Interpreting the memory address in fortran 90 - fortran90

Question

Is there any way to interpret the hex address directly/indirectly in fortran(90). For instance, some times I get memory segmentation issue followed by Failing at address: 0x1fffffff8. A more complete message is something like this:
[amir-desktop] *** Process received signal ***
[amir-desktop] Signal: Segmentation fault (11)
[amir-desktop] Signal code: Address not mapped (1)
[amir-desktop] Failing at address: 0x1fffffff8
[amir-desktop] [ 0] /lib/libpthread.so.0(+0xf8f0) [0x2af5efcb08f0]
[amir-desktop] [ 1] test.exe(__box_mod_MOD_constructlist+0x35c) [0x44052c]
[amir-desktop] [ 2] test.exe(__box_mod_MOD_box_move+0x72) [0x443372]
[amir-desktop] [ 3] test.exe() [0x44c81b]
[amir-desktop] [ 4] test.exe(main+0x1d) [0x41336d]
[amir-desktop] [ 5] /lib/libc.so.6(__libc_start_main+0xfd) [0x2af5f59a9c8d]
[amir-desktop] [ 6] test.exe() [0x41391d]
Is it possible to realize that this address is related to a specific variable, maybe by a simple print statement. I know that loc function prints an integer regarding the address of the variable (Another question arise if they are convertible to hex type address). I would guess that it might be possible to use debugger for that purpose. I would appreciate if some one can comment on this (and some suggestion) as well.
I am currently using gfortran 4.6. under linux distribution.
Thanks

Answer 1

This line:
[amir-desktop] [ 1] test.exe(__box_mod_MOD_constructlist+0x35c) [0x44052c]
tells you that the error was in the function constructlist in the box_mod module. I would start looking there for any obvious errors first.
Try compiling with the following flags:
gfortran -g -fbacktrace -fbounds-check -Wall test.f90
These will (in order): turn on debugging symbols, print out backtraces, check array bounds and print most warnings about unsafe code during compilation.
If that still doesn't solve your problem, you can use gdb on a core dump. This is a record of the state of your program when it crashed. You'll need to enable core dumps first by running ulimit -c unlimited in your terminal. Next, run your code - you should get an error like Segmentation fault (core dumped) - and it will create a file named core or core.pid, where pid is a number. Finally, you can run gdb like so:
gdb ./test.exe core.pid
It will print a whole bunch of information and then leave you with a (gdb) prompt. Run a stack trace with bt to see what happened when your program crashed. quit will exit gdb.