Fix signals/interrupts in sage-doctest

[jdemeyer]

When a doctest gets the dreaded Unhandled SIGSEGV..., the doctest script doesn't know how to handle this if the killed process is the child of /bin/sh, not /bin/sh itself. Newer versions of bash automatically do execve() instead of fork() when there is only one command to run, so the problem does not appear then. The fix is to always exec or even better, not using a shell at all. Consider sigdie.py​. Running this on forking shells gives:

sage -t  "/home/jdemeyer/sigdie.py"
/home/jdemeyer/sage/sage-4.7.alpha5/local/lib/libcsage.so(print_backtrace+0x31)[0x7f66bf2dd415]
[...]
/lib/libpython2.6.so.1.0(Py_Main+0xb4e)[0x7f66c2aa4e7e]
/lib/libc.so.6(__libc_start_main+0xf4)[0x7f66c1da11c4]
python[0x400619]

------------------------------------------------------------------------
Unhandled SIGSEGV: A segmentation fault occurred in Sage.
This probably occurred because a *compiled* component of Sage has a bug
in it and is not properly wrapped with sig_on(), sig_off(). You might
want to run Sage under gdb with 'sage -gdb' to debug this.
Sage will now terminate.
------------------------------------------------------------------------
Segmentation fault

         [2.8 s]

----------------------------------------------------------------------
The following tests failed:


        sage -t  "/home/jdemeyer/sigdie.py"
Total time for all tests: 2.9 seconds

when it should give:

sage -t  "devel/sage/sage/testdoctest/sigdie.py"
The doctested process was killed by signal 11
         [6.3 s]

----------------------------------------------------------------------
The following tests failed:


        sage -t  "devel/sage/sage/testdoctest/sigdie.py" # Killed/crashed
Total time for all tests: 6.4 seconds

Another sympton is that a doctest TIME OUT kills the wrong process: it kills the shell, not the actual Python command.

[jdemeyer]

Doctest showing the failure (1)

[jdemeyer]

Superseded by #12415.