Le 14/02/2017 à 17:43, yla...@apache.org a écrit :
Author: ylavic
Date: Tue Feb 14 16:43:25 2017
New Revision: 1782986
URL:http://svn.apache.org/viewvc?rev=1782986&view=rev
Log:
util_filter: better ap_pass_brigade() vs empty brigades.
Modified:
httpd/httpd/trunk/server/util_filter.c
Modified: httpd/httpd/trunk/server/util_filter.c
URL:http://svn.apache.org/viewvc/httpd/httpd/trunk/server/util_filter.c?rev=1782986&r1=1782985&r2=1782986&view=diff
==============================================================================
--- httpd/httpd/trunk/server/util_filter.c (original)
+++ httpd/httpd/trunk/server/util_filter.c Tue Feb 14 16:43:25 2017
@@ -584,9 +584,9 @@ AP_DECLARE(apr_status_t) ap_pass_brigade
apr_bucket_brigade *bb)
{
if (next) {
- apr_bucket *e;
+ apr_bucket *e = APR_BRIGADE_LAST(bb);
- if ((e = APR_BRIGADE_LAST(bb)) && APR_BUCKET_IS_EOS(e) && next->r) {
+ if (e != APR_BRIGADE_SENTINEL(bb) && APR_BUCKET_IS_EOS(e) && next->r) {
/* This is only safe because HTTP_HEADER filter is always in
* the filter stack. This ensures that there is ALWAYS a
* request-based filter that we can attach this to. If the
Hi,
I don't really understand the change above.
The commit description is clear. 'ap_pass_brigade' should deal better
with empty brigades.
However, if the last bucket is an EOS bucket, how could it be a sentinel?
My understanding is that this change is a no-op, but I may have missed
something.
What is the rational for it?
CJ
