593809 - entity specified crlf in the <pre> is rendered as two newlines

Status: RESOLVED FIXED

(Core :: Layout, defect)

Description

[al_9x]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.8) Gecko/20100722 Firefox/3.6.8
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.2.8) Gecko/20100722 Firefox/3.6.8

<pre>line1&#13;&#10;line2</pre>



Reproducible: Always

Comment 1

[Boris Zbarsky [:bzbarsky]]

Fixed by the HTML5 parser.  I suggest trying a Firefox 4 beta.  ;)

Comment 2

[al_9x]

(In reply to comment #1)
> Fixed by the HTML5 parser.  I suggest trying a Firefox 4 beta.  ;)

These newlines are used in Hotmail "view message source," so this affects many, would be nice if it worked in 3.6

Comment 3

[Boris Zbarsky [:bzbarsky]]

This isn't a security or stability bug, nor a regression (the old HTML parser has behaved this way for literally over a decade), so a fix wouldn't be accepted by the branch drivers even if someone were to create one.  Rightly so: it's a stable branch precisely because the scope of the changes that happen on it is limited.  Changes to HTML parsing are way out of scope.

Comment 4

[al_9x]

(In reply to comment #3)
> Changes to HTML parsing are way out of scope.

HTML parsing and layout is as core functionality as it gets for a browser.  If it doesn't work properly one would think you'd want to fix it, especially since it doesn't require a rearchitecting of the layout engine and is only a minor tweak.  Adding out of process plugins, for instance, was likely far more invasive and destabilizing, and it wasn't even a bug fix.

Comment 5

[Boris Zbarsky [:bzbarsky]]

> If it doesn't work properly one would think you'd want to fix it

Yes.  But there's a difference between "fix" and "backport the fix to stable releases".  The latter doesn't happen for typical parsing/layout bugs.  Again, backports happen for security or some stability fixes or sometimes fixes for things that first broke with the release of the stable branch.

> especially since it doesn't require a rearchitecting of the layout engine and
> is only a minor tweak. 

That's where you're wrong.  Fixing this particular issue requires major changes to the old parser.  Trust me, I've looked into it in the past.

Out of process plugins was a stability fix; it significantly reduced the number of crashes users saw.

Comment 6

[al_9x]

(In reply to comment #5)
> > is only a minor tweak. 
> 
> That's where you're wrong.  Fixing this particular issue requires major changes
> to the old parser.  Trust me, I've looked into it in the past.
> 
Well, I guess that's a valid reason, not to touch it.

> Out of process plugins was a stability fix;

It was absolutely not a bug fix, it was a new feature that improved browser stability by containing plug-in crash bugs.  So it's only logical that if you can introduce new features to improve stability in peripheral functionality, you should not reflexively reject fixing bugs (backporting) in core functionality. Expectation of correct core behavior is the primary reason one uses any given software.

Comment 7

[Boris Zbarsky [:bzbarsky]]

> improve stability in peripheral functionality

Anything that causes more than half the crashes our users experience is not "peripheral functionality" pretty much by definition.

And, as I'm sure you're aware, the decision to backport OOPP was not taken lightly.  Neither are decisions about rendering bugs; if the fix is very safe and the payoff is large enough (e.g. serious compat problems with a major site), they can be backported.  But the fix here would not be safe, so the whole thing is academic.