refactor(pcre): extract composer/pcre into the shirabe-pcre crate

Move `Composer\Pcre` out of shirabe-external-packages and into its own
crate, so the path is `shirabe_pcre::preg::Preg` instead of
`shirabe_external_packages::composer::pcre::preg::Preg`.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

9 files changed, 6 insertions, 377 deletions

diff --git a/crates/shirabe-external-packages/Cargo.toml b/crates/shirabe-external-packages/Cargo.toml
index f274eac9..331efcb7 100644
--- a/crates/shirabe-external-packages/Cargo.toml
+++ b/crates/shirabe-external-packages/Cargo.toml
@@ -4,6 +4,7 @@ version.workspace = true
 edition.workspace = true
 
 [dependencies]
+shirabe-pcre.workspace = true
 shirabe-php-shim.workspace = true
 shirabe-semver.workspace = true
 anyhow.workspace = true
diff --git a/crates/shirabe-external-packages/src/composer.rs b/crates/shirabe-external-packages/src/composer.rs
index 907f9852..da1ab285 100644
--- a/crates/shirabe-external-packages/src/composer.rs
+++ b/crates/shirabe-external-packages/src/composer.rs
@@ -1,7 +1,5 @@
 pub mod ca_bundle;
-pub mod pcre;
 pub mod xdebug_handler;
 
 pub use ca_bundle::*;
-pub use pcre::*;
 pub use xdebug_handler::*;
diff --git a/crates/shirabe-external-packages/src/composer/pcre.rs b/crates/shirabe-external-packages/src/composer/pcre.rs
deleted file mode 100644
index 20c37d02..00000000
--- a/crates/shirabe-external-packages/src/composer/pcre.rs
+++ /dev/null
@@ -1,3 +0,0 @@
-pub mod preg;
-
-pub use preg::*;
diff --git a/crates/shirabe-external-packages/src/composer/pcre/preg.rs b/crates/shirabe-external-packages/src/composer/pcre/preg.rs
deleted file mode 100644
index 06586d2c..00000000
--- a/crates/shirabe-external-packages/src/composer/pcre/preg.rs
+++ /dev/null
@@ -1,367 +0,0 @@
-//! ref: composer/vendor/composer/pcre/src/Preg.php
-//!
-//! The following two exception classes are intentionally not ported:
-//!
-//! - `PcreException`: thrown when a `preg_*()` call returns false. Composer never feeds a pattern
-//!   that fails to compile at runtime, so such a failure would be a programming error rather than
-//!   a recoverable condition; they panic instead.
-//! - `UnexpectedNullMatchException`: thrown by the `Preg::*StrictGroups()` variants when a capture
-//!   group did not participate. Those variants were dropped because Rust's `Option` already
-//!   distinguishes participating from non-participating groups.
-//!
-//! See docs/dev/regex-porting.md for more detailed regex porting rules.
-
-use indexmap::IndexMap;
-pub use shirabe_php_shim::CaptureKey;
-use shirabe_php_shim::{
-    PREG_OFFSET_CAPTURE, PREG_SET_ORDER, PREG_SPLIT_OFFSET_CAPTURE, PREG_UNMATCHED_AS_NULL,
-    PregPattern,
-};
-
-#[derive(Debug)]
-pub struct Preg;
-
-impl Preg {
-    pub fn match3(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, String>>,
-    ) -> bool {
-        Self::match5(pattern, subject, matches, 0, 0)
-    }
-
-    pub fn match5(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, String>>,
-        flags: i64,
-        offset: usize,
-    ) -> bool {
-        Self::check_offset_capture(flags, "matchWithOffsets");
-
-        let mut internal: IndexMap<CaptureKey, Option<String>> = IndexMap::new();
-        let result = shirabe_php_shim::preg_match2(
-            pattern,
-            subject,
-            &mut internal,
-            flags | PREG_UNMATCHED_AS_NULL,
-            offset,
-        );
-
-        if let Some(out) = matches {
-            *out = drop_null_matches(internal);
-        }
-
-        result
-    }
-
-    pub fn match_all(pattern: impl PregPattern, subject: &str) -> usize {
-        Self::match_all5(pattern, subject, None, 0, 0)
-    }
-
-    pub fn match_all3(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, Vec<String>>>,
-    ) -> usize {
-        Self::match_all5(pattern, subject, matches, 0, 0)
-    }
-
-    pub fn match_all5(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, Vec<String>>>,
-        flags: i64,
-        offset: usize,
-    ) -> usize {
-        Self::check_offset_capture(flags, "matchAllWithOffsets");
-        Self::check_set_order(flags);
-
-        let mut internal: IndexMap<CaptureKey, Vec<Option<String>>> = IndexMap::new();
-        let result = shirabe_php_shim::preg_match_all2(
-            pattern,
-            subject,
-            &mut internal,
-            flags | PREG_UNMATCHED_AS_NULL,
-            offset,
-        );
-
-        if let Some(out) = matches {
-            *out = null_to_empty_match_all(internal);
-        }
-
-        result
-    }
-
-    pub fn match_all_with_offsets5(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, Vec<(String, usize)>>>,
-        flags: i64,
-        offset: usize,
-    ) -> usize {
-        Self::check_set_order(flags);
-
-        let mut internal: IndexMap<CaptureKey, Vec<(Option<String>, i64)>> = IndexMap::new();
-        let result = shirabe_php_shim::preg_match_all_offset_capture2(
-            pattern,
-            subject,
-            &mut internal,
-            flags | PREG_UNMATCHED_AS_NULL | PREG_OFFSET_CAPTURE,
-            offset,
-        );
-
-        if let Some(out) = matches {
-            *out = null_to_empty_offset_match_all(internal);
-        }
-
-        result
-    }
-
-    pub fn replace(pattern: impl PregPattern, replacement: &str, subject: &str) -> String {
-        Self::replace_impl(pattern, replacement, subject, -1, None)
-    }
-
-    pub fn replace4(
-        pattern: impl PregPattern,
-        replacement: &str,
-        subject: &str,
-        limit: i64,
-    ) -> String {
-        Self::replace_impl(pattern, replacement, subject, limit, None)
-    }
-
-    pub fn replace5(
-        pattern: impl PregPattern,
-        replacement: &str,
-        subject: &str,
-        limit: i64,
-        count: &mut usize,
-    ) -> String {
-        Self::replace_impl(pattern, replacement, subject, limit, Some(count))
-    }
-
-    fn replace_impl(
-        pattern: impl PregPattern,
-        replacement: &str,
-        subject: &str,
-        limit: i64,
-        count: Option<&mut usize>,
-    ) -> String {
-        // `$subject` is statically a string here, so the is_scalar/is_array
-        // guards (ARRAY_MSG / INVALID_TYPE_MSG) of the PHP original are
-        // unreachable and not reproduced.
-        shirabe_php_shim::preg_replace2(pattern, replacement, subject, limit, count)
-    }
-
-    pub fn replace_callback<F: FnMut(&IndexMap<CaptureKey, String>) -> String>(
-        pattern: impl PregPattern,
-        replacement: F,
-        subject: &str,
-    ) -> String {
-        Self::replace_callback6(pattern, replacement, subject, -1, None, 0)
-    }
-
-    pub fn replace_callback6<F: FnMut(&IndexMap<CaptureKey, String>) -> String>(
-        pattern: impl PregPattern,
-        mut replacement: F,
-        subject: &str,
-        limit: i64,
-        count: Option<&mut usize>,
-        flags: i64,
-    ) -> String {
-        let adapter = |internal: &IndexMap<CaptureKey, Option<String>>| -> String {
-            replacement(&drop_null_matches_ref(internal))
-        };
-
-        shirabe_php_shim::preg_replace_callback2(pattern, adapter, subject, limit, count, flags)
-    }
-
-    pub fn split(pattern: impl PregPattern, subject: &str) -> Vec<String> {
-        Self::split4(pattern, subject, -1, 0)
-    }
-
-    pub fn split4(pattern: impl PregPattern, subject: &str, limit: i64, flags: i64) -> Vec<String> {
-        assert!(
-            flags & PREG_SPLIT_OFFSET_CAPTURE == 0,
-            "PREG_SPLIT_OFFSET_CAPTURE is not supported as it changes the type of $matches, use splitWithOffsets() instead"
-        );
-
-        shirabe_php_shim::preg_split2(pattern, subject, limit, flags)
-    }
-
-    pub fn grep(pattern: impl PregPattern, array: &[&str]) -> Vec<String> {
-        Self::grep3(pattern, array, 0)
-    }
-
-    pub fn grep3(pattern: impl PregPattern, array: &[&str], flags: i64) -> Vec<String> {
-        shirabe_php_shim::preg_grep2(pattern, array, flags)
-    }
-
-    pub fn is_match(pattern: impl PregPattern, subject: &str) -> bool {
-        Self::match5(pattern, subject, None, 0, 0)
-    }
-
-    pub fn is_match3(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, String>>,
-    ) -> bool {
-        Self::match5(pattern, subject, matches, 0, 0)
-    }
-
-    pub fn is_match5(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, String>>,
-        flags: i64,
-        offset: usize,
-    ) -> bool {
-        Self::match5(pattern, subject, matches, flags, offset)
-    }
-
-    pub fn is_match_named(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: &mut IndexMap<String, String>,
-    ) -> bool {
-        let mut internal: IndexMap<CaptureKey, Option<String>> = IndexMap::new();
-        let result = shirabe_php_shim::preg_match2(
-            pattern,
-            subject,
-            &mut internal,
-            PREG_UNMATCHED_AS_NULL,
-            0,
-        );
-
-        matches.clear();
-        for (key, value) in internal {
-            if let (CaptureKey::ByName(name), Some(value)) = (key, value) {
-                matches.insert(name, value);
-            }
-        }
-
-        result
-    }
-
-    pub fn is_match_with_indexed_captures(
-        pattern: impl PregPattern,
-        subject: &str,
-    ) -> Option<Vec<String>> {
-        // Classic preg_match semantics (no PREG_UNMATCHED_AS_NULL): trailing
-        // unmatched groups are truncated, interior unmatched groups become "".
-        let mut internal: IndexMap<CaptureKey, Option<String>> = IndexMap::new();
-        let result = shirabe_php_shim::preg_match2(pattern, subject, &mut internal, 0, 0);
-
-        if !result {
-            return None;
-        }
-
-        let max_index = internal
-            .keys()
-            .filter_map(|key| match key {
-                CaptureKey::ByIndex(index) => Some(*index),
-                CaptureKey::ByName(_) => None,
-            })
-            .max()
-            .unwrap_or(0);
-
-        let mut captures = Vec::with_capacity(max_index + 1);
-        for index in 0..=max_index {
-            let value = internal
-                .get(&CaptureKey::ByIndex(index))
-                .and_then(|value| value.clone())
-                .unwrap_or_default();
-            captures.push(value);
-        }
-
-        Some(captures)
-    }
-
-    pub fn is_match_all3(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, Vec<String>>>,
-    ) -> bool {
-        Self::match_all5(pattern, subject, matches, 0, 0) > 0
-    }
-
-    pub fn is_match_all_with_offsets3(
-        pattern: impl PregPattern,
-        subject: &str,
-        matches: Option<&mut IndexMap<CaptureKey, Vec<(String, usize)>>>,
-    ) -> bool {
-        Self::match_all_with_offsets5(pattern, subject, matches, 0, 0) > 0
-    }
-
-    fn check_offset_capture(flags: i64, use_function_name: &str) {
-        assert!(
-            flags & PREG_OFFSET_CAPTURE == 0,
-            "PREG_OFFSET_CAPTURE is not supported as it changes the type of $matches, use {}() instead",
-            use_function_name
-        );
-    }
-
-    fn check_set_order(flags: i64) {
-        assert!(
-            flags & PREG_SET_ORDER == 0,
-            "PREG_SET_ORDER is not supported as it changes the type of $matches"
-        );
-    }
-}
-
-// Drops `null` (unmatched) groups, mirroring how the public `string`-valued
-// `matches` map represents PHP's `string|null` entries by their absence.
-fn drop_null_matches(
-    matches: IndexMap<CaptureKey, Option<String>>,
-) -> IndexMap<CaptureKey, String> {
-    matches
-        .into_iter()
-        .filter_map(|(key, value)| value.map(|value| (key, value)))
-        .collect()
-}
-
-fn drop_null_matches_ref(
-    matches: &IndexMap<CaptureKey, Option<String>>,
-) -> IndexMap<CaptureKey, String> {
-    matches
-        .iter()
-        .filter_map(|(key, value)| value.clone().map(|value| (key.clone(), value)))
-        .collect()
-}
-
-// In the `Vec<String>`-valued maps a per-iteration `null` cannot be stored, so
-// unmatched groups collapse to "" (the classic non-PREG_UNMATCHED_AS_NULL form).
-fn null_to_empty_match_all(
-    matches: IndexMap<CaptureKey, Vec<Option<String>>>,
-) -> IndexMap<CaptureKey, Vec<String>> {
-    matches
-        .into_iter()
-        .map(|(key, values)| {
-            (
-                key,
-                values
-                    .into_iter()
-                    .map(|value| value.unwrap_or_default())
-                    .collect(),
-            )
-        })
-        .collect()
-}
-
-fn null_to_empty_offset_match_all(
-    matches: IndexMap<CaptureKey, Vec<(Option<String>, i64)>>,
-) -> IndexMap<CaptureKey, Vec<(String, usize)>> {
-    matches
-        .into_iter()
-        .map(|(key, values)| {
-            (
-                key,
-                values
-                    .into_iter()
-                    .map(|(value, offset)| (value.unwrap_or_default(), offset.max(0) as usize))
-                    .collect(),
-            )
-        })
-        .collect()
-}
diff --git a/crates/shirabe-external-packages/src/symfony/console/descriptor/json_descriptor.rs b/crates/shirabe-external-packages/src/symfony/console/descriptor/json_descriptor.rs
index 6c1665e9..3232e6fa 100644
--- a/crates/shirabe-external-packages/src/symfony/console/descriptor/json_descriptor.rs
+++ b/crates/shirabe-external-packages/src/symfony/console/descriptor/json_descriptor.rs
@@ -1,6 +1,5 @@
 //! ref: composer/vendor/symfony/console/Descriptor/JsonDescriptor.php
 
-use crate::composer::pcre::preg::Preg;
 use crate::symfony::console::application::Application;
 use crate::symfony::console::command::command::Command;
 use crate::symfony::console::descriptor::application_description::ApplicationDescription;
@@ -13,6 +12,7 @@ use crate::symfony::console::input::input_definition::InputDefinition;
 use crate::symfony::console::input::input_option::InputOption;
 use crate::symfony::console::output::output_interface::OutputInterface;
 use indexmap::IndexMap;
+use shirabe_pcre::preg::Preg;
 use shirabe_php_shim::{PhpMixed, php_regex};
 
 /// JSON descriptor.
diff --git a/crates/shirabe-external-packages/src/symfony/console/descriptor/markdown_descriptor.rs b/crates/shirabe-external-packages/src/symfony/console/descriptor/markdown_descriptor.rs
index 12afd53b..48514cfe 100644
--- a/crates/shirabe-external-packages/src/symfony/console/descriptor/markdown_descriptor.rs
+++ b/crates/shirabe-external-packages/src/symfony/console/descriptor/markdown_descriptor.rs
@@ -1,6 +1,5 @@
 //! ref: composer/vendor/symfony/console/Descriptor/MarkdownDescriptor.php
 
-use crate::composer::pcre::preg::Preg;
 use crate::symfony::console::application::Application;
 use crate::symfony::console::command::command::Command;
 use crate::symfony::console::descriptor::application_description::ApplicationDescription;
@@ -14,6 +13,7 @@ use crate::symfony::console::input::input_definition::InputDefinition;
 use crate::symfony::console::input::input_option::InputOption;
 use crate::symfony::console::output::output_interface::OutputInterface;
 use indexmap::IndexMap;
+use shirabe_pcre::preg::Preg;
 use shirabe_php_shim::PhpMixed;
 
 /// Markdown descriptor.
diff --git a/crates/shirabe-external-packages/src/symfony/console/descriptor/text_descriptor.rs b/crates/shirabe-external-packages/src/symfony/console/descriptor/text_descriptor.rs
index 4b85844e..147fe73f 100644
--- a/crates/shirabe-external-packages/src/symfony/console/descriptor/text_descriptor.rs
+++ b/crates/shirabe-external-packages/src/symfony/console/descriptor/text_descriptor.rs
@@ -1,6 +1,5 @@
 //! ref: composer/vendor/symfony/console/Descriptor/TextDescriptor.php
 
-use crate::composer::pcre::preg::Preg;
 use crate::symfony::console::application::Application;
 use crate::symfony::console::command::command::Command;
 use crate::symfony::console::descriptor::application_description::ApplicationDescription;
@@ -15,6 +14,7 @@ use crate::symfony::console::input::input_definition::InputDefinition;
 use crate::symfony::console::input::input_option::InputOption;
 use crate::symfony::console::output::output_interface::OutputInterface;
 use indexmap::IndexMap;
+use shirabe_pcre::preg::Preg;
 use shirabe_php_shim::PhpMixed;
 
 /// Text descriptor.
diff --git a/crates/shirabe-external-packages/src/symfony/console/helper/table.rs b/crates/shirabe-external-packages/src/symfony/console/helper/table.rs
index a1f5a11e..650dfffa 100644
--- a/crates/shirabe-external-packages/src/symfony/console/helper/table.rs
+++ b/crates/shirabe-external-packages/src/symfony/console/helper/table.rs
@@ -1,6 +1,5 @@
 //! ref: composer/vendor/symfony/console/Helper/Table.php
 
-use crate::composer::pcre::preg::Preg;
 use crate::symfony::console::exception::invalid_argument_exception::InvalidArgumentException;
 use crate::symfony::console::exception::runtime_exception::RuntimeException;
 use crate::symfony::console::formatter::output_formatter::OutputFormatter;
@@ -14,6 +13,7 @@ use crate::symfony::console::helper::table_style::TableStyle;
 use crate::symfony::console::output::console_section_output::ConsoleSectionOutput;
 use crate::symfony::console::output::output_interface::OutputInterface;
 use indexmap::IndexMap;
+use shirabe_pcre::preg::Preg;
 use shirabe_php_shim::{PhpMixed, php_regex};
 
 /// A single cell within a table row.
diff --git a/crates/shirabe-external-packages/src/symfony/finder/finder.rs b/crates/shirabe-external-packages/src/symfony/finder/finder.rs
index c2b6c4cd..4cdcf392 100644
--- a/crates/shirabe-external-packages/src/symfony/finder/finder.rs
+++ b/crates/shirabe-external-packages/src/symfony/finder/finder.rs
@@ -6,10 +6,10 @@
 //! `RecursiveDirectoryIterator` attaches to each `SplFileInfo` is carried on the
 //! private `Entry` struct so the path/exclude filters keep their exact behavior.
 
-use crate::composer::pcre::{CaptureKey, Preg};
 use crate::symfony::finder::glob::Glob;
 use chrono::{NaiveDate, NaiveDateTime};
 use indexmap::{IndexMap, IndexSet};
+use shirabe_pcre::{CaptureKey, Preg};
 use shirabe_php_shim::{file_exists, glob, is_dir, php_regex, preg_quote, rtrim};
 use std::path::{Path, PathBuf};
 use std::time::UNIX_EPOCH;