Re: Better support for Objective-C

[Jeffrey A Law <law at cygnus dot com>]

  In message <38C6D27A.A17EFE86@doc.com>you write:
  > Ovidiu doesn't have much time now, but none of us know anyone else with
  > enough expertise and time to do any work. Although I have been talking
  > with the author of ffcall, who seems to be interested in helping find a
  > solution. Plus the solutions I'm talking about need the support of GCC
  > guru's because it may involve changes to parts of gcc that aren't
  > associated with the Objective-C part.
At least in my case, given the choice of working on something that helps a
large number of people like an ia32 bug, release issues, etc etc vs working
on something that helps a very small number of people (Objective-C) I'll
most likely choose the thing that helps a larger number of people.

One of the basic things you have to realize about a volunteer project like
GCC is that people make their own decisions about what to work on.  If you
can't find a guru willing to spend their time on your problem, you'll either
have to keep begging or become a guru yourself -- for a large number of us
the latter is how we got involved in GCC to begin with.  We had a problem
that needed solving and eventually we decided to tackle it ourself.


  > Yes. That's exactly the problem we're trying to solve. When I talked of
  > libffi, I was thinking of integrating it into libobjc, like it is
  > integrated into the Java runtime now. 
I wasn't aware that libffi was integrated into libgjc; if it is, then the
legal aspects of including it may be resolved relatively soon.


  > It doesn't work on Solaris. There has even been a patch sent in several
  > times, but it has never been put into gcc sources. It doesn't work
  > correctly on PPC either.
It has been fixed for sparcs.  I don't know about the ppc.

jeff