Re: [PATCH 2/4] pathspec: do exact comparison on the leading non-wildcard part

[Nguyen Thai Ngoc Duy]

On Tue, Nov 20, 2012 at 3:54 AM, Junio C Hamano  wrote:
> How would we protect this optimization from future breakages?
>
> Once we start using FNM_PERIOD, this becomes unsafe, as the simple
> part in "foo/bar*baz" would be "foo/bar" with remainder "*baz".
>
> The pattern "foo/bar*baz" should match "foo/bar.baz" with FNM_PERIOD
> set.  But with this optimization, fnmatch is fed pattern="*baz",
> string=".baz" and they would not match.

The first question is how FNM_PERIOD comes in the first place without
breaking the current syntax. I guess we just turn off the optimization
if FNM_PERIOD is used.

My secret "plan" is to convert all fnmatch() calls to git_fnmatch()
then replace fnmatch() with wildmatch() and move these optimization
down to wildmatch(). I think we can handle FNM_PERIOD case better down
there because string is still "foo/bar.baz".
-- 
Duy
--
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 2/4] pathspec: do exact comparison on the leading non-wildcard part

[Junio C Hamano]

Nguyễn Thái Ngọc Duy   writes:

> Signed-off-by: Nguyễn Thái Ngọc Duy 
> ---
>  dir.c   | 18 +-
>  dir.h   |  8 
>  tree-walk.c |  6 --
>  3 files changed, 29 insertions(+), 3 deletions(-)
>
> diff --git a/dir.c b/dir.c
> index c391d46..e4e6ca1 100644
> --- a/dir.c
> +++ b/dir.c
> @@ -34,6 +34,21 @@ int fnmatch_icase(const char *pattern, const char *string, 
> int flags)
>   return fnmatch(pattern, string, flags | (ignore_case ? FNM_CASEFOLD : 
> 0));
>  }
>  
> +inline int git_fnmatch(const char *pattern, const char *string,
> +int flags, int prefix)
> +{
> + int fnm_flags = 0;
> + if (flags & GF_PATHNAME)
> + fnm_flags |= FNM_PATHNAME;
> + if (prefix > 0) {
> + if (strncmp(pattern, string, prefix))
> + return FNM_NOMATCH;
> + pattern += prefix;
> + string += prefix;
> + }
> + return fnmatch(pattern, string, fnm_flags);

How would we protect this optimization from future breakages?

Once we start using FNM_PERIOD, this becomes unsafe, as the simple
part in "foo/bar*baz" would be "foo/bar" with remainder "*baz".

The pattern "foo/bar*baz" should match "foo/bar.baz" with FNM_PERIOD
set.  But with this optimization, fnmatch is fed pattern="*baz",
string=".baz" and they would not match.
--
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