Re: $r->document_root

[prev] [thread] [next] [Date index for 2005/02/19]

From: Geoffrey Young
Subject: Re: $r->document_root
Date: 13:54 on 19 Feb 2005


> So unless I miss something we have two options:
> 
> 1) nuke that method completely
> 
> 2) either provide a new method that restores the data using a
> preallocated string from the server pool or do that internally in the
> server
> 
> 3) or change $r->document_root method to accept the pool object and
> leave the users the headache of figuring out what they want to do. This
> is because if they use $r->pool and expect it stick beyond that request
> it won't work. If however they pass $r->server->pool and they do that on
> each request they leak memory.
> 
> Comments?

take a look at the implementation in mp1 - why won't something like that
work, so long as we're in prefork?

--Geoff

(message missing)

$r->document_root
Cahill, Earl 23:14 on 14 Feb 2005

Re: $r->document_root
Geoffrey Young 23:21 on 14 Feb 2005

Re: $r->document_root
Geoffrey Young 23:26 on 14 Feb 2005

RE: $r->document_root
Cahill, Earl 20:43 on 15 Feb 2005

Re: $r->document_root
Geoffrey Young 21:10 on 15 Feb 2005

Re: $r->document_root
Stas Bekman 01:48 on 19 Feb 2005

Re: $r->document_root
Stas Bekman 02:02 on 19 Feb 2005

Re: $r->document_root
Geoffrey Young 13:54 on 19 Feb 2005

Re: $r->document_root
Stas Bekman 17:21 on 19 Feb 2005

Re: $r->document_root
Joe Schaefer 14:12 on 19 Feb 2005

Re: $r->document_root
Geoffrey Young 05:26 on 20 Feb 2005

Re: $r->document_root
Stas Bekman 01:29 on 21 Feb 2005

Re: $r->document_root
Joe Schaefer 02:38 on 21 Feb 2005

Re: $r->document_root
Stas Bekman 02:43 on 21 Feb 2005