2 messages in net.sourceforge.lists.courier-usersRe: [courier-users] Possible issue wi...
FromSent OnAttachments
TimMar 20, 2007 9:45 am 
Malcolm WeirMar 20, 2007 10:40 am 
Actions with this message:
Paste this link in email or IM:
Paste this link in email or IM:
Atom feed for this thread
Paste this URL into your reader:
Subject:Re: [courier-users] Possible issue withCourier0.54.2/threadlib/pthread.c and glibc-2.5-25Actions...
From:Malcolm Weir (ma@gelt.org)
Date:Mar 20, 2007 10:40:12 am
List:net.sourceforge.lists.courier-users

-----Original Message----- From: Tim Sent: Tuesday, March 20, 2007 9:46 AM To: cour@lists.sourceforge.net

In the "make check" I get an error and I am not sure it is one I can figure out.

I to have the same problem Malcom Weir had BUT a slight twist. I get a "Success" followed by a "Invalid argument" from the pthread_cound_destroy(gocond): I am wondering if a fix is available or if anyone might know of a work around?

I have tried setting CGLAGS=-D_THREAD_SAFE and other things like that but nothing seems to alter the behavior. My old courier email server is dieing and I am attempting to put up a new one using a current Courier version. Would an older version of Courier help? Can I turn off the "./configure"'s discovery of the fact that I have a multithreading capable machine? Dell Power Edge 700 PROCESSOR, 80546, 2.8, 1M, PENTIUM 4 PRESCOTT DT, 800, D0 The boot messages say they find and SMP MP-table and it sets up two cpu's.

The error messages you get are vaguely random, and unimportant anyway, and the latest development builds of Courier fix the (trivial) bug that leads to that fact!

In the short term, compile the thing without threads: simply add "--without-pthreads" to the configure line.

[ The issue involves pthreads, not multiprocessor setups. ]

And, as of now, I haven't tracked down the cause (the rest of the world intervened).

By the way, what distro, gcc version and kernel are you running?

,,,Tim Lmrgroup.com

Malc.