Issue with nested structs about libffi-rs HOT 7 CLOSED

I'm not sure if the forget() call is the right approach here. I would expect it to leak as element won't have its destructor run. @tov do you perhaps know more about this?

from libffi-rs.

The problem is that the low::ffi_type of element is copied to new, but then element is dropped and the ffi_type deallocated.
I tried changing the loop to *new.add(i) = ffi_type_clone(*element.0); and it also works.

Since Type::drop only deallocates it's type and in this case the type is already copied, I think it should be safe, but I might be wrong.

from libffi-rs.

@xtldr Ah, in that case feel free to set up a pull request and I'll include the changes :)

from libffi-rs.

@YorickPeterse I have set up a pull request. I'm new to GitHub, so if there are any problems with it please let me know.

from libffi-rs.

@xtldr It seems you submitted the PR to your own fork (xtldr#1), instead of this project. I'm not sure if you can change an existing PR, so you may need to set up a new one. IIRC the way to do that is to go to this project, then create a PR from there.

from libffi-rs.

@YorickPeterse Yeah, I got confused when Github offered to create a pull request when I used the online editor for creating the patch. I did a new pull request against this repo, and it seems ok.

from libffi-rs.

Fixed in #23, which has been released on crates.io

from libffi-rs.