On Fri, Feb 19, 2010 at 01:02:49AM -0500, Jeff King wrote:
> But this is a little disturbing still:
>
> $ git init
> $ mkdir dir
> $ touch dir/sub
> $ touch root
> $ echo dir >.gitignore
> $ echo root >>.gitignore
>
> $ git add root
> The following paths are ignored by one of your .gitignore files:
> root
> Use -f if you really want to add them.
> fatal: no files added
> $ echo $?
> 128
>
> $ git add dir
> The following paths are ignored by one of your .gitignore files:
> dir
> Use -f if you really want to add them.
> fatal: no files added
> $ echo $?
> 128
>
> $ git add dir/sub
> $ echo $?
> 0
>
> but we didn't actually add the file.
Junio,
This seems to be caused by dir.c:treat_one_path. In the first few lines:
int exclude = excluded(dir, path, &dtype);
if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
&& in_pathspec(path, *len, simplify))
dir_add_ignored(dir, path, *len);
we see that the prefix "dir" is excluded, but it is not in our pathspec
("dir/sub"), so we do not add it to the ignored list.
This is related to your recent 48ffef9 (ls-files: fix overeager pathspec
optimization, 2010-01-08), as before then we actually didn't consider
"dir/sub" to be ignored at all. The in_pathspec check did not originate
there; it's from my e96980e (builtin-add: simplify (and increase
accuracy of) exclude handling, 2007-06-12). But it is definitely still
necessary.
I'm not sure of the right way to fix this. We can drop further down into
the directory hierarchy when doing COLLECT_IGNORED and look for actual
files, but that may have a negative performance impact. Perhaps we can
go further only if we are a prefix of a pathspec. Or maybe there is some
way to be more clever.
I dunno. I'm out of ideas for the evening, and since you looked at this
not too long ago, I thought you might have some insight.
-Peff
--
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