a step backwards from egcs-980214 to egcs-980221

[Jeffrey A Law]

  In message < 199802231457.OAA08649@quickstep.stanford.edu >you write:
  > when yydebug == 1.  I suspect that this is a bison bug, after stepping
  > through it a bit.  This could be my fault, indirectly; I checked in a
  > regenerated version of cp/parse.c using the bison (1.25) that shipped
  > with my RedHat 5.0 system.  Would anyone care to generate cp/parse.c
  > themselves, and see if this test case behaves any better?  Do we have
  > a recommended bison version?
I rebuilt it with 1.25 (the one I usually use) on my hpux box
and the test still fails in the same manner.

jeff