lldb_{10, 11}: polish the cmakeFlags#100070
Conversation
|
@GrahamcOfBorg build lldb_10 lldb_11 |
|
Diff makes sense to me. Not realistically going to build/test this any time soon. |
|
@primeos what about these? Let me squash, quick. |
disable RTTI in a more idiomatic way
|
Why do you disable RTTI for lldb? I think it would make more sense to enable it, since it is enabled for llvm/clang. |
|
It was disabled before, I didn't change that. What would enabling buy us? LLVM has its own RTTI-ish scheme, and it doesn't need all the features of the C++ RTTI. Or am I missing a detail? |
|
It is set for llvm, because some packages seem to need it (4731485)and it is set on clang, because a mismatch can lead to problems with clang-plugins. |
|
I know a (Rust) program that uses |
primeos
left a comment
There was a problem hiding this comment.
LGTM, thanks :)
Feel free to continue that RTTI discussion (or open an issue/PR) but since it was already disabled this PR should be good to go regardless.
|
I don't know of any programs using lldb in this way, but it leads to linking errors that are hard to track. I only asked, because it is different from llvm. I'll have another look at it later, maybe RTTI isn't needed anymore, quite some time has passed since it was enabled :) |
Motivation for this change
LLVM_ENABLE_RTTIdirectly, not as an environment variableLLDB_CODESIGN_IDENTITY, is ignored on DarwinThings done
sandboxinnix.confon non-NixOS linux)nix-shell -p nixpkgs-review --run "nixpkgs-review wip"./result/bin/)nix path-info -Sbefore and after)