[KLUG Members] Oh, the joys of upgrading!

[Adam Williams]

>>>...In her defense, she's a CIS major, not a CS one :). 
>>Justin..don't get me started on this debate ;}
>Is this the CS vs. CIS debate? If so, it's off-topic, but I'm interested
>in it for another reason. I'll take it up in private e-mail.

Include me if you do, mostly I'm interested in watching this debate. 
Personally,  I've worked with recent graduates with either degrees - and
in general they are both totally useless.

>>This happened with a few folks depending on what they had on their
>>machines prior to this assignment. I had everyone do an Apache source
>>install. Some folks who had a binary ran into some gotchas if they
>>hadn't cleaned up beforehand.

Fascinating;  more evidence that a rebuild helps.  Something is very
fishy here.  Any possibility that a subsequent glibc update or some such
fixed a low level threading/mutex problem and via rebuild we are simply
rolling something into apache?  I haven't looked in detail at the RH8
update history.

>I don't understand how Apache 2.0.x could be released without all the
>authentication modules being redone and tested, in light of the new 
>module API, but that's just me. To the extent this is true, Apache 2
>is not ready for prime time.

Apache 2.0.x was released.  Most of the modules are seperate projects,
not Apache's problem.  So the how makes sense, it is just irritating.

>I ALSO don't understand why Adam had such trouble with compiling PHP,
>but I don't want to preview my reply to his posting to this thread :).

I probably have more --enable's than most people.

>>I haven't had any problems with Apache 2.0.43 and PHP 4.3.0, but the

4.3.0 hasn't been released yet.  I assume you mean 4.3.0rc??.  I'm
eager, but can't do that until it is released - production machines and
all...

>>server isn't doing much but running development stuff I'm working on.
>OK, I'm a wee bit behind ya, with 2.0.40 and 4.2.2, and I have not seen any
>problems with Apache operations or php per se; but I'm going to do some volume
>testing in the next few days. The problems themselves relate to threading,
>so the issue would seem to be that Apache 2, php, and the underlying libs
>are thread-safe. If so, no amount of pounding will reveal problems, if not,
>I should be in for many segfaults, and perhaps a move forward to the 

Watch for Server 500 / Document contains no data errors on the client
side.