Fixing memory leaks

[FJShen]

The old PR #31 is closed. This PR is tracking a dedicated branch "FJShen:leakfix" .

[tgrogers]

Another good change that died on the vine @FJShen - what would it take to move this from "draft" to "done"?

[FJShen]

I think that will involve a lot of work. To merge this PR itself is just about pulling in the latest changes from the dev branch (and resolving any conflicts), but there's no guarantee that this PR solves the memory leak problem. Primary reason is that, there are so many raw "new" allocations everywhere in the code, including cases where the program allocates a vector of elements, but deallocate them individually here and there (if I recall correctly). To rectify these issues involves some vast restructuring. Valgrind does not offer much help.

On the other hand I did not notice significant memory leak at a high level, in the form of continuous memory footprint growth. Reports of memory leak from the community were quite sparse.

[tgrogers]

OK, thanks, Fangjia!

I don't think we have a serious leak problem, but we do have a serious orthogonal over-allocation problem with large traces.
Should we just kill this PR then? Or do you feel there is valuable work here that is worth getting working at some point?

2 other points:

1. Vaggrind should always help if you use it correctly
2. I am not sure who "you" is in your comment. You can't blame me for all the code in the simulator :) Try not to personify code, people tend to get upset.

[FJShen]

Sorry for the misunderstanding. I don't blame you @tgrogers I've modified the comment to de-personify the statement.