labath added inline comments.
================
Comment at: lldb/source/Plugins/SymbolFile/NativePDB/PdbAstBuilder.cpp:631
+ if (name_components.empty() ||
+ base_name != name_components.back()->toString())
+ continue;
----------------
What if the two symbols have the same address (same code) and the same base
name (but are e.g. in different namespaces)?
================
Comment at: lldb/source/Plugins/SymbolFile/NativePDB/PdbAstBuilder.cpp:634-659
+ name_components = name_components.drop_back();
+ if (!name_components.empty()) {
+ // Render the current list of scope nodes as a fully qualified name, and
+ // look it up in the debug info as a type name. If we find something,
+ // this is a type (which may itself be prefixed by a namespace). If we
+ // don't, this is a list of namespaces.
+ std::string qname = RenderScopeList(name_components);
----------------
I know you didn't write this, but I still have to express my surprise at the
existence of this code. My don't have anything like that -- using demangled
names to look up things -- in the DWARF plugin. There, you can find the
containing context by just going "up" in the DIE tree. I don't know if that is
possible in PDB, but the current implementation seems fairly fragile, and I
suspect it could be broken if the user overrides a mangled name of a symbol
(`void f() asm("my_linkage_name");`).
Are you sure it's not possible to do this in some other way?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D133243/new/
https://reviews.llvm.org/D133243
_______________________________________________
lldb-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
