Re: [squid-users] Question on 2.4S2 change re optimization

From: Joe Cooper <joe@dont-contact.us>
Date: Mon, 27 Aug 2001 00:40:24 -0500

I've had no stability problems with any Squid compiled with 2.96-85 and
2.96-95 (from Raw Hide), assuming it actually compiled. However, both
have failed to compile an i686 optimized version in some configurations.

After updating glibc and a few other parts of the system, this
phenomenon disappeared...so I couldn't pin it down to specific options
or versions of anything, and I haven't had time to try to track it more
precisely. Also note that this is a different bug than folks have run
into with 2.95...there, it would compile successfully but crash in randomly.

I'd just say, compile it as you normally would, and then keep a close
eye on it for a few days afterward. If you have no 'random' crashes in
that time, you likely aren't going to fall prey to the compiler bugs
that folks have seen with Squid and gcc 2.95.x. (I usually run through
a number of tests on each Squid revision I release...Polygraph runs, a
big list of sites wgetted through it, browsing through it while doing
either or both of those things, etc.)

Steve Snyder wrote:

> Reading the ChangeLog for version 2.4STABLE2, I see this note:
>
> - Expanded configure's GCC opimization disabling check to
> include GCC 2.95.3
>
> Can any of the Squid developers explain which optimizations cause problems
> and whether the problems are only in gcc v2.95.x? My reason for asking is
> that I'm running Squid on a RedHat v7.1 box, using their, ahem, unique
> version of the gcc compiler. Given this compiler, are there any compiler
> optimization(s) in particular that I should avoid using?
>
> Thanks.

                                   --
                      Joe Cooper <joe@swelltech.com>
                  Affordable Web Caching Proxy Appliances
                         http://www.swelltech.com
Received on Mon Aug 27 2001 - 03:40:00 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:01:55 MST