[PATCH sb/diff-orderfile-config] diff test: reading a directory as a file need not error out
There is no guarantee that strbuf_read_file must error out for directories. On some operating systems (e.g., Debian GNU/kFreeBSD wheezy), reading a directory gives its raw content: $ head -c5 / | cat -A ^AM-|^_^@^L$ As a result, 'git diff -O/' succeeds instead of erroring out on these systems, causing t4056.5 orderfile is a directory to fail. On some weird OS it might even make sense to pass a directory to the -O option and this is not a common user mistake that needs catching. Remove the test. Signed-off-by: Jonathan Nieder jrnie...@gmail.com --- Hi, t4056 is failing on systems using glibc with the kernel of FreeBSD[1]: | expecting success: | test_must_fail git diff -O/ --name-only HEAD^..HEAD | | a.h | b.c | c/Makefile | d.txt | test_must_fail: command succeeded: git diff -O/ --name-only HEAD^..HEAD | not ok 5 - orderfile is a directory How about this patch? Thanks, Jonathan [1] https://buildd.debian.org/status/fetch.php?pkg=gitarch=kfreebsd-amd64ver=1%3A2.0~next.20140107-1stamp=1389379274 t/t4056-diff-order.sh | 4 1 file changed, 4 deletions(-) diff --git a/t/t4056-diff-order.sh b/t/t4056-diff-order.sh index 1ddd226..9e2b29e 100755 --- a/t/t4056-diff-order.sh +++ b/t/t4056-diff-order.sh @@ -68,10 +68,6 @@ test_expect_success POSIXPERM,SANITY 'unreadable orderfile' ' test_must_fail git diff -Ounreadable_file --name-only HEAD^..HEAD ' -test_expect_success 'orderfile is a directory' ' - test_must_fail git diff -O/ --name-only HEAD^..HEAD -' - for i in 1 2 do test_expect_success orderfile using option ($i) ' -- 1.8.5.1 -- To unsubscribe from this list: send the line unsubscribe git in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH sb/diff-orderfile-config] diff test: reading a directory as a file need not error out
Jonathan Nieder jrnie...@gmail.com writes: There is no guarantee that strbuf_read_file must error out for directories. On some operating systems (e.g., Debian GNU/kFreeBSD wheezy), reading a directory gives its raw content: $ head -c5 / | cat -A ^AM-|^_^@^L$ As a result, 'git diff -O/' succeeds instead of erroring out on these systems, causing t4056.5 orderfile is a directory to fail. On some weird OS it might even make sense to pass a directory to the -O option and this is not a common user mistake that needs catching. Remove the test. Signed-off-by: Jonathan Nieder jrnie...@gmail.com --- Hi, t4056 is failing on systems using glibc with the kernel of FreeBSD[1]: | expecting success: | test_must_fail git diff -O/ --name-only HEAD^..HEAD | | a.h | b.c | c/Makefile | d.txt | test_must_fail: command succeeded: git diff -O/ --name-only HEAD^..HEAD | not ok 5 - orderfile is a directory How about this patch? Sounds sensible. Thanks. Thanks, Jonathan [1] https://buildd.debian.org/status/fetch.php?pkg=gitarch=kfreebsd-amd64ver=1%3A2.0~next.20140107-1stamp=1389379274 t/t4056-diff-order.sh | 4 1 file changed, 4 deletions(-) diff --git a/t/t4056-diff-order.sh b/t/t4056-diff-order.sh index 1ddd226..9e2b29e 100755 --- a/t/t4056-diff-order.sh +++ b/t/t4056-diff-order.sh @@ -68,10 +68,6 @@ test_expect_success POSIXPERM,SANITY 'unreadable orderfile' ' test_must_fail git diff -Ounreadable_file --name-only HEAD^..HEAD ' -test_expect_success 'orderfile is a directory' ' - test_must_fail git diff -O/ --name-only HEAD^..HEAD -' - for i in 1 2 do test_expect_success orderfile using option ($i) ' -- To unsubscribe from this list: send the line unsubscribe git in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
