Refactoring to be considered before adding MMTk

[udesou]

This PR contains some refactoring of common functions that were moved to gc-common.c and should be shared between MMTk and Julia's stock GC.

[d-netto]

@udesou, a comment before the review: I think rebase might be preferable over merging master to sync a branch, since it avoids interleaving your commits with the upstream ones, and keeps the log "cleaner" (e.g. all of your commits after HEAD).

[udesou]

@udesou, a comment before the review: I think rebase might be preferable over merging master to sync a branch, since it avoids interleaving your commits with the upstream ones, and keeps the log "cleaner" (e.g. all of your commits after HEAD).

Sounds good. I was just clicking the button on GitHub without changing anything but I think I did it right this time.

[fingolfin]

@udesou if you don't mind, could you perhaps click "Resolve conversation" for comments that you already addressed? That would make it a tad easier to follow what's going on here. Thank you, and thank you for this work!

[d-netto]

Please address the gcext failure on CI before merging.

[d-netto]

@fingolfin: I thought we moved gcext out of "Allow Fail" and into a regular test that needs to pass for CI to be green?

If not, perhaps we should consider doing it given that we'll be working on a bunch of GC PR's in the next weeks and it might be useful to have a very explicit indication (e.g. red CI) of whether we broke gcext.

[udesou]

Please address the gcext failure on CI before merging.

@d-netto I think it should be fixed now, however, some unrelated things seem to be failing (by looking at CI results from the latest commit in the Julia repo). If it's really unrelated, then we should be good to go, I guess.

[d-netto]

Reading through it LGTM, but, as usual, let's wait for CI to merge...

[d-netto]

@udesou. CI looks fine. Could you fix the merge conflict in src/gc-debug.c?