Fix tr.t and multideref.t: Unicode support + strict-refs #65
Conversation
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
- Add support for \N{name} syntax with actual Unicode character names
using UnicodeResolver integration with ICU4J (~30,000+ names supported)
- Fix empty \N{} validation to give proper 'Unknown charname' error
- Fix surrogate pair handling bug that was incorrectly removing characters
by checking isSupplementaryCodePoint() instead of isHighSurrogate()
Test improvements: 256/318 (80.5%) -> 277/318 (87.1%)
- Fixed 21 tests (+6.6%)
- Tests now run to completion (previously died at line 1113)
Examples:
$s =~ tr/\N{LATIN SMALL LETTER E WITH ACUTE}/E/; # now works
$s = "\x{d800}\x{ffff}"; $s =~ tr/\0/A/; # now preserves both chars
This commit properly implements the distinction between strict and non-strict dereference modes, fixing multideref.t and avhv.t without breaking other tests. Key changes: - Added compile-time strict-refs checking in Dereference.java - Implemented *NonStrict methods in RuntimeScalar and RuntimeBaseProxy - Fixed type checking in hashDerefNonStrict/arrayDerefNonStrict to only allow symbolic references for STRING/BYTE_STRING/GLOB types - Fixed context conversion for %$var in scalar context to prevent VerifyError Test results: - multideref.t: 2/65 → 56/65 (+54 tests) - avhv.t: 9/40 → 40/40 (+31 tests) - tr.t: 277 → 277 (maintained) - magic.t: 103 → 103 (no regression) Total: +85 passing tests
This fixes the crash 'Modification of a read-only value attempted'
when using symbolic references like ${"\!"}{ENOENT}.
For RuntimeScalarReadOnly (immutable scalars), lvalue is null.
When NonStrict methods are called, we now delegate to super
instead of calling vivify(), which allows symbolic references
to work without trying to modify read-only values.
Result: magic.t now passes 158/208 tests (was crashing at 103)
Known issue: Double symbolic dereference like ${"foo"}{key}
still returns empty. This needs further investigation.
The fix was actually in RuntimeBaseProxy, not in how the block is evaluated. The original code was correct.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR fixes both
t/op/tr.tandt/op/multideref.twithout breaking other tests.Test Results
Baseline → After Fix:
tr.t: 256/318 → 277/318 (+21 tests) ✅multideref.t: 2/65 → 56/65 (+54 tests) ✅avhv.t: 9/40 → 40/40 (+31 tests) ✅magic.t: 103 → 103 (no regression) ✅Total improvement: +106 passing tests, 0 regressions!
Commits
1. tr/// operator improvements (commit 1)
\N{name}syntax with ICU4J integration (~30,000+ names)$s =~ tr/\N{LATIN SMALL LETTER E WITH ACUTE}/E/;isSupplementaryCodePoint()instead ofisHighSurrogate()\N{}validation: Proper "Unknown charname" error2. Strict-refs support for hash/array dereferences (commit 2)
Dereference.java*NonStrictmethods inRuntimeScalarandRuntimeBaseProxyhashDerefNonStrict/arrayDerefNonStrict:no strict 'refs'%\$varin scalar context to prevent VerifyErrorTechnical Details
The strict-refs fix correctly distinguishes between:
$foo = "bar"; $foo->{key}→ looks up%bar$arr = []; $arr->{key}→ "Not a HASH reference"This matches Perl's behavior and satisfies the Java bytecode verifier.