From: Junio C Hamano <gitster@pobox.com>
Date: Sat, 11 Feb 2012 23:43:34 GMT
Subject: Re: [PATCHv2 1/4] refs: add common refname_match_patterns()
Message-ID: <7vzkcpkkbd.fsf@alter.siamese.dyndns.org>
In-Reply-To: <20120211193742.GD4903@tgrennan-laptop>

Tom Grennan <tmgrennan@gmail.com> writes:

> Yes, I should have stated that this emphasized containment over
> efficiency.  If instead we stipulate that the caller must list exclusion
> patterns before others, this could simply be:

No.

You have to pre-parse and rearrange the pattern[] list *only once* before
matching them against dozens of refs, so instead of forcing the callers do
anything funky, you give a function that gets a pattern[] list and returns
something that can be efficiently used by the match_pattern() function,
and have the caller pass that thing, not the original pattern[] list, to
the match_pattern() function.

That is how pathspec matching side of the logic is arranged.

I keep saying that it is probably not a good idea to directly reuse the
pathspec code, but you would want to study and learn from the overall
structure of it.