newbie willing to contribute...

[François-Xavier Coudert]

> I've already looked at CMPLX.
> In fact I thought I had finished to implement COMPLEX...

Ouch! This is a collision warning from the gfortran radar. This
message is automatically sent because a collision between gfortran
developers is happening :)

As a matter of fact, I'm stuck in a hotel with no wireless access
around and nothing else to do than writing gfortran code (or at least,
nothing else as tempting). Therefore, I'm trying to get the last g77
intrinsics implemented. I've already done COMPLEX, XOR and other
logical functions, FTELL, FGETC/FGET/FPUTC/FPUT. Sorry for not stating
this earlier, I hadn't had time to read the discussion through.

I intend to submit a patch in the next 48 hours, with the intrinsics above.

> f951 compiled.
> But When I tried to make my test case, it appear that the function
> complex_ is still missing.
> Yet, I thought I had done every thing required! at least in  the 4
> files you mentionned and associated headers.
>
> Can you give me an hint about that?

I think COMPLEX is not a usual intrinsic, in the sense that is does
not call a library function called PREFIX(complex). Rather, it should
be handled in trans-intrinsics to build the tree expression that goes
well (see cmplx implentation in that file).

This is a bit trickier than the usual beginner project (but it's very
interesting, and it's a good way to discover the front-end trees).

FX