[Felix-impl] Fwd: SF.net SVN: felix: [1180] felix/trunk

[Rhythmic F. <rfi...@gm...>]

---------- Forwarded message ----------
From: rfi...@us...
To: fel...@li...
Date: Thu, 23 Nov 2006 14:55:33 -0800
Subject: SF.net SVN: felix: [1180] felix/trunk
Revision: 1180
          http://svn.sourceforge.net/felix/?rev=1180&view=rev
Author:   rfistman
Date:     2006-11-23 14:55:33 -0800 (Thu, 23 Nov 2006)

Log Message:
-----------
added surprising comment about win32 mutexes
added #! to autogen.sh for tab completion in cygwin

Modified Paths:
--------------
    felix/trunk/autogen.sh
    felix/trunk/lpsrc/flx_pthread.pak

Modified: felix/trunk/autogen.sh
===================================================================
--- felix/trunk/autogen.sh      2006-11-21 14:55:34 UTC (rev 1179)
+++ felix/trunk/autogen.sh      2006-11-23 22:55:33 UTC (rev 1180)
@@ -1,3 +1,4 @@
+#!/bin/bash
 echo "DEFAULT FELIX BUILD"
 ./configure
 ./mk extract

Modified: felix/trunk/lpsrc/flx_pthread.pak
===================================================================
--- felix/trunk/lpsrc/flx_pthread.pak   2006-11-21 14:55:34 UTC (rev 1179)
+++ felix/trunk/lpsrc/flx_pthread.pak   2006-11-23 22:55:33 UTC (rev 1180)
@@ -515,6 +515,9 @@
   unsigned long tv_nsec;
 };

+// RF: unlike real pthread mutexes, windows mutexes are always recursive.
+// that's annoying because I use deadlock as a debugging tool. the upshot
+// is that recursively acquiring a mutex gives undefined results.
 int pthread_mutex_init (pthread_mutex_t *m, const pthread_mutexattr_t*)
 {
   *m = CreateMutex(NULL,FALSE,NULL);


This was sent by the SourceForge.net collaborative development
platform, the world's largest Open Source development site.