JDevlieghere wrote: > I can see how that could happen, but I'm not sure if it's desirable. IIRC, > creating these functions produces some undesirable artifacts, and we actually > do a fair amount of work to validate that we don't create > lldb_private::Function objects for dead-stripped code. I think it would be > illuminating to see how this function manages to get past the checks in > SymbolFileDWARF::ParseFunction.
It turns out that we reject dead code by comparing `low_pc` against the 8-byte `LLDB_INVALID_ADDRESS`. That happens to equal the tombstone on 64-bit targets, but wasm32 has 4-byte addresses, so its tombstone is `0xffffffff` and we let it pass. We can fix this by comparing against `llvm::dwarf::computeTombstoneAddress(cu->GetAddressByteSize())`. https://github.com/llvm/llvm-project/pull/205691 _______________________________________________ lldb-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
