Re: Bootstrap failure on darwin 7.4.0 due to memory exhaution

[kenner at vlsi1 dot ultra dot nyu dot edu (Richard Kenner)]

    I had been watching Richard Kenner's patches and wanted to see how 
    things were progressing.  Since Kenner had mentioned fixing ACATS 
    failures and this-and-such patch had broken Ada bootstrap 
    <http://gcc.gnu.org/ml/gcc/2004-07/msg00314.html> I thought that there 
    was a good chance Ada *might* work now.  Apparently not.

There are a few issues here:

(1) I have some fixes to Ada files in my tree that aren't in the FSF tree
because Arno and I are sorting out procedures to check things in while
still unable to run ACATS.  Without one of these, I don't think Ada will
bootstrap on *any* machine.

(2) There was a recent change to the Ada front end that causes it to return
variant records directly.  This causes serious gimplification problems.
I *think* I've now fixed that, but am not sure.

(3) There is a bootstrap failure for Ada on x86 that RTH says is due to
major conceptual problems in assign_parms.  He's working on it.

(4) Because testing of patches can't be done with Ada yet, several
recent patches have broken Ada so I'm very much in a "one step forward,
two steps behind" mode.

I'm trying to keep up and make some forward progress, but I've been out of
town for a few days (and will be out of town again next weekend) and that's
been slowing things down too.

The lowest failure rate so far for ACATS has been about 180 failures.  But
it's gone above that again now.