[PATCH] D51081: [DWARF v5] Refactoring range list dumping to fold DWARF v4 functionality into v5 handling (almost NFC).

[David Blaikie via Phabricator via llvm-commits]

dblaikie accepted this revision.
dblaikie added inline comments.


================
Comment at: test/DebugInfo/Generic/cu-ranges.ll:17-20
+; CHECK-NEXT: 0x00000000: 0x0000000000000000, 0x0000000000000011 =>
+; CHECK-SAME: [0x0000000000000000, 0x0000000000000011)
+; CHECK-NEXT: 0x00000010: 0x0000000000000000, 0x0000000000000011 =>
+; CHECK-SAME: [0x0000000000000000, 0x0000000000000011)
----------------
wolfgangp wrote:
> dblaikie wrote:
> > Not sure if this is necessary/improving the check, perhaps? (if this test was going to check for the address ranges - it could've been doing that already, I'm not sure the recent changes to dumping make this more viable/appropriate? Not sure though)
> Perhaps this change would capture the intent of the test better, if we want to verify that ranges for 2 different sections are created, i.e. more relying on the ability to display the section.
At that point, we can probably drop the .debug_ranges part of this test entirely.

And the range checking in the debug_info section can probably skip the address ranges and just be:

  CHECK:  DW_AT_ranges
  CHECK-NEXT: "__TEXT,__foo"
  CHECK-NEXT: ".text")+

Since the addresses aren't checked and this isn't a test for llvm-dwarfdump's syntactic output, but a test for LLVM's ranegs output, assuming dwarfdump's output is as expected.

But up to you what you reckon's more readable there.


https://reviews.llvm.org/D51081