Joshua Slive wrote:
>On 11/17/05, Phil Endecott wrote:
>> Hi Joshua, thanks for the reply. Yes, when I comment-out my
>> ErrorDocument directives I get what I had expected:
>> My ErrorDocuments are the standard ones from the Debian package. Is
>> there anything I can do to fix this? Presumably it's a case of
>> appending rather than setting the Vary header somewhere. If not
>> I'll just stick with the default messages.
> How is your Header directive scoped? Does it work if you apply the
> directive specifically to the error documents directory?
Thanks Joshua, I think I have something that works now.
My ErrorDocuments are declared in the global configuration, while my
Header directive was in the main <Directory> section for one virtual
host. I've copied the Header line into the <Directory> section for the
/error pages and now it seems to work:
1 HTTP/1.1 403 Forbidden
2 Date: Sat, 19 Nov 2005 12:46:34 GMT
3 Server: Apache/2.0
4 Vary: accept-language,accept-charset,X-moz
5 Accept-Ranges: bytes
6 Connection: close
7 Content-Type: text/html; charset=iso-8859-1
8 Content-Language: en
9 Expires: Sat, 19 Nov 2005 12:46:34 GMT
12:46:34 ERROR 403: Forbidden.
I think this is acceptable, though it would be more correct to add the
extra token to Vary: only when the error has come from the one virtual
host. The only way I can think of doing this is to replicate all of the
ErrorDoc stuff (and copy/symlink the actual error pages?) for each
virtual host.
Perhaps the more important issue is that the "always" option to the
Header directive could be better named/documented.
BTW, this is an attempt to defeat the "FasterFox" mozilla extension
which scans pages as you load them and starts fetching all linked-to
pages in the background. One of my sites got a mention on Digg earlier
in the week, with two consequences. First, every FasterFox-using Digg
viewer downloaded my site's front page, whether or not they actually
clicked on it, and those who did click on it downloaded essentially the
entire site rather than just the one page they would look at. My guess
is that this inflated the resources used by a factor of 5 to 10, and
combined with some other issues took the site down for several hours.
Anyone with a site that might get linked to unexpectedly from a popular
blog/news site should consider doing something like this. Of course it
has the disadvantage of also blocking more legitimate uses of
prefetching, such as Google.
Cheers,
--Phil.
---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
" from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
