On 17/03/2012 4:21 a.m., awarecons wrote:
It doesn't matter, `cause request is made so that available to Squid
is http://rutube.ru/player.swf? string only.
Wrong. That is merely what your logs are configured to record. Squid is
given the full path (HTTP) or none of it (HTTPS).
http://www.squid-cache.org/Doc/config/strip_query_terms/
Configure "strip_query_terms off" to log the URL Squid is being given.
There's nothing in
access.log pointing to .iflv-file is being requested by embedded
java-script. The only GET is about '.../player.swf?'
More efficient is:
acl GET method GET
acl rutube_url urlpath_regex -i /player.swf
* You omitted step (2). The above pattern will match the pattern
anywhere in the URL. For example: http://rutube.ru/docs/player?swf=boo.
This is inefficient, a lot of extra work is done scanning the whole path
in case there was a match further down.
* You omitted step (3) and (5). The "." character has special meaning in
regex.
http_access deny GET rutube_url
2012/3/16 Amos Jeffries:
On 17/03/2012 12:34 a.m., awarecons wrote:
acl rutube dstdomain .rutube.ru .vk.com .vkontakte.ru
acl rutube_streams rep_header Content-Type -i
application/x-shockwave-flash
http_reply_access deny rutube rutube_streams
acl rutube_Files url_regex -i \player.swf$
http_access deny rutube rutube_Files
But there is still question - how cut off .iflv instead of cutting
player.swf?
step 1) Use urlpath_regex instead of url_regex.
step 2) Start it with "^" to anchor it to the path beginning.
step 3) Add proper escaping on the URL dot. Making it "\."
step 4) Remove the "$" end-anchor which is saying only catch the request
when there is *no* query string.
step 5) find a basic how-to on regex patterns and start reading.
Amos
Amos
