Signed-off-by: Matthew Ogilvie <mmogilvi_git@miniinfo.net> ---
I've been working on adding crlf attribute support for git-cvsserver, and noticed this little issue. I've included some tests and a fix. The cvsserver stuff is still incomplete and not included.
There are a couple of things I'm not sure about, but seem to make the most sense the way I coded it:
1. A test of the form "attr_check /a/i a/i" (notice the prefixed '/' in the pathname) will "pass" without this patch, but will not "pass" with this patch. This seems slightly more correct, but I'm not sure.
2. The first (non-setup) test in t0003-attributes.sh actually batches up several checks as one test. I figured there should be some limit on how many individual checks get batched into one, so I started a second test, but I could see an argument to just include the new checks in the existing batch.
attr.c | 4 +++- t/t0003-attributes.sh | 8 ++++++++ 2 files changed, 11 insertions(+), 1 deletions(-)
Show changes to 2 files +11 −1
attr.c, t/t0003-attributes.sh
diff --git a/attr.c b/attr.c index 64b77b1..1a15fad 100644 --- a/attr.c +++ b/attr.c @@ -546,7 +546,9 @@ static int path_matches(const char *pathname, int pathlen, (baselen && pathname[baselen] != '/') || strncmp(pathname, base, baselen)) return 0; - return fnmatch(pattern, pathname + baselen + 1, FNM_PATHNAME) == 0; + if (baselen != 0) + baselen++; + return fnmatch(pattern, pathname + baselen, FNM_PATHNAME) == 0; } static int fill_one(const char *what, struct match_attr *a, int rem) diff --git a/t/t0003-attributes.sh b/t/t0003-attributes.sh index 3faf135..c56d2fb 100755 --- a/t/t0003-attributes.sh +++ b/t/t0003-attributes.sh @@ -21,6 +21,7 @@ test_expect_success 'setup' ' mkdir -p a/b/d a/c && ( echo "f test=f" + echo "a/i test=a/i" ) >.gitattributes && ( echo "g test=a/g" && @@ -46,4 +47,11 @@ test_expect_success 'attribute test' ' ' +test_expect_success 'root subdir attribute test' ' + + attr_check a/i a/i && + attr_check subdir/a/i unspecified + +' + test_done
-- 1.5.5.1.57.g5909c