Sep 9 change in gen_lowpart_common causes aborts

[Jeffrey A Law]

  In message < Pine.LNX.4.03.9810021107470.19330-100000@moshier.ne.mediaone.net >
you write:
  > >   > causes current egcs snapshots to abort when trying to compile the 
  > >   > math tests in glibc-2.0.96 (on i586-linux).  Below is a small test
  > >   > program that demonstrates the problem.
  > > 
  > > It's easy to forget that gen_lowpart_common must not fail for integers an
  > ymore
  > > due to changes in how reload works.
  > 
  > The patch was against gcc-2.8.  Your test case does not fail on gcc.
gcc-2.8 does not have the requirement that gen_lowpart_common must not fail for
integers.

That condition was added to deal with some very ugly reload problems on either
the alpha or sparc64 ports.  I don't remember which.

  > I am sorry, but I do not have time to work on two different compilers.
  > Should I delete gcc-2.8 now and change to "egcs?"  Which one is the real 
  > gcc?
That's a subject of great debate.  Just watch gcc2 for differing opinions.

Personally, I work almost exclusively with egcs now.  

jeff