Update test_malloc_free.sh

[robertnen]

• now it checks if the executable give segmentation fault or not and gives 0 points for it
• proof: https://imgur.com/Z2jGTlH

[razvand]

Thanks, @Lucky8boy , for submitting it. It's very good to improve the tests in order to make them more resilient and increase the coverage of cases.

However, checking the executable for Segmentation fault is not the proper way to do it. The proper way is to test for the cause of the PR. In other words, the test should be giving 0 points NOT because a Segmentation fault appears, but because there is something wrong in the program logic.

The proper update to the test would be to give 0 points by noting that the program doesn't do the right thing. If that's a segmentation fault, it's not that relevant.

What you should be looking at what is the cause of the segfault, and why is the current test not considering that cause. Maybe it's not checking for a return value, or anything like that. The process disappearing is a good to check as well.

In all cases, even if segfault is the only way to check this (I'm highly doubtful this is the case), using Valgrind is too heavyweight.

[Alex-deVis]

@razvand, bump on this.