gcc 2.95 bootstrap failure on i686-pc-linux-gnu

[Jeffrey A Law]

  In message < Pine.LNX.4.10.9906061009380.1303-100000@ppro.localdomain >you writ
e:

  > /home/petero/cvs/gcc295/objdir/gcc/xgcc -B/home/petero/cvs/gcc295/objdir/gc
  > c/ -B/home/petero/egcs/i686-pc-linux-gnu/bin/ -fgnu-runtime -c -o linking.o
  >  -I. -I../../../egcs/libobjc   -O2 -Wall -march=pentiumpro -mpentiumpro -DI
  > N_GCC -I../../../egcs/libobjc/objc  -I../../../egcs/libobjc/../gcc -I../../
  > ../egcs/libobjc/../gcc/config -I../../gcc -I../../../egcs/libobjc/../includ
  > e ../../../egcs/libobjc/linking.m
  > In file included from ../../../egcs/libobjc/objc/Object.h:30,
  >                  from ../../../egcs/libobjc/linking.m:28:
  > ../../../egcs/libobjc/objc/objc.h:51: parse error before `const'
  > xgcc: Internal compiler error: program cc1obj got fatal signal 11
You've got inconsistent shared files between the C and ObjC front-end.  Most
likely one (but not both) of the parser files (c-parse.? objc-parse.c) was
rebuilt with bison-1.27.

Use the update_egcs script from the contrib directory when you check out stuff
via CVS to make sure that timestamps are updated to avoid this kind of problem.

jeff