[wleeaf]

NCNAME_STR leaves COMBININGCHAR and EXTENDER as TODOs, so an element or attribute name fails to parse when a later character is a combining mark that isn't [:alpha:], or the middle dot. REXML builds and writes these names itself, but can't read them back:

Name	master	This PR	Expat (Python xml.etree)
café (decomposed é)	Invalid attribute name: <́>	parses	parses
col·leccio (Catalan middle dot)	Invalid attribute name: <·leccio>	parses	parses
เจมส์ (Thai, U+0E4C)	Invalid attribute name: <์>	parses	parses
क्षेत्र (Devanagari virama U+094D)	Invalid attribute name: <्षेत्र>	parses	parses
பெயர் (Tamil virama U+0BCD)	Invalid attribute name: <்>	parses	parses

xml = REXML::Element.new("col·leccio").to_s  # => "<col·leccio/>"
REXML::Document.new(xml)                           # master: REXML::ParseException

[4] NameChar includes CombiningChar and Extender. This change adds:

• the combining marks (\p{M}) that [:alpha:] doesn't already include. Many vowel signs are Alphabetic, but tone marks, viramas and U+0300-U+036F aren't. Using an intersection keeps the character class free of duplicated ranges, so there is no -w warning.
• U+00B7 and U+0387, the only Extender characters that aren't [:alpha:].

Names starting with a mark are still rejected, and no character that was accepted before is rejected now.

Tests: test_combining_character_in_name (element and attribute name) and test_extender_in_name in test/parse/test_element.rb. Both fail before the change and pass after it. bundle exec rake test passes (843 tests), as do the --enable-frozen-string-literal run and bundle exec rake warning:error rdoc. ruby -w shows no new warnings.

Performance: tag scanning with TAG_PATTERN alone is 1-5% slower in a micro-benchmark. Parsing a 5000-element document with 2 attributes per element showed no measurable difference (best of 15 runs: 0.172s vs 0.168s).

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has insufficient remaining quota.