httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Neil A. Hillard" <hilla...@whl.co.uk>
Subject Re: [users@httpd] mod_proxy/mod_proxy_html
Date Tue, 16 Aug 2005 14:52:36 GMT
Hi,

Nick Kew wrote:
> On Tuesday 16 August 2005 14:45, Axel-St├ęphane SMORGRAV wrote:
> 
>>Nick,
>>
>>I submitted the three votes I thought you needed last time this issue came
>>up.
> 
> 
> No, this is different.  Patching a release marked "stable" (i.e. 2.0) requires
> 3 votes from apache committers.  That requires two committers (other than
> myself) to review the code and agree it works (or, crucially, at least doesn't
> break things that work now:-).  With 2.1/2.2, that's not marked stable, and
> I've already committed the patch for that.
> 
> 
>>With respect to the "ProxyPassReverse in Location" problem, I do not really
>>see the relationship to ProxyPassReverseCookie(Domain|Path) unless we are
>>talking about the same patch ??
> 
> The patch I'm now proposing is the same.  Both issues should be fixed, and 
> it's easier to patch both at once than do them separately.
> 
>>Are we talking about backporting your patch (id=11917) or (id=15543) ? I
>>tested the former, not the latter.
> 
> This'll be the patch I posted to the dev mailinglist recently.  See the
> post I referenced.

On the ProxyPassReverse in Location bug - this was a show-stopper, end 
of story, for us.  If Nick hadn't fixed this (under consultancy) we'd 
have been looking at products other than Apache.  It only shows itself 
when there's 'broken' code on the server issuing redirects to partial 
URLs.  I'm sure we know that that's not allowed (according to the HTTP 
standards), but people still do it (including manufacturers of the 
commercial software we're using).  Making ProxyPassReverse work within a 
'Location' resolves this problem and as Nick said we've been using it in 
production since January and not had any problems.

On the cookie rewriting - initially we weren't worried about this as we 
didn't see any need for cookie rewriting.  Just last week I had to add 
another backend server behind our accelerator and it had a path (other 
than '/') in the cookie path.  I then tried applying the patch for bug 
#10722 and needless to say it wouldn't apply cleanly (with the 
ProxyPassReverse patch that Nick supplied) !!!  Either way around, the 
two patches collided.  Obtaining the patch didn't seem overly easy 
either, as it's split into three parts (the patch itself and two 
documentation patches).

Given the revised patch that Nick supplied that resolves both of these 
issues we're once again running things happily in production.

It would make life a lot easier if this was part of the main code. 
Please vote for it !!!


                 Neil.

-- 
Neil Hillard                    hillardn@whl.co.uk
Westland Helicopters Ltd.       http://www.whl.co.uk/

Disclaimer: This message does not necessarily reflect the
             views of Westland Helicopters Ltd.

---------------------------------------------------------------------
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: users-unsubscribe@httpd.apache.org
   "   from the digest: users-digest-unsubscribe@httpd.apache.org
For additional commands, e-mail: users-help@httpd.apache.org


Mime
View raw message