }
 utf8_iter = { version = "1.0.2", default-features = false }
-wasmi = { version = "0.31.2", default-features = false }
 write16 = { version = "1.0.0", default-features = false }
+
+## External Deps Group 2: Heavy Dev and Datagen deps. No default features.
 zip = { version = "0.6.0", default-features = false }
+serde-aux = { version = "4.0.0", default-features = false }
 
-## Dev only
+## External Deps Group 3: Dev and Datagen deps. Include default features.
 arraystring = "0.3.0"
 arrayvec = "0.7.2"
 atoi = "1.0.0"
+bincode = "1.3.1"
+clap = "4.2.0"
+combine = "4.3.1"
 criterion = "0.5.0"
 detone = "1.0.0"
+elsa = "1.10.0"
+erased-serde = "0.3.11"
+eyre = "0.6.0"
 getrandom = "0.2"
 iai = "0.1.1"
+indexmap = "1.0.0"
+itertools = "0.10.3"
+ndarray = "0.15.0"
+nom = "7.0.0"
+once_cell = "1.15.0"
+proc-macro2 = "1.0.61"
+quote = "1.0.28"
 rand = "0.8"
 rand_distr = "0.4"
 rand_pcg = "0.3"
+rayon = "1.3.0"
 rkyv = "0.7"
+serde_json = "1.0.45"
+simple_logger = "4.0.0"
+syn = "2.0.21"
+synstructure = "0.13.0"
+toml = "0.5.8"
+twox-hash = "1.4.2"
+ureq = "2.3.0"
 walkdir = "2.3.2"
+wasmi = "0.31.2"
 
 # Size optimized builds
 [profile.release-opt-size]

components/calendar/src/any_calendar.rs

@@ -48,7 +48,7 @@ use core::fmt;
 ///
 /// let calendar = AnyCalendar::new_for_locale(&locale.into());
 /// let calendar = Rc::new(calendar); // Avoid cloning it each time
-///                                   // If everything is a local reference, you may use icu_calendar::Ref instead.
+///                                   // If everything is a local reference, you may use icu::calendar::Ref instead.
 ///
 /// // manually construct a datetime in this calendar
 /// let manual_time = Time::try_new(12, 33, 12, 0).expect("failed to construct Time");

components/calendar/src/calendar.rs

@@ -88,7 +88,7 @@ pub trait Calendar {
     fn day_of_year_info(&self, date: &Self::DateInner) -> types::DayOfYearInfo;
 
     /// The [`AnyCalendarKind`] corresponding to this calendar,
-    /// if one exists. Implementors outside of `icu_calendar` should return `None`
+    /// if one exists. Implementors outside of `icu::calendar` should return `None`
     fn any_calendar_kind(&self) -> Option<AnyCalendarKind> {
         None
     }

components/calendar/src/duration.rs

@@ -13,7 +13,7 @@ use core::marker::PhantomData;
 /// # Example
 ///
 /// ```rust
-/// use icu_calendar::{
+/// use icu::calendar::{
 ///     types::IsoWeekday, Date, DateDuration, DateDurationUnit,
 /// };
 ///
@@ -114,7 +114,7 @@ impl<C: Calendar + ?Sized> DateDuration<C> {
     /// Construct a DateDuration
     ///
     /// ```rust
-    /// # use icu_calendar::*;
+    /// # use icu::calendar::*;
     /// // two years, three months, and five days
     /// let duration: DateDuration<Iso> = DateDuration::new(2, 3, 0, 5);
     /// ```

components/calendar/src/error.rs

@@ -76,8 +76,8 @@ impl CalendarError {
     /// # Examples
     ///
     /// ```
-    /// use icu_calendar::AnyCalendarKind;
-    /// use icu_calendar::CalendarError;
+    /// use icu::calendar::AnyCalendarKind;
+    /// use icu::calendar::CalendarError;
     ///
     /// let cal_str = "maori";
     ///

components/calendar/src/hebrew.rs

@@ -367,7 +367,7 @@ impl Date<Hebrew> {
     ///
     ///
     /// ```rust
-    /// use icu_calendar::Date;
+    /// use icu::calendar::Date;
     ///
     /// let date_hebrew =
     ///     Date::try_new_hebrew_date(3425, 4, 25)
@@ -470,7 +470,7 @@ impl<A: AsCalendar<Calendar = Hebrew>> DateTime<A> {
 mod tests {
 
     use super::*;
-    use crate::types::MonthCode;
+    use crate::types::{Era, MonthCode};
     use calendrical_calculations::hebrew_keviyah::*;
 
     // Sentinel value for Adar I
@@ -587,4 +587,17 @@ mod tests {
         let yi = YearInfo::compute_for(88369);
         assert_eq!(yi.keviyah.year_length(), 383);
     }
+
+    #[test]
+    fn test_weekdays() {
+        // https://github.com/unicode-org/icu4x/issues/4893
+        let cal = Hebrew::new();
+        let era = "am".parse::<Era>().unwrap();
+        let month_code = "M01".parse::<MonthCode>().unwrap();
+        let dt = cal.date_from_codes(era, 3760, month_code, 1).unwrap();
+
+        // Should be Saturday per:
+        // https://www.hebcal.com/converter?hd=1&hm=Tishrei&hy=3760&h2g=1
+        assert_eq!(6, cal.day_of_week(&dt) as usize);
+    }
 }

components/calendar/src/iso.rs

@@ -135,7 +135,9 @@ impl Calendar for Iso {
 
         // The days of the week are the same every 400 years
         // so we normalize to the nearest multiple of 400
-        let years_since_400 = date.0.year % 400;
+        let years_since_400 = date.0.year.rem_euclid(400);
+        debug_assert!(years_since_400 >= 0); // rem_euclid returns positive numbers
+        let years_since_400 = years_since_400 as u32;
         let leap_years_since_400 = years_since_400 / 4 - years_since_400 / 100;
         // The number of days to the current year
         // Can never cause an overflow because years_since_400 has a maximum value of 399.
@@ -169,7 +171,7 @@ impl Calendar for Iso {
             }
         };
         let january_1_2000 = 5; // Saturday
-        let day_offset = (january_1_2000 + year_offset + month_offset + date.0.day as i32) % 7;
+        let day_offset = (january_1_2000 + year_offset + month_offset + date.0.day as u32) % 7;
 
         // We calculated in a zero-indexed fashion, but ISO specifies one-indexed
         types::IsoWeekday::from((day_offset + 1) as usize)
@@ -832,4 +834,49 @@ mod test {
         check(-1439, 1969, 12, 31, 0, 1);
         check(-2879, 1969, 12, 30, 0, 1);
     }
+
+    #[test]
+    fn test_continuity_near_year_zero() {
+        // https://github.com/unicode-org/icu4x/issues/4893
+        const ONE_DAY_DURATION: DateDuration<Iso> = DateDuration {
+            years: 0,
+            months: 0,
+            weeks: 0,
+            days: 1,
+            marker: core::marker::PhantomData,
+        };
+
+        let mut iso_date = Date::try_new_iso_date(-10, 1, 1).unwrap();
+        let mut rata_die = iso_date.to_fixed();
+        let mut weekday = iso_date.day_of_week();
+
+        for _ in 0..(366 * 20) {
+            let next_iso_date = iso_date.added(ONE_DAY_DURATION);
+            let next_rata_die = next_iso_date.to_fixed();
+            assert_eq!(
+                next_rata_die,
+                rata_die + 1,
+                "{iso_date:?}..{next_iso_date:?}"
+            );
+            let next_weekday = next_iso_date.day_of_week();
+            assert_eq!(
+                (next_weekday as usize) % 7,
+                (weekday as usize + 1) % 7,
+                "{iso_date:?}..{next_iso_date:?}"
+            );
+            if iso_date.month().code.parsed().unwrap().0 == 2 && iso_date.day_of_month().0 == 28 {
+                assert_eq!(next_iso_date.is_in_leap_year(), iso_date.is_in_leap_year());
+                if iso_date.is_in_leap_year() {
+                    assert_eq!(next_iso_date.month().code.parsed().unwrap().0, 2);
+                    assert_eq!(next_iso_date.day_of_month().0, 29);
+                } else {
+                    assert_eq!(next_iso_date.month().code.parsed().unwrap().0, 3);
+                    assert_eq!(next_iso_date.day_of_month().0, 1);
+                }
+            }
+            iso_date = next_iso_date;
+            rata_die = next_rata_die;
+            weekday = next_weekday;
+        }
+    }
 }

components/calendar/src/japanese.rs

@@ -509,7 +509,7 @@ impl Date<JapaneseExtended> {
         Ok(Date::from_raw(inner, japanext_calendar))
     }
 
-    /// For testing era fallback in icu_datetime
+    // For testing era fallback in icu_datetime
     #[doc(hidden)]
     pub fn into_japanese_date(self) -> Date<Japanese> {
         Date::from_raw(self.inner, self.calendar.0)