Bug 5823 – REGRESSION: text at www.zverskienovosti.ru no longer decoded properly

• Status: VERIFIED
• Resolution: FIXED
• Component: DOM

[Alexey Proskuryakov]

Because of what looks like an incorrect encoding being used, the main text (above photographs) at this 
page is no longer displayed properly in ToT.

Interestingly, the text in sidebar (that is displayed just fine in ToT), was broken in Safari 2.0.2.

[Alexey Proskuryakov]

This looks like a duplicate of bug 5548 (and results in a regression because of other good changes). But 
before closing it as such, I want to re-check (currently, it's a bit too late here to be sure of anything :) ).

Note that this behavior depends on the document being gzipped during transfer, so the headers have to 
be examined like this:

curl -I --header "Accept-Encoding: gzip" "http://www.zverskienovosti.ru/index.php?
option=content&task=view&id=382&Itemid=2"

[Alexey Proskuryakov]

Yes, the patch in bug 5548 does contain a fix for this. However, I don't think the bug itself is a duplicate, 
so I'll just add a dependency, and mark fixed after the patch for bug 5548 gets landed.

Here, the encoding gets wrong because it's an XHTML file transferred as text/html with an http-equiv 
meta, and without an explicit encoding in the XML encoding declaration. With a fix to bug 3809, WebKit 
now defaults to UTF-8, and a fix to bug 5548 allows meta to override the default. The problem with side 
bar was fixed in bug 5140.