This is an archive of the discontinued LLVM Phabricator instance.

Remove -generate-dwarf-pub-sections flag.
ClosedPublic

Authored by pcc on Sep 11 2017, 5:05 PM.

Details

Summary

This flag is unnecessary for testing because we can get the coverage
we need by adjusting CU attributes.

Depends on D37655

Diff Detail

Repository
rL LLVM

Event Timeline

pcc created this revision.Sep 11 2017, 5:05 PM
dblaikie added inline comments.Sep 11 2017, 5:15 PM
llvm/test/DebugInfo/Generic/dwarf-public-names.ll
3 ↗(On Diff #114739)

I think it'd be better to change the IR to enable gnu-pubnames here (& in other tests where this patch is currently adding -debugger-tune) - makes it more explicit what the purpose is (I mean, yeah, it may get lost in the noise of the debug info metadata... but still)

llvm/test/DebugInfo/X86/gnu-public-names-gmlt.ll
2 ↗(On Diff #114739)

What's this for? Maybe this removes the purpose of this test (it seems to be testing gnu-public-names and gmlt, which is now impossible). Perhaps the test should be deleted?

Though looking at the original commit (r303894) that added this test, maybe there's something to be concerned about that either this patch or the previous one may've regressed.

Probably worth checking if this does work or did work when it was committed, I guess.

pcc added inline comments.Sep 11 2017, 5:47 PM
llvm/test/DebugInfo/Generic/dwarf-public-names.ll
3 ↗(On Diff #114739)

I think this test (and the others like it) are intended to cover non-GNU pubnames, though.

llvm/test/DebugInfo/X86/gnu-public-names-gmlt.ll
2 ↗(On Diff #114739)

it seems to be testing gnu-public-names and gmlt, which is now impossible

Actually, I think the change makes non-GNU public-names + gmlt impossible.

Perhaps the test should be deleted?

Removing this line seems fine to me. If the purpose of this test is to test things related to gmlt I guess it wouldn't make sense to disable gmlt.

Probably worth checking if this does work or did work when it was committed, I guess.

It isn't clear to me what you mean by "this". Do you mean whether gold accepts the object file?

dblaikie accepted this revision.Sep 12 2017, 2:17 PM
dblaikie added inline comments.
llvm/test/DebugInfo/X86/gnu-public-names-gmlt.ll
2 ↗(On Diff #114739)

By "this" I mean the functionality committed with/supposedly verified by this test:

"DebugInfo: Produce debug_{gnu_}pub{names,types} entries when explicitly requested, even in -gmlt or when empty"

Though admittedly I was mostly looking at gnu pubnames there, not sure if gold uses non-gnu pubnames for gdb-index or not (I think not). So maybe the original generalization over gnu and non-gnu pubnames isn't important (in which case removing this RUN line and only testing the gnu case would be fine), but it does seem a bit inconsistenty, I guess.

This revision is now accepted and ready to land.Sep 12 2017, 2:17 PM
pcc added inline comments.Sep 12 2017, 2:32 PM
llvm/test/DebugInfo/X86/gnu-public-names-gmlt.ll
2 ↗(On Diff #114739)

Okay, looks like that is being tested for by the first line. I'll go ahead and remove the second line, we can always bring it back if we add a supported way of doing gmlt + non-gnu pubnames.

This revision was automatically updated to reflect the committed changes.