[PATCH] D69847: DWARFDebugLoc(v4): Add an incremental parsing function

[Pavel Labath via Phabricator via llvm-commits]

labath added a comment.

In D69847#1754282 <https://reviews.llvm.org/D69847#1754282>, @dblaikie wrote:

> Ah, sure - that was a comment I didn't realize I'd written a while back and had failed to hit "send" on, feel free to ignore it if it doesn't make any more sense today or the like.


It's still relevant. The deed is now kind of done, so changing that would mean re-updating every test, but it still can be done. And you did remind me that I still want to add support for parsing/dumping of the section indexes, so this would be a good opportunity to revisit that..

Overall, I don't think the format of debug_loc matters that much. They (unlike loclists) are very simple, and so printing them as `[a,b)` could be fine even if they are not real ranges. I sort of like how the leading `(` subtly hints that this is not a real range, but that hint may be too subtle/confusing. If I take the consistency with debug_loclists into consideration, then I think that `{}` might be the best -- debug_loc could be printed as `{a, b}`, and debug_loclists as `{DW_LLE_whatever, a, b}`. It would be good to hear if anyone else has an opinion too, to avoid flipping this back and forth.

I also want to address the range/loclist discrepancy somehow, but I first want to finish the work with the location lists. The two things that I know I'm still missing are: parsing section indices; and parsing other DW_LLE_ forms.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D69847/new/

https://reviews.llvm.org/D69847