Changelog

Copy Markdown View Source

As of unicode_set version 1.4.0, Elixir 1.12 or later is required.

Unicode Set 1.7.0

This is the changelog for Unicode Set 1.7.0 released on July 9th, 2026. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Backslash-letter escapes are decoded correctly: \a \b \e \f \v \t \n \r map to their control codes, and any other \<char> maps to that literal character. Previously \a–\f silently decoded to the wrong codepoint and \g–\z / \U / \w raised.

  • Unicode.Set.parse/1 returns {:error, _} for genuinely unsupported syntax (\N{...}, \p{emoji=value}, multi-codepoint \u{...}) instead of raising, restoring its documented tagged-tuple contract.

  • The empty set [-] now reduces to {:in, []} and its consumers (to_pattern/1, to_utf8_char/1, to_regex_string/1) return an empty result or a never-matching regex instead of crashing.

  • A union of complements such as [[^a][^b]] now reduces correctly (per De Morgan) and no longer crashes to_regex_string/1 or the match?/2 search-tree path.

  • to_pattern/1 and compile_pattern/1 return a tagged error for complement ([^...]) sets rather than raising; the ! variants continue to raise.

  • Unicode.Set.match?/2 and the search tree no longer crash when matched against an empty string, and generate_matches/2 no longer crashes on a complement set.

  • Set operations now bind strictly left-to-right, including across an implicit-union boundary: [[a-f]-[b][g]&[g-z]] is now {g} and the README precedence example evaluates as documented. Previously a trailing & or - bound only to its immediate neighbour.

  • union/2 merges overlapping and adjacent ranges, so a union feeding a difference or intersection no longer retains codepoints that should have been subtracted; symmetric_difference/2 is likewise correct for overlapping inputs.

  • Reversed character ranges ([z-a]) and mismatched-length string ranges ([{abc}-{de}]) are rejected with a clear error instead of being silently accepted.

  • String members and string ranges are PCRE-escaped when emitted as a regex, so [{a.c}] matches the literal string and sets containing regex metacharacters no longer produce an uncompilable pattern (RE-1).

  • Sets of multiple string members no longer emit a bogus empty [[][]] class, and string-range alternations are wrapped in (?:...) so they compose correctly when embedded in a larger regex (SR-1, RE-4).

  • Character ranges with a surrogate endpoint are clipped rather than emitting a dangling - or dropping codepoints, and a surrogate-only set emits a never-matching (?!) instead of the uncompilable [] (RE-2, RE-5).

  • The regex splitter correctly handles a character class containing an escaped backslash such as [\\] (RS-2), and passes \Q...\E literal spans and (?#...) comments through verbatim rather than expanding any [...] inside them (RS-1, RS-4).

  • The Is<name> prefix now resolves as a script, general category or binary property before falling back to a block, so \p{IsAlphabetic}, \p{IsLatin} and [:IsLowercase:] resolve instead of erroring; Is<Block> names such as \p{IsBasicLatin} still resolve to their block (GAP-ISPREFIX).

  • Digit-bearing block names such as \p{block=Latin-1 Supplement} now resolve, working around a Unicode.Block.fetch/1 bug present in the unicode dependency (PS-7).

  • An empty property value such as \p{gc=} now returns a clean error instead of silently mis-parsing as a literal character plus a string member.

  • Multi-codepoint string members are emitted before the character class in a set's regex, longest first, so a string such as i̯ in [ij{i̯}] matches as a whole rather than being shadowed by the bare i in the class.

  • A negated set containing string members such as [^abc{de}] now drops the strings (which can never match a single position under negation) and keeps the code-point ranges, instead of returning an "unsupported" error.

Enhancements

  • Added the \UHHHHHHHH (8 hex digit) escape, the single-digit \xH escape, and single-codepoint bracketed \u{...} / \x{...} escapes (including astral codepoints such as \u{1F600}).

  • Added octal \0ooo escapes and \cX control escapes.

  • Multi-codepoint bracketed escapes such as \u{41 42 43} are now a string member (equivalent to {ABC}).

  • Implemented single-quote quoting: text within '...' is literal and '' is a literal quote (CLDR TR35).

  • \N{NAME} now resolves to its codepoint when built against unicode ~> 2.0 (which provides the character-name table); on earlier versions it returns a clean error.

  • Whitespace immediately after [ or [^ is now ignored, consistent with whitespace elsewhere in a set.

  • Hyphens are now accepted and ignored in property names per UAX44-LM3, so \p{White-Space} and [:Quotation-Mark:] resolve (PS-1).

  • Accept the Java-style In<Block> prefix, so \p{InBasicLatin} resolves to the block while genuine In... names such as \p{Inherited} are unaffected (PS-8).

  • The empty set is now written [] as well as [-], the empty-string member [{}] is supported, and a hyphen at the start or end of a set ([-a], [a-], [a-z-]) is treated as a literal hyphen, matching ICU.

  • The LC / Cased_Letter group general category (Lu | Ll | Lt) now resolves, for both \p{Cased_Letter} and \p{gc=LC}.

Changes

  • Corrected README examples: the block name Sundanese (was sudanese), a working \p{General_Category=...} property spelling, the single-dash print compatibility definition, and the to_regex_string/1 doc example.

  • Removed the unused :parse_many parser combinator.

  • Moved the Dialyzer ignore list to the term-format .dialyzer_ignore.exs.

  • Added a "Conformance" section to the README documenting supported syntax, deliberate tailorings, and current limitations, and a note explaining the POSIX-compatible [:punct:] definition.

  • Allow unicode ~> 2.0.

  • Add Credo, test coverage and CI hardening across the Elixir 1.17 to 1.20 / OTP 27 to 29 matrix.

Unicode Set 1.6.1

This is the changelog for Unicode Set 1.6.1 released on March 16th, 2026. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Fix a bug where a set starting with "-" was interpreted as negation.

Unicode Set 1.6.0

This is the changelog for Unicode Set 1.6.0 released on January 19th, 2026. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Fix tests for OTP 28 and Elixir 1.20.

Enhancements

Unicode Set 1.5.0

This is the changelog for Unicode Set 1.5.0 released on March 29th, 2025. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Converts all compile-time regex compilation to runtime to be compatible with OTP 28. Performance implications are not yet known.

Unicode Set 1.4.1

This is the changelog for Unicode Set 1.4.1 released on January 1st, 2025. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Work around the Elixir type checker for now.

Unicode Set 1.4.0

This is the changelog for Unicode Set 1.4.0 released on May 26th, 2024. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Fix warnings for Elixir 1.17. Thanks to @alco for the PR.

Unicode Set 1.3.0

This is the changelog for Unicode Set 1.3.0 released on February 18th, 2023. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Correct the code examples in README.md. Thanks to @DianaOlympos for the PR. Closes #9.

Enhancements

Unicode Set 1.2.0

This is the changelog for Unicode Set 1.2.0 released on September 15th, 2022. For older changelogs please consult the release tag on GitHub

Enhancements

  • Update parsing code to ensure compatibility against future deprecations. Thanks to @josevalim.

  • Fix library name in doc links. Thanks to @zmaril for the PR.

  • Update dependencies. Thanks to @kianmeng.

Unicode Set 1.1.0

This is the changelog for Unicode Set 1.1.0 released on September 15th, 2021. For older changelogs please consult the release tag on GitHub

Enhancements

  • ex_unicode is renamed to unicode in collaboration with @Qqwy and therefore this release updates the dependency name.

Unicode Set 1.0.0

This is the changelog for Unicode Set 1.0.0 released on September 14th, 2021. For older changelogs please consult the release tag on GitHub

Enhancements

Unicode Set 0.13.1

This is the changelog for Unicode Set 0.13.1 released on May 25th, 2021. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Update dependency configuration to mark ex_doc and benchee as optional. Thanks to @fireproofsocks.

Unicode Set 0.13.0

This is the changelog for Unicode Set 0.13.0 released on April 4th, 2021. For older changelogs please consult the release tag on GitHub

Enhancements

  • Adds Unicode.Set.to_generate_matches/1 that returns a tuple whose first element is the AST of a guard clause and the second element is a list of strings. This function is marked private and is implemented to support unicode_transform which uses this information to generate optimised code for matching unicode sets in a case expression.

Unicode Set 0.12.0

This is the changelog for Unicode Set 0.12.0 released on February 23rd, 2021. For older changelogs please consult the release tag on GitHub

Enhancements

  • Adds support for "isBlockName" Perl and POSIX regex syntax. Used in a regex as [[:isLatin1]] or \p{isLatin1} or their inverse forms [[:^isLatin1]] and \P{isLatin1}.

Unicode Set 0.11.0

This is the changelog for Unicode Set 0.11.0 released on October 5th, 2020. For older changelogs please consult the release tag on GitHub

Enhancements

  • Add recursively defined sets to support compatibility with Posix classes. See Unicode.Set.Property.

Bug Fixes

  • Fix various bugs in set operations for Union, Difference, Intersection abd Complement

  • Correctly parse and interpret set complements such as [^[:^Sc:]] and more complex sets such as [^[[:Sc:]-[:^Lu:]]]

Unicode Set 0.10.0

This is the changelog for Unicode Set 0.10.0 released on October 2nd, 2020. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Fix list composition in Unicode.Set.to_uft8_char/1

Unicode Set 0.9.0

This is the changelog for Unicode Set 0.9.0 released on October 2nd, 2020. For older changelogs please consult the release tag on GitHub

Enhancements

  • Support nimble_parsec version 1.x. Thanks to @josevalim for the PR.

Unicode Set 0.8.0

This is the changelog for Unicode Set 0.8.0 released on July 12th, 2020. For older changelogs please consult the release tag on GitHub

Enhancements

  • Rewrite Unicode.Regex module to better extract character classes, process unicode sets and build Elixir regexs. Now also supports string ranges.

  • Supports the property East Asian Width (short name ea) which is required for implementing the Unicode segmentation algorithms. Also bumps the minimum requirement for ex_unicode version 1.8.

Unicode Set 0.7.0

This is the changelog for Unicode Set v.07.0 released on May 18th, 2020. For older changelogs please consult the release tag on GitHub

Enhancements

  • Add Unicode.Set.character_class/1 which returns a string compatible with Regex.compile/2. This supports the idea of expanded Unicode Sets being used in standard Elixir/erlang regular expressions and will underpin implementation of Unicode Transforms in the package unicode_transform

  • Add Unicode.Regex.compile/2 to pre-process a regex to expand Unicode Sets and the compile it with Regex.compile/2. Unicode.Regex.compile!/2 is also added.

Bug Fixes

  • Fixes a bug whereby a Unicode Set intersection would fail with a character class that starts at the same codepoint as the Unicode set.

Unicode Set 0.6.0

This is the changelog for Unicode Set v.06.0 released on May 13th, 2020. For older changelogs please consult the release tag on GitHub

Enhancements

  • Unicode sets are now a %Unicode.Set{} struct

  • Add Unicode.Set.Sigil implementing sigil_u

  • Add support for String.Chars and Inspect protocols

Bug Fixes

  • Fixes parsing sets to ignore non-encoded whitespace

  • Fixes intersection and difference set operations for sets that include string ranges like {abc}

Unicode Set 0.5.1

This is the changelog for Unicode Set v.05.1 released on March 14th, 2020. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Compacts tuple-ranges in order to minimize the number of generated clauses in guards. Requires at least ex_unicode version 1.5.0.

Unicode Set 0.5.0

This is the changelog for Unicode Set v.05.0 released on March 11th, 2020. For older changelogs please consult the release tag on GitHub

Enhancements

  • Updates ex_unicode to version 1.4.0 which includes support for Unicode version 13.0 as well as support for several derived categories related to quote marks.

Unicode Set 0.4.2

This is the changelog for Unicode Set v.04.2 released on February 25th, 2020. For older changelogs please consult the release tag on GitHub

Bug Fixes

  • Allow \n, \t and \r, \s as part of character classes

Unicode Set 0.4.1

This is the changelog for Unicode Set v.04.1 released on January 8th, 2020. For older changelogs please consult the release tag on GitHub

Bug Fixes

Unicode Set 0.4.0

This is the changelog for Unicode Set v.04.0 released on November 27th, 2019. For older changelogs please consult the release tag on GitHub

Enhancements

  • Bump to ex_unicode to version 1.3.0 to support an expanded set of properties resolved by unicode_set.

Unicode Set 0.3.0

This is the changelog for Unicode Set v.03.0 released on November 26th, 2019. For older changelogs please consult the release tag on GitHub

Enhancements

  • Support string ranges expressed as {abc} or {abc}-{def}

  • Note that the supported proporties in this release are script, block, category and combining class.

Unicode Set 0.2.0

This is the changelog for Unicode Set v.02.0 released on November 24th, 2019. For older changelogs please consult the release tag on GitHub

Enhancements

Set the README for example usage.

Unicode Set 0.1.0

This is the changelog for Unicode Set v.01.0 released on November 23rd, 2019. For older changelogs please consult the release tag on GitHub

Initial release.