[mesa-users] debugging modified code

[Robert FARMER]

Bill i think there leftover from the seg fault tracking, I don't see it in
my utils/makefile_header

Also we don't want to to add it to FCdebug as that requires setting all
*/make/makefiles to be COMPILE=COMPILE_TO_TEST while they currently are set
to COMPILE=COMPILE_TO_DEPLOY. What we could do is move the logic for
COMPILE=( COMPILE_TO_TEST or COMPILE_TO_DEPLOY ) to utils/makefile_header
or just put them in FCbasic.

Rob



On Wed, Aug 5, 2015 at 5:17 PM, Bill Paxton <paxton at kitp.ucsb.edu> wrote:

> hmmm,  now that I look in makefile_header, I find this:
> FCbasic = -fno-range-check -fmax-errors=100 $(SPECIAL_FC_FLAGS)
> -fprotect-parens -fno-sign-zero -fbacktrace -ggdb
>
> Do you find -fbacktrace in your compile commands?
>
> b
>
>
> On Aug 5, 2015, at 5:07 PM, Aaron Dotter wrote:
>
> Hi mesa-users,
>
> Could we add -fbacktrace to FCdebug in utils/makefile_header?  It might
> have once been there and later removed.
>
> FCdebug = -g -fbacktrace
>
> This will, if used to compile the modules, produce a much more informative
> backtrace in situations like Kenny presented without having to invoke gdb.
> The user would have a fighting chance of tracking down the error.
>
> Aaron
>
> On Thu, Aug 6, 2015 at 3:51 AM, Robert FARMER <rjfarmer at asu.edu> wrote:
>
>> Hi
>>
>> The problem is in
>> #3  0x42ABCC in __star_utils_MOD_em1
>> star_utils, backtraces you should read from the top down.
>>
>> In the makefile add -ggdb to the FC_BASIC compile options then ./clean
>> and ./mk.
>>
>> Then when you run the code the line number will either appear in the
>> backtrace or you need to run the code under gdb
>>
>> cd $MESA_DIR/star/test
>> gdb ./test_star
>> type r at the prompt and when it crashes type bt that should tell you
>> which line its crashing at.
>>
>> Rob
>>
>> On Wed, Aug 5, 2015 at 10:13 AM, Kenny Van <kvan at ualberta.ca> wrote:
>>
>>> Hi, I am currently trying to modify some parts of MESA codes and the
>>> code compiles. However the test fails during the star phase of installing
>>> MESA with a SIGSEGV.
>>>
>>> run_star_extras.o run_star.o  run.o -L../../make -lstar -L../../../lib
>>> -lionization -latm -lcolors -lmlt -lnse -lnet -leos -lkap -lrates -lneu
>>> -lchem -linterp_2d -linterp_1d -lnum -lf2crlibm -lcrlibm -lmtx -lconst
>>> -lutils -lmesaklu -lmesalapack -lmesablas
>>>
>>>
>>> Program received signal SIGSEGV: Segmentation fault - invalid memory
>>> reference.
>>>
>>> Backtrace for this error:
>>> #0  0x7FE3A4E00777
>>> #1  0x7FE3A4E00D7E
>>> #2  0x7FE3A40F1D3F
>>> #3  0x42ABCC in __star_utils_MOD_em1
>>> #4  0x66A327 in __hydro_eqns_MOD_do1_dlnt_dm_eqn
>>> #5  0x670ED4 in __hydro_eqns_MOD_eval_equ_for_solver._omp_fn.0 at
>>> hydro_eqns.f90:?
>>> #6  0x7FE3A48DA349
>>> #7  0x7FE3A3EA5181
>>> #8  0x7FE3A41B547C
>>> #9  0xFFFFFFFFFFFFFFFF
>>> ./test_star: line 6: 24055 Segmentation fault      (core dumped) ./star
>>> > log
>>> 0a1
>>> > stop because star_age >= max_age
>>>
>>> Judging from the error the problem is in my hydro_eqns.f90 code. Does
>>> anyone know a way to narrow down the approximate line where the code fails?
>>>
>>> Thanks
>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> mesa-users mailing list
>>> mesa-users at lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>>>
>>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> mesa-users mailing list
>> mesa-users at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/mesa-users
>>
>>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> mesa-users mailing list
> mesa-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/mesa-users
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mesastar.org/pipermail/mesa-users/attachments/20150805/016e5d9f/attachment.html>