| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
Now it behaves as HTTP. Also fixed a bug in dillo with lingering cache
entries (when the file transfer was already done at abort time).
|
|
When a redirection page comes in, obviously we haven't followed the
redirection yet, so we can't find the entry. This meant that the
content type from the HTTP header could not be set.
This was not so great for my new error message.
Maybe circumstances where we'd want to follow the redirection here will
reveal themselves, but I'd be surprised.
|
|
|
|
|
|
|
|
|
|
|
|
Was triggered by the URL data:text/plain;charset=iso-8859-2,%B5%B5
because dillo does not currently think that's likely to be text.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
e.g. some links at http://git.kernel.org/gitweb.cgi didn't render.
e.g. #2 This page didn't render:
<html>
<head>
<meta http-equiv="content-type" content="; charset=utf-8"/>
<title></title>
</head>
<body>
Generating....
</body>
</html>
|
|
|
|
Race condition example: www.blisty.cz
charset after CSS ex. : http://www.daemonnews.org
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This patch gets rid of a series of valgrind complains with this page:
http://selenic.com/pipermail/mercurial/
|
|
|
|
|
|
|
|
- Added use of inttypes.h when stdint.h isn't found.
|
|
|
|
|
|
|
|
|
|
- Cleaned up a_Url_new().
|
|
|
|
|
|
|
|
|
|
|
|
- Fixed data guesser to detect ASCII, LATIN1, UTF8, KOI8-R, CP-1251 as text.
|
|
|
|
|
|
|
|
|
|
|
|
|
