[squid-users] ETA for Bug 3775

[Amos Jeffries]

On 21/08/2015 1:14 p.m., Nicolaas Hyatt wrote:
> Hey guys,
> 
> I have been paying close attention to the list for a while and am just
> beginning to realize the scale of things the squid team has in front of
> them. So please understand that I'm _NOT_ begging here. I understand other
> priority issues take precedence, and my little issue is way on down the
> line. I was wondering if there was some sort of schedule as to when this may
> be examined in case I need to provide any more dumps.
> 
>  
> 
> http://bugs.squid-cache.org/show_bug.cgi?id=3775
> 

Unknown again I'm afraid. That latest feedback was a painfully
disappointing result from a very hopeful looking approach.

If you have a dump that provides any new backtrace sequence to what we
have already they may be useful.

So far it all still seems to link back to a DeferredReader pausing a
transaction read, then something else playing around with its I/O
socket/FD/connection, then defer resumes and boom. But there is
potentially some minutes of delay between those actions.

PS. you may want to help with the testing of todays snapshot when its
ready, it contains some patches to HTTPS and FTP connection handling
that may have an impact on this bug (or not, grasping at straws here).

Amos