Re: [BUG?] read_directory(), core.ignorecase and pathspec

Nguyen Thai Ngoc Duy [off-list ref] writes:

I was looking at how read_directory() uses pathspec, and it seems that
simplify_away(), the function that cuts out directories early if
they're definitely outside given pathspec, does exact match (ie.
memcmp) regardless core.ignorecase. This means "git add -- '*.c'" may
not work as expected when core.ignorecase is on.

I don't think for '*.c' it would make any difference, but it is very
plausible that 'frotz/*.c' will cull 'Frotz/anything' as "never going to
match".

The whole simplifying thing in read_directory() will eventually be
replaced with real pathspec matching as we put more magic in pathspec.

That should be done not "as" but "before". Get the foundation right before
going fancier.

On Thu, May 12, 2011 at 11:04 PM, Junio C Hamano [off-list ref] wrote:

Nguyen Thai Ngoc Duy [off-list ref] writes:

quoted

I was looking at how read_directory() uses pathspec, and it seems that
simplify_away(), the function that cuts out directories early if
they're definitely outside given pathspec, does exact match (ie.
memcmp) regardless core.ignorecase. This means "git add -- '*.c'" may
not work as expected when core.ignorecase is on.

I don't think for '*.c' it would make any difference, but it is very
plausible that 'frotz/*.c' will cull 'Frotz/anything' as "never going to
match".

Right, it only checks prefix.

quoted

The whole simplifying thing in read_directory() will eventually be
replaced with real pathspec matching as we put more magic in pathspec.

That should be done not "as" but "before". Get the foundation right before
going fancier.

Right, again.
-- 
Duy