clean-ignore: support utf-8 filenames found in .gitignore

If someone names files with emoji, and puts those files in their .gitignore, ensure we correctly parse and handle those files. Do so by passing -z to `git check-ignore`, so that we don't have to worry about weird quoting of special characters. Reported-by: Yufan Lou (@louy2 on github) Reported-by: Martin Ranieri (@maffinyep on github) Reported-by: Aral Balkan <[email protected]> Signed-off-by: Elijah Newren <[email protected]>
newren · Jul 3, 2024 · 2800bcc · 2800bcc
1 parent 5c54a53
commit 2800bcc
Show file tree

Hide file tree

Showing 3 changed files with 49 additions and 5 deletions.
diff --git a/contrib/filter-repo-demos/clean-ignore b/contrib/filter-repo-demos/clean-ignore
@@ -29,7 +29,7 @@ class CheckIgnores:
     self.ignored = set()
     self.okay = set()
 
-    cmd = 'git check-ignore --stdin --verbose --non-matching --no-index'
+    cmd = 'git check-ignore --stdin --verbose --non-matching --no-index -z'
     self.check_ignore_process = subprocess.Popen(cmd.split(),
                                                  stdin=subprocess.PIPE,
                                                  stdout=subprocess.PIPE)
@@ -46,13 +46,13 @@ class CheckIgnores:
       elif name in self.okay:
         continue
       else:
-        self.check_ignore_process.stdin.write(name+b'\n')
+        self.check_ignore_process.stdin.write(name+b'\0')
         self.check_ignore_process.stdin.flush()
-        result = self.check_ignore_process.stdout.readline().rstrip(b'\n')
-        (rest, pathname) = result.split(b"\t")
+        result = os.read(self.check_ignore_process.stdout.fileno(), 65535).rstrip(b'\0')
+        (source, linenum, pattern, pathname) = result.split(b"\0")
         if name != pathname:
           raise SystemExit("Error: Passed {} but got {}".format(name, pathname))
-        if rest == b'::':
+        if not source and not linenum and not pattern:
           self.okay.add(name)
         else:
           self.ignored.add(name)

diff --git a/t/t9391-filter-repo-lib-usage.sh b/t/t9391-filter-repo-lib-usage.sh
@@ -16,6 +16,8 @@ esac
 export PYTHONDONTWRITEBYTECODE=1
 export CONTRIB_DIR=$TEST_DIRECTORY/../contrib/filter-repo-demos
 
+DATA="$TEST_DIRECTORY/t9391"
+
 setup()
 {
 	git init $1 &&
@@ -203,4 +205,17 @@ test_expect_success 'lint-history' '
 	)
 '
 
+test_expect_success 'clean-ignore with emoji in filenames' '
+	test_create_repo clean-ignore &&
+	(
+		cd clean-ignore &&
+		git fast-import --quiet <$DATA/emoji-repo &&
+		git reset --hard &&
+		$CONTRIB_DIR/clean-ignore --force &&
+		printf ".gitignore\nfilename\n" >expect &&
+		git ls-files >actual &&
+		test_cmp expect actual
+	)
+'
+
 test_done
diff --git a/t/t9391/emoji-repo b/t/t9391/emoji-repo
@@ -0,0 +1,29 @@
+feature done
+blob
+mark :1
+data 8
+initial
+
+blob
+mark :2
+data 5
+lock
+
+blob
+mark :3
+data 11
+*.bak
+🔒
+
+reset refs/heads/master
+commit refs/heads/master
+mark :4
+author Little O. Me <[email protected]> 1535228562 -0700
+committer Little O. Me <[email protected]> 1535228562 -0700
+data 10
+My commit
+M 100644 :1 filename
+M 100644 :2 🔒
+M 100644 :3 .gitignore
+
+done