Re: Dead Horse

--MimeMultipartBoundary
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

libc5 _does_ leak. Back when I worked for OM, one of my engineers came
up grumbling with a memory-leak exploit for libc5, and proved it to me.
Don't know if I still have the thing (what with all the on-premises code
being not-my-property(tm)) but I'll see.

I'm told early libc6 leaked as well. Don't know about more recent
versions.

Chris Wedgwood wrote:
>
> > Having two tags is confusing. However, both of them are important.
> > Perhaps, we should have one "the amount of memory the squid should use"
> > tag and then another tag that would specify the proportion between "cache
> > memory" and "pooled memory". However, it is still unclear if one can track
> > all the memory in use precisely (I doubt that) and what to do if Squid
> > runs over the limit..
>
> Not necessary. Some people just need to learn to read more carefully....
>
> Although, the squid config. that comes with beta20 has memory_pools on and
> no size set, so it defaults to zero with is treated as 2GB, which mean its
> will grow and grow and grow if you leave this part of the config. unedited.
>
> (I'm still pretty sure I am seeing a libc leak with one one machine though,
> but the other is fine).
>
> -Chris

-- 
-----BEGIN GEEK CODE BLOCK-----
Version: 3.1
GAT d- s++: a C++++$ UL++++B+++S+++C++H++U++V+++$ P+++$ L+++ E-
W+++(--)$ N++ w++$>--- t+ 5++ X+() R+ tv b++++ DI+++ e- h-@ 
------END GEEK CODE BLOCK------
--MimeMultipartBoundary--