Daniel Wyatt
Daniel Wyatt
> As a workround I would recommend instead of `--without-os-feature=threads` using `--disable-modules=thread_utils` since `without-os-feature` also disables all provisions for locking etc so if someone used it in a MT app...
(Still hangs with both patches) EDIT: *either, to be clear
> maybe just > `memset(&s, 0, sizeof some_struct_t); ` > ? It's not very C++-y, it relies on all members remaining POD, and it seems maybe more likely to be...
This is already available via `rnp_ffi_set_log_fd` for functions that use `FFI_LOG`, right? That was the original idea there, we just didn't update everything to use `FFI_LOG` yet.
> @dewyatt `FFI_LOG()` requires `rnp_ffi_t` instance, while `RNP_LOG` works on lower layers, where it will be unavailable. However, `rnp_ffi_set_log_fd()` could also set lower-layer global variable, what do you think about...
This is a good start but I think it's incomplete. * Personally I'd prefer an argument name like `password-attempts` vs `num-tries`. * I'm thinking `rnpkeys` should also accept and utilize...