Re: FYI: Don't hard wire c.m4

[Akim Demaille]

>>> "Paul" == Paul Eggert <address@hidden> writes:

 > That patch broke Bison bootstrapping, since it breaks the calc++ example.
 > To reproduce the problem, check out a fresh version of Bison from CVS,
 > run ./bootstrap, ./configure, make.

 > Here are the symptoms:

 > make[4]: Entering directory `/home/eggert/src/gnu/bison-tmp/examples/calc++'
 > if g++ -DHAVE_CONFIG_H -I. -I. -I../..     -g -O2 -MT calc++-scanner.o -MD 
 > -MP -MF ".deps/calc++-scanner.Tpo" -c -o calc++-scanner.o calc++-scanner.cc; 
 > \
 > then mv -f ".deps/calc++-scanner.Tpo" ".deps/calc++-scanner.Po"; else rm -f 
 > ".deps/calc++-scanner.Tpo"; exit 1; fi
 > ./calc++-parser.hh:121: error: 'driver_yyarg' has not been declared
 > ./calc++-parser.hh:281: error: ISO C++ forbids declaration of 'driver' with 
 > no type
 > ./calc++-parser.hh:281: error: declaration of 'int yy::calcxx_parser::driver'
 > ./calc++-parser.hh:280: error: conflicts with previous declaration 
 > 'calcxx_driver& yy::calcxx_parser::driver'

I don't see that problem anymore.  It might have been related to the
double inclusion of c++.m4 in location.cc, which caused b4_parse_param
to be underquoted.  Sorry for the instable state I caused bw yesterday
and today.