fix: source handling with remoteRoot

[connor4312]

Fixes #383. There were a two things that caused issues attaching to
remote processes:

1. We requested sourcemaps from their remote location, which doesn't
   exist locally. This PR rebases the path correctly.
2. We used url length as the heuristic for which to narrow the 'correct'
   file at an absolute path. With a different remote root, though, the
   urls might be totally different. We now also check whether the
   original path is the sourcemap parent of the generated path. This was
   the problem in the linked issue, where ts-node was used.

[jasonwilliams]

Should there be tests for this?
Btw thanks soo much for fixing it!

[connor4312]

Yea, I was thinking about how to do this, I think I can do it well enough by having a test with a duplicate of the workspace running the process out of os.tmpdir(). After booting the process it could wipe that tmpdir location which would trigger the first issue. Won't be able to test cross-platform things (e.g. Windows running Linux containers) but that would work for this, at least.

[jasonwilliams]

I think that should be good enough.
A test for this is certainly worth while as I’m not sure you would be aware if it broke unless you’re actively using it with containers.

[connor4312]

Yep, good reasoning.

I unintentionally practiced proper TDD, by adding a test and not realizing I had changed back to master, then spending a while debugging why it was failing.

[jasonwilliams]

Seems to be working great today :D
Thank you

[connor4312]

Awesome thanks for confirming!