Hi Michael, > > Here is an update on AIX. I found why sage startup was crashing in > > real_mpfi.so. I rebuilt libmpfi.so and recompiled real_mpfi.pyx and > > complex_double.pyx. The illegal instruction error was due to an error > > in building libmpfi.so caused by arguments given by libtool. By > > testing with the supplied testcases I was able to narrow it down. > > Ok, can you give some details what had to be changed? I just did a manual:
% cd spkg/build; cp -r mpfi* mympfi (the package builds without error and deletes the sources, so it is important to copy it midway while it is building); % cd mympfi; make; % gcc -shared -o libmpfi.so ../src/.libs/*.o -L/usr/local/PET/src/ build/sage-3.1.1/local/lib -lmpfr -lgmp cp libmpfi $SAGE_LOCAL/lib and tested with the tests/test_mpfi to make sure it didn't crash with the illegal instruction. > > Also, for the list, I edited sage/rings/polynomial/ > > mult_polynomial_libsingular.pyx and replaced the arguments to dlopen > > call to > > > handle = dlopen(lib, 65536+4); > > > It was using 256+1 earlier (RTLD_GLOBAL | RTLD_LAZY). These values > > are defined differently in AIX's dlfcn.h as shown above. > > This will be fixed in Sage 3.2 and a patch is > athttp://trac.sagemath.org/sage_trac/ticket/4427 > > > Now I am running into another error:> > > /usr/local/PET/src/build/sage-3.1.1/local/lib/python2.5/site-packages/sage/ > > matrix/matrix_space.py(38)<module>() > > > -> import matrix_modn_dense > > (Pdb) s > > --Call--> > > /usr/local/PET/src/build/sage-3.1.1/local/lib/python2.5/site-packages/sage/ > > libs/linbox/__init__.py(1)<module>() > > (Pdb) s > > > /usr/local/PET/src/build/sage-3.1.1/local/lib/python2.5/site-packages/sage/ > > > libs/linbox/__init__.py(1)<module>() > > > (Pdb) s > > --Return--> > > /usr/local/PET/src/build/sage-3.1.1/local/lib/python2.5/site-packages/sage/ > > libs/linbox/__init__.py(1)<module>()->None > > > (Pdb) s > > /usr/local/PET/src/build/sage-3.1.1/local/bin/sage-sage: line 226: > > 28756 Illegal instruction (core dumped) sage-ipython "$@" -c > > "$SAGE_STARTUP_COMMAND;" > > Ok, that could be the BLAS or something in LinBox itself. How did you > build ATLAS? > > > I checked the __init__.py file and it has 0 bytes:> ls -lt /usr/local/ > > PET/src/build/sage-3.1.1/local/lib/python2.5/site-packages/sage/libs/ > > linbox/__init__.py > > -rw-r----- 1 sameer 0 Dec 20 2007 /usr/local/PET/ > > src/build/sage-3.1.1/local/lib/python2.5/site-packages/sage/libs/ > > linbox/__init__.py > > > Is this normal? > > That should be ok, all that is required is that file to exist IRC. > > > I think there may be another library (linbox) that > > probably needs to be rebuilt, but what I was wondering if this null > > init.py file is designed to be this way or if it is just an error in > > my installation. I think the illegal instruction probably points to an > > incorrectly installed liblinbox.so on which linbox.so is based, but I > > thought I would check with you. > > You should get a svn checkout of LinBox and run "make check" - in Sage > we patch the commentator and that breaks "make check". Should LinBox > pass its make check we can keep on digging. > Thanks I will rebuild it! - Sameer --~--~---------~--~----~------------~-------~--~----~ To post to this group, send email to sage-support@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/sage-support URLs: http://www.sagemath.org -~----------~----~----~----~------~----~------~--~---
