Commit 02c5631 (difftool --dir-diff: symlink all files matching the
working tree, 2013-03-14) does not handle the case where a file that is
being compared does not exist in the working tree. Fix this by checking
for existence explicitly before running git-hash-object.
Reported-by: Kevin Bracey <kevin@xxxxxxxxx>
Signed-off-by: John Keeping <john@xxxxxxxxxxxxx>
---
On Fri, May 17, 2013 at 11:10:40AM -0700, Junio C Hamano wrote:
> John Keeping <john@xxxxxxxxxxxxx> writes:
>
> > Commit 02c5631 (difftool --dir-diff: symlink all files matching the
> > working tree, 2013-03-14) does not handle the case where a file that is
> > being compared does not exist in the working tree. Fix this by checking
> > for existence explicitly before running git-hash-object.
> >
> > Reported-by: Kevin Bracey <kevin@xxxxxxxxx>
> > Signed-off-by: John Keeping <john@xxxxxxxxxxxxx>
> > ---
> > This fixes a regression in 1.8.3-rc0.
> >
> > git-difftool.perl | 9 ++++++++-
> > t/t7800-difftool.sh | 7 +++++++
> > 2 files changed, 15 insertions(+), 1 deletion(-)
> >
> > diff --git a/git-difftool.perl b/git-difftool.perl
> > index 6780292..0a1cb0a 100755
> > --- a/git-difftool.perl
> > +++ b/git-difftool.perl
> > @@ -92,7 +92,14 @@ sub use_wt_file
> > return 0;
> > }
> >
> > - my $wt_sha1 = $repo->command_oneline('hash-object', "$workdir/$file");
> > + my $wt_sha1;
> > + if (-e "$workdir/$file") {
> > + $wt_sha1 = $repo->command_oneline('hash-object', "$workdir/$file");
> > + } else {
> > + # If the file doesn't exist in the working tree, use something
> > + # that cannot match a SHA-1.
> > + $wt_sha1 = '';
>
> Yuck.
>
> "that cannot match" might be a good justification to say "this does
> not break the next line to set $use and forces it to false", but
> "when we return false in $use, the value of $wt_sha1 is not used"
> needs to be said to convince why this is a safe change.
>
> But if $sha1 is $null_sha1, we do end up setting $use to true and
> the caller would stuff the empty $wt_sha1 to form:
>
> $wtindex .= "$rmode \$dst_path\0";
>
> Is that what we want to do here, or is it a "will never happen"
> condition? If the latter, the reason need to be described in this
> comment (and in the log).
It can't ever happen because we only call this if the mode is non-zero
in which case the SHA-1 is only null if there are unstaged changes.
However, I think this revised version is much clearer.
git-difftool.perl | 6 ++++++
t/t7800-difftool.sh | 7 +++++++
2 files changed, 13 insertions(+)
diff --git a/git-difftool.perl b/git-difftool.perl
index 6780292..8a75205 100755
--- a/git-difftool.perl
+++ b/git-difftool.perl
@@ -92,6 +92,12 @@ sub use_wt_file
return 0;
}
+ if (! -e "$workdir/$file") {
+ # If the file doesn't exist in the working tree, we cannot
+ # use it.
+ return (0, $null_sha1);
+ }
+
my $wt_sha1 = $repo->command_oneline('hash-object', "$workdir/$file");
my $use = ($sha1 eq $null_sha1) || ($sha1 eq $wt_sha1);
return ($use, $wt_sha1);
diff --git a/t/t7800-difftool.sh b/t/t7800-difftool.sh
index a6bd99e..d46f041 100755
--- a/t/t7800-difftool.sh
+++ b/t/t7800-difftool.sh
@@ -356,6 +356,13 @@ run_dir_diff_test 'difftool --dir-diff from subdirectory' '
)
'
+run_dir_diff_test 'difftool --dir-diff when worktree file is missing' '
+ test_when_finished git reset --hard &&
+ rm file2 &&
+ git difftool --dir-diff $symlinks --extcmd ls branch master >output &&
+ grep file2 output
+'
+
write_script .git/CHECK_SYMLINKS <<\EOF
for f in file file2 sub/sub
do
--
1.8.3.rc2.285.gfc18c2c
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html