-
1. Re: mod_proxy_cluster.c:3325: possible bad if test ?
jfclere Nov 15, 2013 5:40 AM (in response to dcb314)sure create a JIRA and provide a patch. Actually removing the sessionid is not used often.
-
2. Re: mod_proxy_cluster.c:3325: possible bad if test ?
rhusar Nov 15, 2013 8:22 AM (in response to jfclere)Correct me if I am wrong, but that would actually result in a segfault?
while(*path!='&'||*path!='\0')
path++;
-
3. Re: mod_proxy_cluster.c:3325: possible bad if test ?
dcb314 Nov 15, 2013 8:42 AM (in response to rhusar)>Correct me if I am wrong, but that would actually result in a segfault?
I think so. The while loop cannot finish, so it runs
until it reaches an invalid address and then segfaults.
It also looks as if there needs to be an additional test case in
the test suite to cover this code.
-
4. Re: mod_proxy_cluster.c:3325: possible bad if test ?
rhusar Nov 15, 2013 9:17 AM (in response to dcb314)If I am roughly reading that this logic can be triggered:
* HTTPd version higher than 2.3.11
* selecting a worker for failover to a different domain
* sticky-session-force OFF
* sticky-session-remove ON
Which is not exercised very frequently.
David, so please Jira and see if you can attach a test code as well. Thanks!