On 02/02/11 17:11, Gustaf Josefsson wrote:
Hello!
I just installed mediawiki on a new site. I've used it on other servers before and
never encountered this problem.
For some reason, i get "Error 330 (net::ERR_CONTENT_DECODING_FAILED): Unknown
error.", when i try to enter a new url (/index.php?title=asd). This seem to be
similar to the problem described here:
http://www.mediawiki.org/wiki/Help_talk:Starting_a_new_page#Page_Not_Found_…
I cannot for the life of me figured out what the problem is. I reinstalled from scratch,
without any customizations, so this has to be a server config problem. It's not my
server, but my guess it that it probably something I can fix by changing some php option
in my .htaccess.
Anyone have a good idea?
My site is at:
http://www.incrediblemusicmachine.se/
For some reason, the edit link returns pure HTML but mark it as encoded
with gzip. This confuse the browser which is expecting gzipped content.
I reproduce it in both Safari and Firefox. IE might be able to guess it.
$ curl -I -H 'Accept-Encoding: gzip'
'http://www.incrediblemusicmachine.se/index.php?title=FOOBAR'
HTTP/1.1 404 Not Found
Date: Wed, 02 Feb 2011 22:32:55 GMT
Server: Apache/1.3.37 (Unix) mod_gzip/1.3.26.1a mod_fastcgi/FSDATA-1.1
mod_jk/1.1.0 Embperl/2.0b8 mod_perl/1.29 PHP/4.4.3 mod_ssl/2.8.28
OpenSSL/0.9.8b
X-Powered-By: PHP/5.2.14
Content-language: en
Vary: Accept-Encoding,Cookie
Expires: Thu, 01 Jan 1970 00:00:00 GMT
Cache-Control: private, must-revalidate, max-age=0
Content-Encoding: gzip
X-Powered-By: PHP/4.4.3
Content-Type: text/html
I have noticed a strange character at the top of the output:
curl -H 'Accept-Encoding: gzip'
'http://www.incrediblemusicmachine.se/index.php?title=FOOBAR' | hexdump -C
00000000 0a 3c 21 44 4f 43 54 59 50 45 20 68 74 6d |.<!DOCTYPE htm|
^^^^^
Not really helpful, but at least it gives the cause.
However, the Main Page is delivered compressed without errors.
Check the files for a that newline before the <?php Have you tried to
manual edit any of them?
Although I find strange that a new line breaks the gziping process but
doesn't also fail the header() call.