Re: FATAL_EXIT_CODE - why 33?

[Mark Mitchell <mark at codesourcery dot com>]

>>>>> "Richard" == Richard Henderson <rth@redhat.com> writes:

    Richard> I'd been thinking that it oughtn't necessarily be in a
    Richard> main.c, but force each language to define main.  I'd been
    Richard> thinking that it would allow the elimination of some
    Richard> other hook, but looking I don't immediately see which.
    Richard> Perhaps nothing...

There are two basic directions:

  - Create a series of hooks (function-pointers) that are called 
    whenever some non-standard behavior might be wanted.

    (Note: we really want to use function-pointers, not
           function-names, both to reduce interface clutter and
	   also to allow sharing of hooks.)

    Let's call this the "Emacs approach", since Emacs works this
    way.

    The nice thing about this is that the algorithmic structure
    remains in the main code; you just give these hooks a chance
    to fire off every now and then.

    That would argue for putting in a `lang_main' hook which is
    called from `main'.

  - Make a library out of GCC.

    Essentially, create documented, self-contained interfaces
    making GCC suitable for use as a component of a larger piece
    of software.  

    I don't have a cute name for this approach, but it's the 
    GNOME/OLE approach.

    This approach would argument for renaming main to `gcc_main',
    and making it callable.  Going one step further would be to make
    it reentrant and thread-safe...
	
I'm not sure what to do here.

Sorry,

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com