W3C home > Mailing lists > Public > public-qa-dev@w3.org > April 2008

Re: Outline feature in validator (was: profiling the Markup Validator with Devel::NYTProf)

From: Dominique Hazael-Massieux <dom@w3.org>
Date: Mon, 07 Apr 2008 13:32:04 +0200
To: olivier Thereaux <ot@w3.org>
Cc: public-qa-dev hacking list <public-qa-dev@w3.org>
Message-Id: <1207567924.5148.337.camel@localhost>

Le lundi 07 avril 2008 à 07:30 -0400, olivier Thereaux a écrit :
> On Apr 7, 2008, at 02:56 , Dominique Hazael-Massieux wrote:
> > (probably not once per character, except in worse case scenarii; in
> > general, while there is no guarantee that characters be all stuck
> > together in the data call in the SAX API, it doesn't imply that each
> > character generates such a call either)
> 
> Note that the name is misleading. It's really an opensp API handler,  
> which looks like SAX.

Oh, OK, I was clearly mistaken then :)

> > Err... Is this API call done even when the show outline option isn't
> > enabled? If so, the best solution might be to make sure it isn't  
> > called
> > unless it is enabled, no?

> The way the API works is to define a handler module with a subroutine  
> for each sort of event, and short of making several handler modules,  
> one with data subroutine and one without (and we can imagine other  
> branching situations) I don't think it's possible to just say "don't  
> use your data routine for data events".

Well, if the cost of using a single handler is so high, I would
definitely recommend defining several handler modules; I expect one can
inherit from another, and thus, there wouldn't be any code duplication,
right?

Dom
Received on Monday, 7 April 2008 11:34:06 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 19 August 2010 18:12:49 GMT