HAIFA scheduler bug?

[Jeffrey A Law]

  In message < 199803262227.OAA08984@rtl.cygnus.com >you write:
  > 	Actually we made some changes to reduce the change that jump will bogus
  > ly
  > 	use a REG_DEAD note after reload.  I suspect we didn't catch all of
  > 	them, but we may have caught the one in question.
  > 
  > What changes exactly are these?  I can't see anything obvious looking at
  > the ChangeLog file.
  > 
  > There haven't been any changes at all to delete_computation which is where
  > the bug is.  It is still using REG_DEAD notes which are obviously unsafe.
  > It happens to work by accident for the testcase, but the original bug report
  > points out that the current sources still fail to compile makeinfo correctly
  > so there is still a bug present that needs to be fixed.
Opps.  Sorry, the change was to dead_or_set_p in rtlanal.c.

jeff