Le 13 sept. 2009 à 22:01, Joel E. Denny a écrit :
I'd like to push this patch to master, branch-2.5, and branch-2.4.2 so
that push parser users are informed that they can't set
YYSTACK_USE_ALLOCA=1. The only way I can see to do this is with
#error in
yacc.c. However, I don't see any other use of #error in the
skeletons.
Is there any reason not to use it?
In a distant past, when free(0) could not be trusted, there was also
compilers (well, preprocessors), that die on #unknown directives, even
if inside a skipped #if section.
I suppose this is no longer the case, so let's do it. But I wouldn't
in 2.4.2 though. Let's drop knr completely in 2.5, not before.
