Merge branch 'main' into bidi

Author: robertbastian

CHANGELOG.md

@@ -37,6 +37,9 @@
     - `zerofrom`
     - `zerotrie`
     - `zerovec`
+      - This release has multiple changes that affect the bit representation of various types. Do not update to this release if you wish to retain stable data formats.
+        - Change the `VarZeroVecFormat` values shipped by default to use the same index and length width. This breaks data layout for all `VarZeroVec`s. (https://github.com/unicode-org/icu4x/pull/5594)
+        - Optimize `MultiFieldsULE` to not store a length anymore. This breaks data layout for any `#[make_varule]`-using struct with multiple variable-sized fields. (https://github.com/unicode-org/icu4x/pull/5593)
     - `writeable`
 
 

CODEOWNERS

@@ -11,7 +11,7 @@ components/calendar/              @Manishearth @sffc
 components/casemap/               @Manishearth
 components/collator/              @hsivonen @echeran
 components/collections/           @echeran
-components/datetime/              @zbraniecki @nordzilla
+components/datetime/              @sffc @zbraniecki
 components/decimal/               @sffc
 components/experimental/src/compactdecimal/      @eggrobin
 components/experimental/src/dimension/           @younies
@@ -24,8 +24,8 @@ components/locale_core/           @zbraniecki @nciric
 components/normalizer/            @hsivonen @echeran
 components/plurals/               @zbraniecki @sffc
 components/segmenter/             @aethanyc @makotokato @sffc
-components/timezone/              @nordzilla
-tutorials/gn/                @sffc
+components/timezone/              @nekevss @robertbastian @sffc
+tutorials/gn/                     @sffc
 ffi/capi/                         @Manishearth
 ffi/ecma402/                      @filmil
 ffi/harfbuzz/                     @hsivonen
@@ -39,13 +39,15 @@ tools/benchmark/binsize/          @gnrunge
 tools/depcheck/                   @Manishearth
 tools/depcheck/src/allowlist.rs   @unicode-org/icu4x-owners
 tools/make/                       @Manishearth @sffc
-utils/bies/                @sffc
+utils/bies/                       @sffc
 utils/databake                    @robertbastian @Manishearth
 utils/deduplicating_array         @robertbastian
 utils/fixed_decimal/              @sffc @younies
+utils/izdtf                       @nekevss @sffc
 utils/litemap/                    @Manishearth @sffc
 utils/pattern/                    @zbraniecki
 utils/tinystr                     @zbraniecki @sffc
+utils/tzif                        @nordzilla
 utils/writeable/                  @sffc @robertbastian
 utils/yoke/                       @Manishearth @sffc
 utils/zerofrom                    @Manishearth @sffc

Cargo.lock

@@ -587,10 +587,10 @@ pub(crate) struct RawDateTimeNamesBorrowed<'l> {
     dayperiod_names: OptionalNames<FieldLength, &'l LinearNamesV1<'l>>,
     zone_essentials: OptionalNames<(), &'l tz::EssentialsV1<'l>>,
     exemplar_cities: OptionalNames<(), &'l tz::ExemplarCitiesV1<'l>>,
-    mz_generic_long: OptionalNames<(), &'l tz::MzGenericLongV1<'l>>,
-    mz_generic_short: OptionalNames<(), &'l tz::MzGenericShortV1<'l>>,
-    mz_specific_long: OptionalNames<(), &'l tz::MzSpecificLongV1<'l>>,
-    mz_specific_short: OptionalNames<(), &'l tz::MzSpecificShortV1<'l>>,
+    mz_generic_long: OptionalNames<(), &'l tz::MzGenericV1<'l>>,
+    mz_generic_short: OptionalNames<(), &'l tz::MzGenericV1<'l>>,
+    mz_specific_long: OptionalNames<(), &'l tz::MzSpecificV1<'l>>,
+    mz_specific_short: OptionalNames<(), &'l tz::MzSpecificV1<'l>>,
     pub(crate) fixed_decimal_formatter: Option<&'l FixedDecimalFormatter>,
     pub(crate) week_calculator: Option<&'l WeekCalculator>,
 }

components/datetime/src/format/neo.rs

@@ -1017,13 +1017,13 @@ pub trait ZoneMarkers: private::Sealed {
     /// Marker for loading exemplar city names for time zone formatting
     type ExemplarCitiesV1Marker: DataMarker<DataStruct = tz::ExemplarCitiesV1<'static>>;
     /// Marker for loading generic short time zone names.
-    type GenericLongV1Marker: DataMarker<DataStruct = tz::MzGenericLongV1<'static>>;
+    type GenericLongV1Marker: DataMarker<DataStruct = tz::MzGenericV1<'static>>;
     /// Marker for loading generic short time zone names.
-    type GenericShortV1Marker: DataMarker<DataStruct = tz::MzGenericShortV1<'static>>;
+    type GenericShortV1Marker: DataMarker<DataStruct = tz::MzGenericV1<'static>>;
     /// Marker for loading generic short time zone names.
-    type SpecificLongV1Marker: DataMarker<DataStruct = tz::MzSpecificLongV1<'static>>;
+    type SpecificLongV1Marker: DataMarker<DataStruct = tz::MzSpecificV1<'static>>;
     /// Marker for loading generic short time zone names.
-    type SpecificShortV1Marker: DataMarker<DataStruct = tz::MzSpecificShortV1<'static>>;
+    type SpecificShortV1Marker: DataMarker<DataStruct = tz::MzSpecificV1<'static>>;
 }
 
 /// A trait associating constants and types implementing various other traits
@@ -1145,10 +1145,10 @@ impl ZoneMarkers for NeoNeverMarker {
     type TimeZoneVariantInput = NeverField;
     type EssentialsV1Marker = NeverMarker<tz::EssentialsV1<'static>>;
     type ExemplarCitiesV1Marker = NeverMarker<tz::ExemplarCitiesV1<'static>>;
-    type GenericLongV1Marker = NeverMarker<tz::MzGenericLongV1<'static>>;
-    type GenericShortV1Marker = NeverMarker<tz::MzGenericShortV1<'static>>;
-    type SpecificLongV1Marker = NeverMarker<tz::MzSpecificLongV1<'static>>;
-    type SpecificShortV1Marker = NeverMarker<tz::MzSpecificShortV1<'static>>;
+    type GenericLongV1Marker = NeverMarker<tz::MzGenericV1<'static>>;
+    type GenericShortV1Marker = NeverMarker<tz::MzGenericV1<'static>>;
+    type SpecificLongV1Marker = NeverMarker<tz::MzSpecificV1<'static>>;
+    type SpecificShortV1Marker = NeverMarker<tz::MzSpecificV1<'static>>;
 }
 
 /// A struct that supports formatting both a date and a time.
@@ -1504,16 +1504,16 @@ macro_rules! datetime_marker_helper {
         NeverMarker<tz::ExemplarCitiesV1<'static>>
     };
     (@data/zone/generic_long,) => {
-        NeverMarker<tz::MzGenericLongV1<'static>>
+        NeverMarker<tz::MzGenericV1<'static>>
     };
     (@data/zone/generic_short,) => {
-        NeverMarker<tz::MzGenericShortV1<'static>>
+        NeverMarker<tz::MzGenericV1<'static>>
     };
     (@data/zone/specific_long,) => {
-        NeverMarker<tz::MzSpecificLongV1<'static>>
+        NeverMarker<tz::MzSpecificV1<'static>>
     };
     (@data/zone/specific_short,) => {
-        NeverMarker<tz::MzSpecificShortV1<'static>>
+        NeverMarker<tz::MzSpecificV1<'static>>
     };
     (@names/year, yes) => {
         YearNamesV1Marker

components/datetime/src/neo_marker.rs

@@ -7,6 +7,7 @@
 use super::super::{reference, PatternError, PatternItem, TimeGranularity};
 use alloc::vec::Vec;
 use core::str::FromStr;
+use icu_plurals::provider::FourBitMetadata;
 use icu_provider::prelude::*;
 use zerovec::{ule::AsULE, ZeroSlice, ZeroVec};
 
@@ -75,6 +76,15 @@ impl PatternMetadata {
     pub(crate) fn set_time_granularity(&mut self, time_granularity: TimeGranularity) {
         self.0 = time_granularity.ordinal();
     }
+
+    pub(crate) fn to_four_bit_metadata(self) -> FourBitMetadata {
+        #[allow(clippy::unwrap_used)] // valid values for self.0 are 0, 1, 2, 3, or 4
+        FourBitMetadata::try_from_byte(self.0).unwrap()
+    }
+
+    pub(crate) fn from_u8(other: u8) -> Self {
+        Self(TimeGranularity::from_ordinal(other).ordinal())
+    }
 }
 
 impl Default for PatternMetadata {

components/datetime/src/pattern/runtime/pattern.rs

@@ -7,28 +7,18 @@ use crate::{
     skeleton::{reference::Skeleton, SkeletonError},
 };
 use core::convert::TryFrom;
-use icu_provider::prelude::*;
 use litemap::LiteMap;
 
-size_test!(DateSkeletonPatternsV1, date_skeleton_patterns_v1_size, 24);
-
 /// Skeleton data for dates and times, along with the corresponding plural pattern
 /// information.
-#[doc = date_skeleton_patterns_v1_size!()]
 ///
 /// <div class="stab unstable">
 /// 🚧 This code is considered unstable; it may change at any time, in breaking or non-breaking ways,
 /// including in SemVer minor releases. While the serde representation of data structs is guaranteed
 /// to be stable, their Rust representation might not be. Use with caution.
 /// </div>
-#[icu_provider::data_struct(DateSkeletonPatternsV1Marker = "datetime/skeletons@1")]
 #[derive(Debug, PartialEq, Clone, Default)]
-#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
-pub struct DateSkeletonPatternsV1<'data>(
-    #[cfg_attr(feature = "serde", serde(borrow))]
-    #[zerofrom(clone)]
-    pub LiteMap<SkeletonV1, PatternPlurals<'data>>,
-);
+pub struct DateSkeletonPatternsV1<'data>(pub LiteMap<SkeletonV1, PatternPlurals<'data>>);
 
 /// This struct is a public wrapper around the internal `Skeleton` struct. This allows
 /// access to the serialization and deserialization capabilities, without exposing the
@@ -43,7 +33,6 @@ pub struct DateSkeletonPatternsV1<'data>(
 /// to be stable, their Rust representation might not be. Use with caution.
 /// </div>
 #[derive(Debug, PartialEq, Eq, PartialOrd, Ord, Clone)]
-#[cfg_attr(feature = "serde", derive(serde::Deserialize))]
 pub struct SkeletonV1(pub Skeleton);
 
 impl TryFrom<&str> for SkeletonV1 {

components/datetime/src/provider/calendar/skeletons.rs

@@ -17,6 +17,7 @@
 
 pub mod calendar;
 pub(crate) mod date_time;
+pub mod packed_pattern;
 pub mod time_zones;
 
 /// Module for new DateSymbols design