v0.4.1

Other

• implement h11/h12/h23/h24 hour cycles (0-vs-1 origin)
• derive default hour cycle from locale data; keep weekday in date combos

v0.4.0

Other

• rustfmt the ECMA-402 number/datetime code + codegen emit_dates
• ECMA-402 DateTimeFormat options + formatToParts
• ECMA-402 NumberFormat options + formatToParts
• add sub-second DateTime.millisecond field
• raw-source Gregorian date data from official CLDR v48
• linear-time connector folding (quadratic)
• truncate normalization buffer on char boundary (wrong-fallback)
• validate POSIX TZ rule field ranges (reject malformed M-rules)
• document that to_unicode output is unvalidated (not for security decisions)
• bounds-check names.bin parser (panic on malformed blob)
• bound push_pend against MAX_DECOMP (latent OOB)
• wrapping_add for regional-indicator parity counters (debug overflow)
• guard forward conversions against i64 overflow (panic on extreme inputs)
• bound find/contains unaligned-start scan (quadratic DoS)

v0.3.1

Other

• use unsigned_abs in format_gmt_offset (i32::MIN panic)
• guard empty separator in parse_decimal (latent hang) and saturating compact width
• cap U-label code points before Punycode encode (encoder DoS)
• memoize SB8 sentence lookahead (quadratic DoS)
• bound window_decision scan (quadratic find/contains DoS)

v0.3.0

Other

• mark Error #[non_exhaustive]
• declare bidi feature dependency (fixes component-subsets CI)
• enforce VerifyDnsLength (reject empty/trailing-root labels)
• apply UTS #39 augmented (CJK) script sets in is_single_script
• satisfy CI gates for security fixes (fmt, clippy, public-api snapshot)
• resolve is_single_script via Script_Extensions intersection (UTS #39)
• IDNA tests: adapt to fallible to_unicode; raise conformance rejection bar to 480
• enforce full IDNA2008 validity (CheckBidi/ContextJ/Hyphens/V1/V5/V6) and re-canonicalize xn--
• do not drop combining marks past MAX_COMBINING in NFC recomposition
• O(1) matched-PDI lookup (latent quadratic)
• linear find/contains (quadratic-scan DoS)
• avoid O(n) Vec::remove in discontiguous matching
• strip Default_Ignorable code points in UTS #39 skeleton (confusable-detection miss)
• cap MessageFormat parse recursion depth (stack-overflow DoS)
• use checked arithmetic in POSIX TZ offset parsing (i32 overflow panic/wrap)
• cap punycode decode output (decompression-bomb DoS) and enforce 253-octet domain limit

v0.2.3

Other

• reject malformed labels; make the conformance test honest
• revert the eos-skip hack; add exhaustive BidiTest.txt conformance
• gate out CLDR locales the parser mis-orders (correctness fix)
• Remove ROADMAP.md (every item complete)

v0.2.2

Other

• Collation for_locale: data-driven from official CLDR rules
• Collation tailoring: unbounded weight allocation (pair-encoded primaries)
• transliteration component delivered (mark done)
• profiling addressed via benchmarks + size tracking
• add Armenian + Georgian romanizations
• Document collation tailoring capacity (gap-insertion limit)
• add alphabetic index (index_labels / index_bucket)
• Format spellout_ordinal regression test (rustfmt)
• Fix RBNF stack overflow / runaway on adversarial input
• add primary-strength string search (find/contains)
• add ordinal spell-out (spell_ordinal)

v0.2.1

Other

• use then_some over then(||...) (clippy)
• set quantifiers ([..]+/*/?) and $0 match-reference
• Bundle more collation locales + fix 3-letter subtag resolution
• full conformance (91707/91707) via isolate-boundary eos fix
• Per-component Unicode features + iana-tz/full/names default (MSRV 1.86)
• Default to everything: full + alloc + names (opt out for size)
• Line breaking: full conformance (19338/19338) via LB15b EAW tailoring
• Bundle more locale collation tailorings (hu, ro, sq, uk, vi)
• add character-set sources ([abc x-z] > t)
• add before/after context (ICU b { src } a > tgt syntax)
• Add the full tabulated character Name database (names feature)
• line-break now 19335/19338 after LB21a fix
• Fix line-break LB21a: Hebrew-letter rule is HY-only, not HY|BA
• Add cargo-public-api surface guard (CI)
• multi-char targets (digraphs) + many more locale rules

v0.2.0

Other

• Add a rule-based Transform to transliteration (x > y rewrites)
• Add collation expansions to the tailoring engine (ä → "ae")
• document collation strength/numeric/tailoring
• Add Tailoring::for_locale for well-known collations
• handle <<, <<<, = relation levels (not just primary)
• Add locale-tailored collation engine (Tailoring, primary reordering)
• Add numeric collation (Collator::with_numeric, natural sort)
• Add translit::any_ascii convenience (mixed-script -> ASCII)
• Add collation strength levels (Collator::with_strength)
• Add Greek->Latin transliteration (ELOT 743 / ISO 843)
• Add Cyrillic->Latin transliteration (ISO 9)
• Add remove_diacritics to the transliteration module
• Add transliteration module: Latin-ASCII fold (translit::latin_ascii)
• Fix format_compact panic on non-finite values
• Add ordinal number formatting (number::format_ordinal)
• Parse the compact c/e plural operand (PluralOperands)
• Add cargo-fuzz harness (unicode + formatters targets)
• Add bidi mirroring (bidi_mirror + is_bidi_mirrored)
• Add Changes_When_* predicates (UAX #44)
• Fix doc-comment placement on lowercase_str_lang
• Add Lithuanian locale casing (retained dot above)
• Track compiled binary size per tier in CI (sizeprobe example)
• Gate differential test on the bmp tier
• Format differential test (rustfmt)
• Add differential tests against std (case mapping + predicates)
• reject the trie migration; keep the paged match tables
• Add Joining_Group property (UAX #9)
• Add char_name for algorithmically-named characters (Hangul + ideographs)
• Add algorithmic Hangul syllable names
• Add MSRV CI job (Rust 1.70); fix f64::abs not in core on MSRV
• Add Turkic (tr/az) locale-aware casing
• Fix codegen clippy: allow too_many_arguments on emit_value_enum
• Add context-sensitive lowercasing with Greek Final_Sigma
• Add Indic_Positional_Category; factor a value-enum codegen helper
• Add Indic_Syllabic_Category property (UAX #44)
• Add Joining_Type property (Arabic shaping, UAX #9)
• Add the Chinese (lunisolar) calendar
• Add duration formatting; correct stale CLDR/formatter roadmap status
• Add Age and Block character properties (UCD)
• Fix panics on unvalidated input; fuzz every input-facing API
• Add full IANA time-zone support via the timezone-data crate
• Add native digit systems to number formatting
• Add compact number notation (number::format_compact)
• Add scientific number notation (number::format_scientific)
• Add locale-aware number parsing (number::parse_decimal)
• Add locale-driven RBNF cardinal spell-out
• Add POSIX TZ time zones (intl::timezone)
• Update the crate-level docs to describe the full library
• Render localized Persian dates; share the non-Gregorian renderer
• Add date arithmetic to DateTime (add_seconds/add_days/weekday)
• Render localized Islamic (Hijri) dates
• report embedded-data footprint (Phase 2 size tracking)
• Add Hebrew calendar (Dershowitz-Reingold arithmetic)
• Add Persian (Solar Hijri) and Japanese calendars
• Add localized GMT offset formatting (CLDR time zones, data-light)
• Add ISO-8601 timestamp parse/format to DateTime
• Add calendar conversions (no_std): Gregorian, Islamic, ISO week
• Parse BCP-47 extension and private-use subtags in Locale
• Add locale negotiation (best-match)
• Add likely-subtags maximize/minimize to Locale (CLDR / UTS #35)
• Add flexible date skeleton formatting (CLDR availableFormats)
• Remove English-only spellout
• Add English number spell-out (RBNF)
• Add Gregorian date/time formatting (CLDR / UTS #35)
• Add ICU MessageFormat subset (intl::message)
• Add measurement-unit formatting (CLDR / UTS #35)
• Add locale display names (CLDR / UTS #35)
• Embed CLDR formatter tables as no_std binary blobs; add currency

v0.1.4

Other

• Add relative-time formatting + fix non-alloc tier builds
• Add locale-aware list formatting (CLDR / UTS #35)
• Add locale-aware number formatting (CLDR / UTS #35)
• Add CLDR ordinal plural rules
• Add CLDR cardinal plural rules (UTS #35)
• Implement the full bidi algorithm (UAX #9)
• Add common boolean properties (Math, Dash, Hex_Digit, …)
• Add BCP-47 locale parsing (Phase 3 foundation)
• build on a bare-metal no_std target
• Add Bidi_Class property + paragraph base direction (UAX #9)
• Add IDNA / Punycode (UTS #46 + RFC 3492)
• Add confusable / spoof detection (UTS #39)
• Phase 2: add robustness fuzzing + criterion benchmarks
• Add Unicode title-casing
• Add Unicode identifiers (UAX #31)
• Add line breaking (UAX #14)
• Add ROADMAP.md (path to ICU feature parity)
• Add sentence boundary segmentation (UAX #29)
• Add word boundary segmentation (UAX #29)
• Add grapheme cluster segmentation (UAX #29)

v0.1.3

Other

• add crates.io, docs.rs, CI, and license badges
• Add whole-string case adaptors: uppercase/lowercase/fold
• Format codegen (separate package, not covered by cargo fmt --all)
• Add UTS #10 collation (DUCET root collation)