kamleshbhalui wrote: > The implementation of Variable::IsThreadLocal is most likely a non-starter, > as it introduces a relatively expensive operation on the common variable > update path for all variables. Doing this (once) at variable creation would > be better, but I still have very big reservations about this approach. > AFAICT, there's no way to guarantee that the variable you find this way will > be the one that is actually described by the DWARF. You could even have more > than one variable with the same name if they're `static thread_local`. > > This really sounds like something where the compiler should give us more to > go on -- instead of us trying to divine this information out of thin air.
I could pass that info to variable creation to simplify the check on update path and about second question even gdb could not debug static thread_local. https://github.com/llvm/llvm-project/pull/110822 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
