[CIG-SHORT] Re: PyLith BC help

[Leif Strand]

> Architecture of file not recognized.

According to Apple's docs, there is only one way to debug a PowerPC 
program on an Intel Mac.  I've never done it myself.  It's documented here:

http://developer.apple.com/documentation/MacOSX/Conceptual/universal_binary/universal_binary_exec_a/chapter_950_section_8.html


> [0]0:Return code = 0, signaled with Abort trap

Here's my $0.02 regarding the original symptom, an abort trap:

I remember seeing this with the PyLith binaries, even on a PowerPC.  If 
I remember correctly, the problem is this:  if the C++ code throws any 
exception whatsoever, the process aborts (even if there is a matching 
exception handler further up the stack).  I think you can reproduce this 
problem by simply running the code in an empty directory.  I haven't 
spent much time programming under Mac OS X, so I have no idea what's wrong.

So, for what it's worth, the original symptom is consistent with a C++ 
exception being thrown somewhere in the code... of course, you'd still 
want to use a debugger to find out where.

We should try to get Mac Intel binaries for PyLith at some point. 
Currently, this is blocked by the fact that PETSc cannot be configured 
on our Intel Mac:

     http://tinyurl.com/3bhzm7

The error is:

     Runaway process exceeded time limit of 1800.0s

But this is bogus: if I copy & paste the offending preprocessing command 
into a command shell, it runs instantly.  'configure.py' is doing some 
fancy thread and pipe stuff here, so I suspect there is either a bug in 
'configure.py', or Python itself.

--Leif