Re: Mark Mitchell's 15-May toplev.c patch breaks some Fortran statement functions

[Jason Merrill <jason at cygnus dot com>]

>>>>> Mark Mitchell <mark@markmitchell.com> writes:

>   Why exactly did you check in this patch?  I dimly remember preparing
> it, and claiming it had a purpose, but I don't any more remember what
> it was.

It's not actually needed by C++ now.  It would be needed by C++ if
decl_function_context worked properly for C++ and we removed the hacks
currently in the frontend to work around the fact that it doesn't work.

The problem is that if a nested function is deferred, we don't get back to
it until end-of-file, at which point its RTL has been thrown away.  On the
other hand, this hasn't been a problem for C, so perhaps it's just a matter
of fixing C++ to handle forcing out nested functions like C does.

I don't know why the change would be breaking Fortran, tho.  If it's
difficult to fix, go ahead and revert the change.

Jason