[Octave-bug-tracker] [bug #55363] Global variable in a MEX file

[Guillaume]

Follow-up Comment #1, bug #55363 (project octave):

And here is the output of nm on the compiled MEX file.

On OpenSUSE 15.0, gcc 7.3.1:


nm mytest1.mex
0000000000201034 B __bss_start
0000000000201034 b completed.6989
                 w __cxa_finalize@@GLIBC_2.2.5
0000000000000660 t deregister_tm_clones
00000000000006f0 t __do_global_dtors_aux
0000000000200d78 t __do_global_dtors_aux_fini_array_entry
0000000000201028 d __dso_handle
0000000000200d80 d _DYNAMIC
0000000000201034 D _edata
0000000000201038 B _end
0000000000000780 T _fini
0000000000000730 t frame_dummy
0000000000200d70 t __frame_dummy_init_array_entry
0000000000000830 r __FRAME_END__
0000000000201000 d _GLOBAL_OFFSET_TABLE_
                 w __gmon_start__
000000000000079c r __GNU_EH_FRAME_HDR
0000000000000608 T _init
                 w _ITM_deregisterTMCloneTable
                 w _ITM_registerTMCloneTable
0000000000000740 T mexFunction
                 U mexPrintf
                 U mxGetScalar
0000000000201030 D myglobalvar
00000000000007a0 r __offload_funcs_end
00000000000007a0 r __offload_func_table
00000000000007a0 r __offload_vars_end
00000000000007a0 r __offload_var_table
00000000000006a0 t register_tm_clones
0000000000201038 d __TMC_END__


On Ubuntu 18.10, gcc 8.2.0: 


nm mytest1.mex
0000000000003e48 d _DYNAMIC
0000000000004000 d _GLOBAL_OFFSET_TABLE_
                 w _ITM_deregisterTMCloneTable
                 w _ITM_registerTMCloneTable
00000000000020a8 r __FRAME_END__
0000000000002014 r __GNU_EH_FRAME_HDR
0000000000004038 d __TMC_END__
0000000000004034 B __bss_start
                 w __cxa_finalize
00000000000010d0 t __do_global_dtors_aux
0000000000003e40 t __do_global_dtors_aux_fini_array_entry
0000000000004028 d __dso_handle
0000000000003e38 t __frame_dummy_init_array_entry
                 w __gmon_start__
0000000000002018 r __offload_func_table
0000000000002018 r __offload_funcs_end
0000000000002018 r __offload_var_table
0000000000002018 r __offload_vars_end
0000000000004034 D _edata
0000000000004038 B _end
0000000000001160 t _fini
0000000000001000 t _init
0000000000004034 b completed.7930
0000000000001060 t deregister_tm_clones
0000000000001110 t frame_dummy
0000000000001120 T mexFunction
                 U mexPrintf
                 U mxGetScalar
0000000000004030 D myglobalvar
0000000000001090 t register_tm_clones


and with Matlab:


nm mytest1.mexa64 
0000000000201034 b __bss_start
0000000000201034 b completed.6989
                 w __cxa_finalize@@GLIBC_2.2.5
00000000000005e0 t deregister_tm_clones
0000000000000670 t __do_global_dtors_aux
0000000000200d88 t __do_global_dtors_aux_fini_array_entry
0000000000201028 d __dso_handle
0000000000200d90 d _DYNAMIC
0000000000201034 d _edata
0000000000201038 b _end
0000000000000704 t _fini
00000000000006b0 t frame_dummy
0000000000200d80 t __frame_dummy_init_array_entry
00000000000007f8 r __FRAME_END__
0000000000201000 d _GLOBAL_OFFSET_TABLE_
                 w __gmon_start__
0000000000000734 r __GNU_EH_FRAME_HDR
0000000000000588 t _init
                 w _ITM_deregisterTMCloneTable
                 w _ITM_registerTMCloneTable
0000000000000000 A MEX
00000000000006ba T mexFunction
                 U mexPrintf
00000000000006f5 T mexVersion
                 U mxGetScalar
0000000000201030 d myglobalvar
0000000000000620 t register_tm_clones
0000000000201038 d __TMC_END__


So, interestingly, the variable myglobalvar is 'd: initialized data, local'
with Matlab while it is 'D: initialized data, global' with Octave.

    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?55363>

_______________________________________________
  Message sent via Savannah
  https://savannah.gnu.org/