The size of the offsets in the table depends on the DWARF format.
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Looks good.
(note to self/anyone: Be nice if this code wasn't loclist specific/was instead shared with the rnglist handling too)
llvm/test/CodeGen/X86/debug-loclists.ll | ||
---|---|---|
26 | Should this dump the loclist = XXX value with a wider encoding, since it's stored wider? (I think that's generally the idea you've had for dumping, that the dumper should dump the width of the encoded field?) |
llvm/test/CodeGen/X86/debug-loclists.ll | ||
---|---|---|
26 | In general, yes, but with some exceptions. In particular, when we show the offset of the information we dump, like the case we have here. The actual encoding of the value in the field is ULEB, and the stored value is 1, while it indirectly points to offset 0x39 in .debug_loclists. |
Should this dump the loclist = XXX value with a wider encoding, since it's stored wider? (I think that's generally the idea you've had for dumping, that the dumper should dump the width of the encoded field?)