Skip generating DWARF info for dead code

[adamrk]

Here's an attempt at fixing #3403

When encountering a subprogram that is dead code (as indicated by the
dead code proposal
https://dwarfstd.org/ShowIssue.php?issue=200609.1), don't generate debug
output for the subprogram or any of its children.

[alexcrichton]

I'm not personally really all that familiar with this code, but it seems to me like skip_children should be reset back to false at some point after all children have been skipped?

[adamrk]

Ugh, of course. I'll fix that and update the test to make sure it's done properly.

[alexcrichton]

Reading over this some more it appears that the depth_delta and skip_at_depth variables are used as an alternative form of skipping here? Is that something where these two methods of skipping can get unified into one which uses next_sibling and/or next_dfs?

[adamrk]

Is that something where these two methods of skipping can get unified into one which uses next_sibling and/or next_dfs?

Yeah good point. Looking at skip_at_depth more carefully, I think it is used for skipping a node and all its children, so we can just reuse that here. I added some comments describing how it works.

[alexcrichton]

Thanks!