Repository navigation
fix(runtime): correct vstring numification for dotted-numeric literals (#1683) - #1697
Merged
Merged
Conversation
Vstrings (e.g. `49.46.48` or `v49.46.48`) store their content as raw
character bytes. Their string value is already the correct content
(codepoints 49, 46, 48 → "1.0"), but numification was building a
version-number representation ("49.046_048") instead of converting the
character content to a number.
Changes:
- Remove the VSTRING special case in `NumberParser.parseNumber` that
built "49.046_048" from codepoints; let it fall through to the normal
string path which calls `toString()` → "1.0" → 1.0 ✓
- Fix `RuntimeScalar.getInt/getLong/getDouble` VSTRING fast-paths that
were hardcoded to 0/0L/0.0; delegate to `NumberParser.parseNumber`
so arithmetic operators (not just `==`) get the right value too.
- Add `unit/vstring_dotted_numeric.t` regression test (issue #1683).
Fixes #1683
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…strings The previous commit (fix #1683) removed the VSTRING special case from `NumberParser.parseNumber` entirely, breaking `op/override.t` test 7: BEGIN { *CORE::GLOBAL::require = sub { $r = shift; 1 } } require v5.6; ok( abs($r - 5.006) < 0.001 && $r eq "\x05\x06" ); `require v5.6` (bytes [5,6]) passes the vstring to the override. Perl pre-caches the version NV (5.006) in the SV before calling the override, so the argument should numerify as ~5.006. Without the VSTRING special case, it numerified as 0 and the test failed. The fix uses a heuristic in `parseNumber`: - If the vstring's character content starts with a digit or sign (e.g. v49.46.48 → "1.0"), use normal string numification. This fixes issue #1683 (isvstring.t `dotted num` assertion). - If it starts with a non-printable byte (e.g. v5.6 → "\x05\x06"), use the version-number representation (e.g. "5.006"), matching what Perl's `require` pre-caches. Also expand the regression unit test to cover the require-override case (tests 8-9: `require v5.6` with CORE::GLOBAL::require override) and the printable-start vstring case (tests 10-13). Fixes regression in op/override.t introduced by commit 1047fbe. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…old peers Commit 57fbfea incorrectly removed the code path in Regex.java that renders caseless negated ASCII singletons with no fold peers as NEXACTb nodes. It matched system Perl 5.42 (NANYOFM) rather than blead Perl 5.45 (NEXACTb). Restore the removed guard and correct the unit test assertions that were updated to expect the wrong output. Fixes anyof.t test 559 regression. Validated against blead Perl at perl5/perl (v5.45.4): `(?i:[^:])` -> NEXACTb[:] (no fold peers: NEXACTb) `(?i:[^a])` -> NANYOFM[Aa] (has fold peers: NANYOFM) Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
fglock
force-pushed
the
fix/1683-vstring-dotted-numeric-numification
branch
from
October 8, 2026 15:24
b842791 to
ca2bca3
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Fixes #1683:
Scalar-List-Utils isvstring.twas failing thedotted numassertion because vstring numification built a version-number string ("49.046_048") instead of converting the vstring's character content ("1.0") to a number.Root cause
NumberParser.parseNumberhad a special VSTRING branch that combined each codepoint byte into a dotted-decimal representation (e.g. codepoints 49, 46, 48 →"49.046_048"). That format is forsprintf "%vd", not for numeric comparison.RuntimeScalar.getInt/getLong/getDoublehad hardcodedcase VSTRING -> 0shortcuts that also bypassed string numification.Fix
The fix uses a heuristic based on the first byte of the vstring's character content:
v49.46.48→"1.0", first byte = '1'): use normal string numification viatoString(). This fixes theisvstring.tdotted numassertion.v5.6→"\x05\x06", first byte = 0x05): use the version-number representation ("5.006"). This preservesop/override.ttest 7 compatibility:require v5.6with aCORE::GLOBAL::requireoverride passes the vstring with the version NV pre-cached (Perl stores 5.006 in the SV before calling the override).Also replace the
case VSTRING -> 0/0L/0.0shortcuts ingetInt/getLong/getDoublewith calls toNumberParser.parseNumber(this), so arithmetic operators also get the correct value.Tests
src/test/resources/unit/vstring_dotted_numeric.t(13 tests, passes on system Perl and on PerlOnJava, both backends):$vs == "1.0") — the exact failing assertion from Scalar-List-Utils isvstring.t fails the dotted-num assertion #1683isvstringreturns true for dotted-numerics and v-strings, false for plain strings and floatsrequire v5.6withCORE::GLOBAL::requireoverride: argument stringifies as"\x05\x06"AND numifies as ~5.006 (regression test forop/override.ttest 7)v49.46.48numifies as 1.0🤖 Generated with Claude Code