From: Tom Grennan <tmgrennan@gmail.com>
Date: Thu, 23 Feb 2012 00:59:32 GMT
Subject: Re: [PATCHv3 1/5] refs: add match_pattern()
Message-ID: <20120223005932.GF2410@tgrennan-laptop>
In-Reply-To: <7v62ey8ktp.fsf@alter.siamese.dyndns.org>

On Wed, Feb 22, 2012 at 04:17:22PM -0800, Junio C Hamano wrote:
>Tom Grennan <tmgrennan@gmail.com> writes:
>
>> Yes, I tried to minimize change but the conversion of argv's to
>> string_list's won't add that much.
>
>How about _not_ using string_list?  After all, string_list is not just a
>collection of strings, but is a table to hold strings with attributes.  I
>thought argv_array is more appropriate abstraction for the purpose of your
>patch.

OK, It looks like I should also add a common parse_opt_argv_array() to
parse-options-cb.  Of course that would be in a separate, dependent
patch.

>>>	ref_match_pattern("refs/heads/master",
>>>        		  ["maste?", NULL],
>>>                          ["refs/heads/", NULL],
>>>                          (REF_MATCH_FNM_PATH|REF_MATCH_LEADING));
>>>
>>>Note that the above "ref_match_pattern()" gives the same "flags" for the
>>>call to match_one() for elements in both positive and negative array and
>>>it is very deliberate.  See review comment to [3/5] for the reasoning.
>>
>> OK, I think that I understand, but please confirm, you'd expect no output in
>> the above example, right?
>
>"maste?" would match with FNM_PATHNAME with "refs/heads/master" but
>the negative "refs/heads/" matches with it, so yeah, I expect that the
>function would return false.

thanks, that's a good test case.

-- 
TomG