Patch to XMLHttpRequest and decoder, auto detects scripts charset a little bit better

[Fredrik Johansson]

Hi !

Triggered by the quickness that my other patches got merged into branch, I 
thought I cook up a new one :-)

Issues adressed:
1. When using XMLHttpRequest from javascript a http status 304 is converted to 
200 on a async call but not on a sync call. 
IE and moz converts both to 200

2. When using XMLHttpRequest and the server sends a Content-type: some/mime; 
charset=something, it still didnt find the charset.

3. When using XMLHttpRequest to grab only text (such as scripts) and it comes 
from cache, there are no Content-type header but there might be a charset 
info from KIOSlaves, I grabbed that info, if there still is a Content-type 
with charset info it overrules the one from KIOSlaves

4. Sometimes a server doesnt send charset info in Content-type and there is no 
clue in the content, such as a scripts, and the file has no BOM bytes, 
decoder goes for latin1. So UTF-8 string gets messed up.
If the autodetection has failed and we are about to default for latin1, I 
included a last minute UTF-8 check. If that fails then it goes back to latin1
This issue has a bug ticket: 
http://bugs.kde.org/show_bug.cgi?id=130234


I also included the test files I used.
By the way, the test dir in the khtml dir, is that the only tests used by 
khtml?
Seems kind of few tests for a large project like khtml.

Anyways I wonder if anybody would  want to take a look at patch

Regards
Fredrik Johansson
-------------- next part --------------
A non-text attachment was scrubbed...
Name: charset.patch.tar.gz
Type: application/x-tgz
Size: 3855 bytes
Desc: not available
URL: <https://mail.kde.org/mailman/private/kfm-devel/attachments/20060709/4e41c7a7/attachment.bin>