Neil Horman
Neil Horman
@baentsch yes, that seems like a reasonable approach I think.
@ashman-p yes, I would think so, I'm a bit tied up at the moment, but given what we've found I would imagine that the leak would occur in that configuration...
I think something went wrong with your squash here. Theres only a comment change in this PR
can you rebase this so its a bit easier to review please? FWIW, the change to the return code looks fine to me, just needs rebasing, ack when thats done
failed ci is related to https://github.com/openssl/project/issues/930, being handled there
can you test again with the latest on the master or 3.5 branch? 864333b455eb36ba84562d6482547bf4c8b49581 may have fixed this
any results from your testing? Did the problem subside?
awesome, thank you!
This is very odd. From what I'm reading here, this sounds to me like perhaps macos has some problem with its symbol exporting mechanism in the compiler. the symbols that...
ah, you're right, I think you're referring to the -exported_symbols_list linker option. Ok, so something can probably be done with that, though it won't work with the constructs we currently...