Re: [[PATCH v3] 1/2] [submodule] handle multibyte characters in name

Fredrik Gustafsson [off-list ref] writes:

... The
correct approach to solve the problem for all pathnames may be to use
"ls-files -z" and tell the Perl script that reads its output to read NUL
separated records by using $/ = "\0".

I've tentatively queued the attached without 2/2; the scriptlet is
small enough not to matter in an eventual rewrite, so it shouldn't
make a difference either way.

-- >8 --
From: Fredrik Gustafsson <redacted>
Subject: [PATCH] handle multibyte characters in name

Many "git submodule" operations do not work on a submodule at a path whose
name is not in ASCII.

This is because "git ls-files" is used to find which paths are bound to
submodules to the current working tree, and the output is C-quoted by default
for non ASCII pathnames.

Tell "git ls-files" to not C-quote its output, which is easier than unwrapping
C-quote ourselves.

Signed-off-by: Fredrik Gustafsson <redacted>
Signed-off-by: Junio C Hamano <redacted>
---
 git-submodule.sh           |  3 ++-
 t/t7400-submodule-basic.sh | 12 ++++++++++++
 2 files changed, 14 insertions(+), 1 deletion(-)

diff --git a/git-submodule.sh b/git-submodule.sh
index 79bfaac..48bdf84 100755
--- a/git-submodule.sh
+++ b/git-submodule.sh

@@ -113,7 +113,7 @@ resolve_relative_url ()
 module_list()
 {
 	(
-		git ls-files --error-unmatch --stage -- "$@" ||
+		git ls-files -z --error-unmatch --stage -- "$@" ||
 		echo "unmatched pathspec exists"
 	) |
 	perl -e '

@@ -121,6 +121,7 @@ module_list()
 	my ($null_sha1) = ("0" x 40);
 	my @out = ();
 	my $unmatched = 0;
+	$/ = "\0";
 	while (<STDIN>) {
 		if (/^unmatched pathspec/) {
 			$unmatched = 1;

diff --git a/t/t7400-submodule-basic.sh b/t/t7400-submodule-basic.sh
index ff26535..d5743ee 100755
--- a/t/t7400-submodule-basic.sh
+++ b/t/t7400-submodule-basic.sh

@@ -868,4 +868,16 @@ test_expect_success 'submodule deinit fails when submodule has a .git directory
 	test -n "$(git config --get-regexp "submodule\.example\.")"
 '
 
+test_expect_success 'submodule with strange name works "å äö"' '
+	mkdir "å äö" &&
+	(
+		cd "å äö" &&
+		git init &&
+		touch sub
+		git add sub
+		git commit -m "init sub"
+	)
+	git submodule add "/å äö" &&
+	test -n "$(git submodule | grep "å äö")"
+'
 test_done

-- 
1.8.3.1-538-gb4d04a7

On Fri, Jun 14, 2013 at 10:23:52AM -0700, Junio C Hamano wrote:

Fredrik Gustafsson [off-list ref] writes:

quoted

... The
correct approach to solve the problem for all pathnames may be to use
"ls-files -z" and tell the Perl script that reads its output to read NUL
separated records by using $/ = "\0".

I've tentatively queued the attached without 2/2; the scriptlet is
small enough not to matter in an eventual rewrite, so it shouldn't
make a difference either way.

Sorry, I didn't knew enough perl to understand that that was a
suggestion rather than a hint to a future developer.

Now when I see how you meant it's looks like the best solution to me.
To me it looks like we now should be able to handle the multiline case
here. However, git submodule add doesn't handle newline yet, so it
really doesn't matter for now.

Thanks for the help!

-- 
Med vänliga hälsningar
Fredrik Gustafsson

tel: 0733-608274
e-post: iveqy@iveqy.com