From: Johan Herland <johan@herland.net>
Date: Tue, 23 Sep 2008 02:05:04 GMT
Subject: Re: [PATCH] diff funcname_pattern: Allow HTML header tags without attributes
Message-ID: <200809230405.04471.johan@herland.net>
In-Reply-To: <7v7i93ws64.fsf@gitster.siamese.dyndns.org>

On Tuesday 23 September 2008, Junio C Hamano wrote:
> Johan Herland <johan@herland.net> writes:
> > After looking over this once more, I think the HTML regexp should be
> > changed as follows. This fixes a buglet that was part of my original
> > HTML pattern, and although this patch textually depends on Brandon's
> > work, it is conceptually independent of his refactorization.
> > ...
> > -	{ "html", "^[ \t]*(<[Hh][1-6][ \t].*>.*)$", REG_EXTENDED },
> > +	{ "html", "^[ \t]*(<[Hh][1-6]([ \t].*)?>.*)$", REG_EXTENDED },
>
> I do not think these two particularly would make much difference.  Why
> isn't it simply...
>
> 	"<[Hh][1-6].*"
>
> without even any capture or anchor?
>
> It would falsely hit oddball cases like <h1foo> which is not <h1>, but
> anybody who uses such a nonstandard thing deserves it, imnvho ;-).

Ok. I agree. Go ahead.


...Johan

-- 
Johan Herland, <johan@herland.net>
www.herland.net