fixed bug in stringlist

[aman-godara]

If the inserted slist has length 0, then insert_at can give seg-fault error in this code:

do i = work_idxn + post_length - pre_length, post_length
         list%stringarray(inew) = slist%stringarray(i)
         inew = inew + 1
end do

Resolved by: adding an if-else statement to check for length of slist

[aman-godara]

Any reason why test cases won't pass specifically for macos compiler. I rechecked everything many times but couldn't find any error in logic, I wanted to know if I am making some mistake by making some assumption which is invalid for macos's compiler.

PS: all test cases pass for my gfortran 10.2.0 windows, ran cmake --build build --target clean as well.

[aman-godara]

As per my analysis:

call number 14 to function compare_list is failing because:

work_list <- ["-100", "-99", ..., "99", "100"] (201 items)

I am trying to insert work_list at the head of work_list, expecting an output: ["-100", "-99", ..., "99", "100", "-100", "-99", ..., "99", "100"] (402 items).

This is what I am getting as the output from macos compiler: ["", "", ..., "", "", "-100", "-99", ..., "99", "100"] (402 items)

[jvdp1]

What is the status of this PR? How can we help to move forward?

[aman-godara]

@jvdp1 Thanks for contacting. PR is stuck here, I tried many different things. I have mentioned the details of the problem here.

I am planning to make these test cases readable for others (I should have thought to do that earlier).
But I don't understand why all test cases passes on my GCC compiler but not here on GitHub. This is also preventing me from diving deep into the issue because I have to make a commit everytime and push it to GitHub to see if it works or not. That's why you will see those "temp commits" in the logs.

[jvdp1]

Strange...
I also tested this commit bb30a55a53b3728232727a3afdce168900b0be66 using Fedora with GCC 11.2.0, and all tests passed.
I'll try to look further in the next days...

[aman-godara]

Any luck? @jvdp1

[jvdp1]

Any luck? @jvdp1

No :( And then the holidays arrived. It is also difficult to run tests as I don't have a MacOS computer. Anyone with a MaCOS to test this issue?