vyzo

Did you run the tests with the `LD_LIBRARY_PATH` set too?

So it seems we need those definitions included in `libcrypto.scm` on MacOS. Can you add a darwin guard here: https://github.com/vyzo/gerbil/blob/master/src/std/crypto/libcrypto.scm#L303 ?

this is very weird, I am not sure what's going on. @ober any ideas?

I don't know how productive it is to be fighting with gambit head, it seems rather unstable. The problem is somewhere in openssl.

That's good to hear. I would like to bisect what goes wrong in source installation nonetheless.

That would be very cool, but it's quite a bit of work. It would make a great contribution if you feel like hacking something like this.

You mean creating a cons cell from ffi? You can use `___CONS`: ``` return ___CONS(___FIX(w.ws_row), ___FIX(w.ws_col)) ```

Also, declare the method as return a `scheme-object`.

That's interesting, you can only use `___CONS` in an inline block it seems because of the heap pointer. We'll have to ask Marc about this, it used to work!

There is also the function `___make_pair`, which you can call to make a pair, but it needs a pstate as argument.