[PATCH] D78208: [DebugInfo] Report the format of the DWARF64 sections.

[James Henderson via Phabricator via llvm-commits]

jhenderson added inline comments.


================
Comment at: llvm/lib/DebugInfo/DWARF/DWARFCompileUnit.cpp:20-21
+     << " length = " << format("0x%08" PRIx64, getLength());
+  if (getFormat() == dwarf::DWARF64)
+    OS << " format = DWARF64";
+  OS << " version = " << format("0x%04x", getVersion());
----------------
I wonder if it would make sense to print DWARF32 for when we're not using DWARF64? That would make the output more regular and parseable. What do you think?


================
Comment at: llvm/test/DebugInfo/X86/debug-frame-dwarf64.s:1
+# RUN: llvm-mc -triple x86_64 %s -filetype=obj -o - | \
+# RUN:   llvm-dwarfdump -debug-frame - | \
----------------
Is there no testing already for DWARF64 debug_frame? If not, you should probably test the length fields and any other bits that are DWARF64 related too. If you don't want to do it as part of this change, that's fine, but perhaps it should be a prerequisite?


================
Comment at: llvm/test/DebugInfo/X86/dwarfdump-rnglists-dwarf64.s:212
 
+# RNGLISTS:      .debug_rnglists contents:
+# RNGLISTS:      range list header:
----------------
This block might belong below the ERR block to match order of usage.


================
Comment at: llvm/test/DebugInfo/X86/dwarfdump-rnglists-dwarf64.s:247
+# RNGLISTS-NEXT: <End of list>
+
+
----------------
Nit: too many blank lines.


================
Comment at: llvm/test/DebugInfo/dwarfdump-64-bit-dwarf.test:2
 RUN: llvm-dwarfdump %p/Inputs/dwarfdump.elf-mips64-64-bit-dwarf \
-RUN:   --debug-line | FileCheck %s
+RUN:  --debug-line | FileCheck %s
 
----------------
Nit: unrelated (and bad?) change.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D78208/new/

https://reviews.llvm.org/D78208