From 8b57e90f5c563bbc1a6060d9f5e94b5d3cd91324 Mon Sep 17 00:00:00 2001
From: nsfisis <nsfisis@gmail.com>
Date: Mon, 17 Aug 2026 03:09:04 +0900
Subject: fix(pcre): preserve unmatched groups in Preg::match_all*()

PHP's Preg::matchAll() and matchAllWithOffsets() always set
PREG_UNMATCHED_AS_NULL, so a non-participating group is `null` and its
offset is -1. The Rust wrappers collapsed those to "" and 0, so callers
could not tell a group that did not participate from one that matched an
empty string at offset 0, and the offset value matched no PHP mode at
all.

Hand the shim's representation through unchanged and let each caller
mirror what the PHP original does with it: `isset()` and `(string)`
casts stay lenient, while `assert(is_string(...))` and the
*StrictGroups() variants become `expect()`.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---
 crates/shirabe/src/downloader/git_downloader.rs | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

(limited to 'crates/shirabe/src/downloader')

diff --git a/crates/shirabe/src/downloader/git_downloader.rs b/crates/shirabe/src/downloader/git_downloader.rs
index 4394cc95..b34f7800 100644
--- a/crates/shirabe/src/downloader/git_downloader.rs
+++ b/crates/shirabe/src/downloader/git_downloader.rs
@@ -109,7 +109,7 @@ impl GitDownloader {
             .cloned()
             .unwrap_or_default();
 
-        let mut branches_match: IndexMap<CaptureKey, Vec<String>> = IndexMap::new();
+        let mut branches_match: IndexMap<CaptureKey, Vec<Option<String>>> = IndexMap::new();
         if !Preg::is_match_all3(
             format!("{{^{} refs/heads/(.+)$}}mi", preg_quote(&head_ref, None)),
             &refs,
@@ -121,7 +121,10 @@ impl GitDownloader {
         let candidate_branches: Vec<String> = branches_match
             .get(&CaptureKey::ByIndex(1))
             .cloned()
-            .unwrap_or_default();
+            .unwrap_or_default()
+            .into_iter()
+            .map(|branch| branch.expect("group 1 participates whenever the pattern matches"))
+            .collect();
 
         // use the first match as branch name for now
         let mut branch = candidate_branches[0].clone();
@@ -134,7 +137,7 @@ impl GitDownloader {
 
             // try to find matching branch names in remote repos
             for candidate in &candidate_branches {
-                let mut m: IndexMap<CaptureKey, Vec<String>> = IndexMap::new();
+                let mut m: IndexMap<CaptureKey, Vec<Option<String>>> = IndexMap::new();
                 if Preg::is_match_all3(
                     format!(
                         "{{^[a-f0-9]+ refs/remotes/((?:[^/]+)/{})$}}mi",
@@ -143,11 +146,13 @@ impl GitDownloader {
                     &refs,
                     Some(&mut m),
                 ) {
-                    let matches: Vec<String> =
+                    let matches: Vec<Option<String>> =
                         m.get(&CaptureKey::ByIndex(1)).cloned().unwrap_or_default();
                     for match_ in matches {
                         branch = candidate.clone();
-                        remote_branches.push(match_);
+                        remote_branches.push(
+                            match_.expect("group 1 participates whenever the pattern matches"),
+                        );
                     }
                     break;
                 }
-- 
cgit v1.3.1-4-g156e