You are viewing a plain text version of this content. The canonical link for it is here.
Posted to cvs@httpd.apache.org by tr...@apache.org on 2011/04/30 14:22:24 UTC
svn commit: r1098105 - /httpd/httpd/branches/2.2.x/STATUS
Author: trawick
Date: Sat Apr 30 12:22:23 2011
New Revision: 1098105
URL: http://svn.apache.org/viewvc?rev=1098105&view=rev
Log:
vote, note
Modified:
httpd/httpd/branches/2.2.x/STATUS
Modified: httpd/httpd/branches/2.2.x/STATUS
URL: http://svn.apache.org/viewvc/httpd/httpd/branches/2.2.x/STATUS?rev=1098105&r1=1098104&r2=1098105&view=diff
==============================================================================
--- httpd/httpd/branches/2.2.x/STATUS (original)
+++ httpd/httpd/branches/2.2.x/STATUS Sat Apr 30 12:22:23 2011
@@ -124,7 +124,7 @@ PATCHES PROPOSED TO BACKPORT FROM TRUNK:
http://svn.apache.org/viewvc?view=revision&revision=937858
http://svn.apache.org/viewvc?view=revision&revision=938265
2.2.x patch: http://people.apache.org/~sf/408.diff
- +1: sf
+ +1: sf, trawick
* mod_cache: Realign the cache_quick_handler() to behave identically
to the default_handler() when reacting to errors when writing to the
@@ -132,6 +132,8 @@ PATCHES PROPOSED TO BACKPORT FROM TRUNK:
Trunk patches: http://svn.apache.org/viewvc?view=revision&revision=1003913
2.2.x patch: http://people.apache.org/~minfrin/httpd-mod_cache-errorfix-22.patch
+1: minfrin
+ trawick: any reason it shouldn't be completely aligned with default_handler's
+ choice to return OK vs. 500?
PATCHES/ISSUES THAT ARE STALLED
Re: svn commit: r1098105 - /httpd/httpd/branches/2.2.x/STATUS
Posted by Graham Leggett <mi...@sharp.fm>.
On 30 Apr 2011, at 2:22 PM, trawick@apache.org wrote:
> * mod_cache: Realign the cache_quick_handler() to behave identically
> to the default_handler() when reacting to errors when writing to
> the
> @@ -132,6 +132,8 @@ PATCHES PROPOSED TO BACKPORT FROM TRUNK:
> Trunk patches: http://svn.apache.org/viewvc?view=revision&revision=1003913
> 2.2.x patch: http://people.apache.org/~minfrin/httpd-mod_cache-errorfix-22.patch
> +1: minfrin
> + trawick: any reason it shouldn't be completely aligned with
> default_handler's
> + choice to return OK vs. 500?
Hmmm...
In the cache, we care about AP_FILTER_ERROR, and pass that back if
present. In the default handler, we ignore AP_FILTER_ERROR completely,
and instead check for this following:
if (status == APR_SUCCESS
|| r->status != HTTP_OK
|| c->aborted) {
return OK;
}
I'm not sure whether merging these is safe enough to backport?
The problem the original patch solves as that an APR error is
appearing in the access_log, and that has caused much confusion for
some people: http://www.google.co.uk/search?q=http+error+103
Ideally, this should be fixed at the very least, and we can then look
at aligning the behaviour to be closer matched. Does this make sense,
or am I being excessively paranoid?
Regards,
Graham
--