Fixed bug in linked list where the lastnode wasn't being updated when…

[stevendargaville]

… the last value was popped off the end of the list. Causes a segfault the next time you try and access the lastnode after a pop_last.

[stephankramer]

I'm a little confused: if you don't update lastnode then it presumably still points to node (the last current node it has found by looping through the list), which has now been deallocated?

The issue I see with the current code is that if the length of the current list is one, it never goes into the loop, so prev_node is never initialised. There should probably be a special case where it leaves a valid 0-length list behind in that case.

[stevendargaville]

I'm a little confused: if you don't update lastnode then it presumably still points to node (the last current node it has found by looping through the list), which has now been deallocated?

Yes, so I think that if you call pop_last and then immediately call insert for a non-zero length list, line 264 (node => list%lastnode) in insert will segfault?

The issue I see with the current code is that if the length of the current list is one, it never goes into the loop, so prev_node is never initialised. There should probably be a special case where it leaves a valid 0-length list behind in that case.

Yes you're right, there should be a guard on the assignment of lastnode with a list of size one, I'll add that now.