[PATCH] D41718: [lld] Correctly link S_FILESTATIC records

[Zachary Turner via Phabricator via llvm-commits]

zturner added inline comments.


================
Comment at: lld/COFF/PDB.cpp:440-441
+  case SymbolKind::S_DEFRANGE_SUBFIELD:
+    log("Not fixing up string table reference in S_DEFRANGE / "
+        "S_DEFRANGE_SUBFIELD record");
+    break;
----------------
rnk wrote:
> I guess we could implement this, but I take it we've never encountered an object file with these records?
Yea, I'd like to figure out how to generate them, perhaps running `llvm-pdbutil` against a large corpus of object files and PDBs could turn something up.  I plan to try it on chrome's PDBs one of these days.


================
Comment at: lld/test/COFF/Inputs/pdb-file-statics-a.yaml:43
+              Type:            4193
+              UDTName:         '__vc_attributes::event_sourceAttribute'
+          - Kind:            S_UDT
----------------
rnk wrote:
> I hate, hate, hate this __vc_attributes pollution from C++ debug info test case. Oh well. It's not reasonable to do the work to remove it.
Me too.  I wish I knew what they were for.


https://reviews.llvm.org/D41718