W3C home > Mailing lists > Public > www-style@w3.org > January 2010

RE: Text selector [was Re: breaking overflow]

From: Belov, Charles <Charles.Belov@sfmta.com>
Date: Fri, 8 Jan 2010 13:03:03 -0800
Message-ID: <E17F75B6E86AE842A57B4534F82D0376659FBB@MTAMAIL.muni.sfgov.org>
To: "Tab Atkins Jr." <jackalmage@gmail.com>
Cc: "www-style list" <www-style@w3.org>
> -----Original Message-----
> From: Tab Atkins Jr. [mailto:jackalmage@gmail.com] 
> Sent: Friday, January 08, 2010 11:36 AM
> To: Belov, Charles
> Cc: www-style list; Boris Zbarsky; Brad Kemper; Aryeh Gregor; 
> James Hopkins
> Subject: Re: Text selector [was Re: breaking overflow]
> 
> On Fri, Jan 8, 2010 at 1:17 PM, Belov, Charles 
> <Charles.Belov@sfmta.com> wrote:
> > -----Original Message-----
> > From: www-style-request@w3.org [mailto:www-style-request@w3.org] On 
> > Behalf Of Aryeh On Mon, Jan 4, 2010 at 1:28 PM, James Hopkins 
> > <james@idreamincode.co.uk>
> > wrote:
> >>If a text node is split across two packets, apparently browsers will
> > split it into two >text nodes, at least sometimes.  
> Obviously, having 
> > behavior change noticeably because
> >>of that would be a really bad idea, since it's not feasible 
> to control
> > where packet
> >>boundaries are.
> >
> > Then one of the user agent requirements for supporting the ::text 
> > pseudoselector would be that if a style sheet in the 
> cascade used such 
> > a selector the browser would have to first patch together any text 
> > node that it had previously arbitrarily split.
> >
> > That would be a programming issue rather than a CSS issue, 
> but it is 
> > good to know that user-agent programmers would have to look for it.
> > There would need to be a way for programmers to intentionally 
> > construct such packets for testing purposes, but I would 
> imagine that is doable.
> 
> It shouldn't be necessary to test for this specifically.  A 
> split text node doesn't know *why* it was split, and neither 
> does the CSS engine attempting to match against it.  As long 
> as you can manufacture a split text node through some means, 
> such as DOM scripting, you should be good.

This is an example of where trimming of prior messages turns the 
thread into a game of Telephone.

The subject was inadvertent splitting of consecutive text into separate 
text nodes by browsers in response to the transport layer protocol 
splitting running text across multiple packets, thus potentially causing 
interference with the ability of a ::text selector to work properly.

Hope this helps,
Charles Belov
SFMTA Webmaster
Received on Friday, 8 January 2010 21:03:58 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:23 GMT