Final testing of the cbdata branch accidently revealed a potential
source of problems in earlier Squid versions. commSetTimeout is not
cbdata protected, so any forgotten timeout could trigger on
non-existant data, or in most cases on data now being used for some
other.
One of the apparent uses of commSetTimeout where I think this may be
a source of problems is in the management of persistent connections.
Anyway, I will be cleaning up this in the cbdata branch. Now need to
track down the few locations where commSetTimeout is called with a
non-cbdata variable.
Maybe this should be fixed in 2.5 as well. Unfortunately the cbdata
branch is a bit too intrusive for 2.5 at this time I think..
Regards
Henrik
Received on Mon Apr 08 2002 - 09:54:44 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:14:58 MST