Ok - another data point here... if I build with --disable-shared ex13 runs ok in debug mode... but I do still get printing weirdness like this:

#########
 Mesh Information:
  mesh_dimension()=
 EquationSystems
  n_systems()=
#########

Going to upgrade to 3.2.4 now....

Derek

On Fri, Sep 17, 2010 at 8:22 AM, Derek Gaston <friedmud@gmail.com> wrote:
Has anyone around here been able to successfully use libMesh with XCode 3.2.3 on OSX?

We've held off from upgrading to 3.2.3 for a while because of an incompatibility with the Intel compilers.... but that incompatibility was resolved so we have now tried 3.2.3... with epic failure.

At this point I'm just using 3.2.3 with libMesh by itself.  I'm not using the Intel compiler... nor Petsc.  I'm just letting it use all the default stuff on OSX (including the built in MPI, etc).  I'm not passing any options to libmesh's configure.

So if you want to reproduce this... just "upgrade" to XCode 3.2.3, checkout libmesh do "./configure" and make in debug mode... then go build ex13 (just chosen by random) in debug mode and you should see something like:

#########
 Mesh Information:
  mesh_dimension()=
 EquationSystems
  n_systems()=
ERROR: cannot convert system "Navier-Stokes" to requested type!
[0] /Users/gastdr/projects/libmesh/include/systems/equation_systems.h, line 671, compiled Sep 17 2010 at 08:06:04
terminate called after throwing an instance of 'libMesh::LogicError'
  what():  Error in libMesh internal logic
[Derek-Gastons-MacBook-Pro:13048] *** Process received signal ***
[Derek-Gastons-MacBook-Pro:13048] Signal: Abort trap (6)
[Derek-Gastons-MacBook-Pro:13048] Signal code:  (0)
[Derek-Gastons-MacBook-Pro:13048] [ 0] 2   libSystem.B.dylib                   0x00007fff803e735a _sigtramp + 26
[Derek-Gastons-MacBook-Pro:13048] [ 1] 3   ???                                 0x00007fff5fbfde00 0x0 + 140734799797760
[Derek-Gastons-MacBook-Pro:13048] [ 2] 4   libstdc++.6.dylib                   0x00007fff84f465d2 __tcf_0 + 0
[Derek-Gastons-MacBook-Pro:13048] [ 3] 5   libobjc.A.dylib                     0x00007fff819e4d3d _objc_terminate + 120
[Derek-Gastons-MacBook-Pro:13048] [ 4] 6   libstdc++.6.dylib                   0x00007fff84f44ae1 _ZN10__cxxabiv111__terminateEPFvvE + 11
[Derek-Gastons-MacBook-Pro:13048] [ 5] 7   libstdc++.6.dylib                   0x00007fff84f44b16 _ZN10__cxxabiv112__unexpectedEPFvvE + 0
[Derek-Gastons-MacBook-Pro:13048] [ 6] 8   libstdc++.6.dylib                   0x00007fff84f44bfc _ZL23__gxx_exception_cleanup19_Unwind_Reason_CodeP17_Unwind_Exception + 0
[Derek-Gastons-MacBook-Pro:13048] [ 7] 9   ex13-dbg                            0x0000000100022d71 _ZN7libMesh15EquationSystems10get_systemINS_15TransientSystemINS_20LinearImplicitSystemEEEEERT_RKSs + 869
[Derek-Gastons-MacBook-Pro:13048] [ 8] 10  ex13-dbg                            0x0000000100005c9d main + 1646
[Derek-Gastons-MacBook-Pro:13048] [ 9] 11  ex13-dbg                            0x00000001000026dc start + 52
#########


After looking at the various problems for the past few days it appears to be having trouble dynamic casting things... but I cannot see why.  BUT... that's not the only problem.  Notice the weirdness where the "Mesh Information" block didn't get printed completely.  Some of the examples will run just fine.. but exhibit weird behavior like incomplete printing of information, etc.

All in all... it's total weirdness... and I just want to make sure we're not the only ones seeing it.

I saw this morning that there is an XCode 3.2.4 out as of a week ago... so I'm going to check that out today... but looking over the release notes I don't see anything that would lead me to believe that this is fixed.  I'll report back.

If 3.2.4 doesn't work I'm probably just going to have to downgrade to 3.2.2.  That's the situation we've been in for a while... but it is trouble because we have to tell all of our customers with Macs NOT to upgrade to 3.2.3 (which the OSX software update thing will happily tell them to do all the time).  So it would be nice if we could boil this down to something that XCode is doing wrong so I can send it off to my Apple contacts so this can get fixed once and for all.

Derek