| Age | Commit message (Collapse) | Author |
|---|
|
The two functions ran the same search and differed only in how they
reported a non-participating group: one as ("", 0) in a bespoke struct,
the other as (null, -1) in a capture-key map. Neither shape covered both
callers, because PHP reaches preg_match_all() with different flags from
each: Preg::matchAllWithOffsets() always ORs in PREG_UNMATCHED_AS_NULL,
while OutputFormatter::formatAndWrap() passes PREG_OFFSET_CAPTURE alone.
Keep the capture-key map, which also exposes named groups, and take the
flags that decide between null and "" for an unmatched group. The offset
is -1 either way, so the ("", 0) approximation is gone.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
|
The sole caller, Preg::match_all_with_offsets5(), always passed flags=0
and offset=0, so PREG_UNMATCHED_AS_NULL is now unconditional and the
subject is scanned from the beginning. That method is only reached from
Preg::match_all_with_offsets(), so it is no longer public either.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
|
Every caller reached preg_match_all2() through Preg::match_all5(), which
always passed flags = PREG_UNMATCHED_AS_NULL and offset = 0. Inline those
constants and make match_all5() private, since the two public wrappers are
its only callers.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
|
Callers had to build a temporary Vec<&str> at every call site to satisfy
the &[&str] parameter. Taking IntoIterator and yielding the matched items
lets them pass owned or borrowed strings directly.
The flags variant and PREG_GREP_INVERT go away with it: no caller passes
flags, and Composer's Preg::grep() has no such parameter either.
|
|
|
|
preg_replace_callback2 had one caller, Preg::replace_callback6, which
was itself reached only from Preg::replace_callback with the default
limit, count and flags. Fold the two shim functions into one and drop
those parameters along with replace_callback6.
The surviving callback takes callback2's IndexMap of matches: it can be
keyed by group name, and it omits trailing non-participating groups the
way PHP does, which is what the callbacks in Process and ProgressBar
test for.
|
|
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
|
Retag every Shirabe-authored TODO comment to one of the fixed tags:
phase-c, phase-d, plugin, php-runtime, phase-e.
Upstream-authored TODO comments from Composer/Symfony are left
untouched to preserve the ported code shape.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
|
|
regex::Regex::clone() does not share the underlying meta engine's
search-cache pool, so every fresh clone pays a ~10us warmup cost on
its first use. Two changes together eliminate this across nearly all
preg_* call sites:
- A php_regex! macro resolves PHP-style patterns to a per-call-site
&'static regex::Regex (via regex-macro's LazyLock), applied at the
majority of call sites throughout the codebase.
- Call sites still passing dynamic pattern strings go through
PATTERN_CACHE, which now stores Arc<(Regex, bool)> and hands out
Arc::clone()s instead of cloning the Regex itself.
PregPattern::resolve() returns a ResolvedPattern enum (Arc or
'static reference) rather than an owned Regex, so neither path ever
clones the Regex proper.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
Composer's classmap generator re-issues the same PHP-derived pattern
string for every scanned file (and every token within it), relying on
PCRE's built-in compiled-pattern cache to make that free. shirabe had
no equivalent, so every Preg::* call recompiled the pattern from
scratch via regex::Regex::new(), making `composer create-project`
autoload generation ~130x slower than Composer on a fresh laravel/laravel
install (130s vs ~1s). Memoizing compiled patterns by their raw string
in compile_php_pattern closes that gap.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
|
|
|
|
Port autoload_generator (24), bitbucket (14), suggested_packages (11),
git_driver (6), archive_manager (3), and a bump command test. Fix the
ApplicationTester output-capture root cause (php://memory streams must be
readable regardless of fopen mode). Implement posix_getuid/geteuid, the PCRE
'A' anchored modifier, php_strip_whitespace, stream_get_wrappers, is_callable
scalars; fix preg_quote angle-bracket escaping and class-map parser regexes.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
|
|
|
|
|
|
The optional matches output was always passed Some(&mut ...) at every
call site, so the Option wrapper added no value. Take &mut directly and
inline the former internal helpers.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
preg_match2 returns bool and the preg_match_all* helpers return usize
(the match count is never negative), matching how callers use them.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
preg_match can only return 1 or 0 now that compile failure panics, so
return bool and update all call sites accordingly.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
The preg_* shim helpers wrapped their results in Option/Result solely to
signal a regex that failed to compile. Composer never feeds a pattern
that fails at runtime, so such a failure is a programming error: panic
instead and drop the Option/Result wrappers, updating all callers.
preg_replace_callback keeps its Result return type since the callback
itself is fallible. preg_match_groups is removed in favor of preg_match
at its sole call site.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
The preg.rs shim had several near-duplicate functions: simple helpers
re-implementing logic already covered by their full-featured `2`
variants. Delegate or remove the redundant ones while preserving
behavior, and migrate the affected callers:
- preg_replace / preg_split now delegate to preg_replace2 / preg_split2
- preg_match_all_simple removed; its caller uses preg_match_all
- preg_split_chars removed; its caller uses a char-boundary iterator
- preg_match_offset removed; its callers use preg_match2 directly
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
|
The free preg_* functions, the CaptureKey type, and the PREG_* constants
move from the pcre crate into shirabe-php-shim, where the rest of the
PHP standard-library shims live. Colliding names take a 2 suffix
(preg_match2, preg_split2, etc.) since the shim already exposes
differently-shaped variants. Preg keeps its public API and delegates to
the relocated functions; CaptureKey is re-exported from composer::pcre
so consumers are unaffected.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
|
