Sure is needed.
Wonder how long Squid needs to run before cbdata gets upset on this
one.. (cbdata won't like at all if the lock count wraps).. Apparently
very long as there has been no complaints on this one from what I can
tell.
Regards
Henrik
On Friday 25 October 2002 01.23, Adrian Chadd wrote:
> Could someone also review this and see if its needed in
> squid2-HEAD and squid-2.5 ?
>
> Ta!
>
>
>
> Adrian
>
> ----- Forwarded message from adrian@squid-cache.org -----
>
> Date: 24 Oct 2002 23:22:30 -0000
> From: adrian@squid-cache.org
> To: squid-cvs@squid-cache.org
> Subject: cvs commit: squid3/src acl.cc
>
> adrian 2002/10/24 17:22:30 MDT
>
> Modified files:
> src acl.cc
> Log:
> Add a cbdataReferenceDone() call to match the cbdataReference()
> call performed in aclChecklistCreate().
>
> In a nutshell, this wasn't a memory leak, but it was a lock-leak
> - the acl itself wasn't being freed, but the refcount slowly crept
> upwards.
>
> Revision Changes Path
> 1.291 +2 -1 squid3/src/acl.cc
>
>
> ----- End forwarded message -----
Received on Sat Nov 09 2002 - 20:09:46 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:18:39 MST