opensubscriber
   Find in this group all groups
 
Unknown more information…

b : bison-patches@gnu.org 2 August 2012 • 7:48PM -0400

Lint in 2012
by Akim Demaille

REPLY TO AUTHOR
 
REPLY TO GROUP




Hi Paul,

There are plenty of ARGSUSED in the generated code.  This dates back to

commit 12ce2df60d16961eaa03a5aa009eeaa645e4e1cb
Author: Paul Eggert <eggert@cs.u...>
Date:   Fri Dec 9 22:14:36 2005 +0000

    Cleanups so that Bison-generated parsers have less lint.
    * data/c.m4 (b4_yydestruct_generate, b4_yysymprint_generate):
    Prepend /*ARGSUSED*/, for lint's sake.
    * data/glr.c (YYUSE): Properly parenthesize, and use an alternate
    definition if 'lint' is defined.
    (YYID): New macro (or function, if lint).
    All uses of /*CONSTCOND*/0 replaced by YYID(0).
    * data/yacc.c: Likewise.
    * data/glr.c (yyuserAction, yyuserMerge, yy_reduce_print):
    (yyrecoverSyntaxError): Prepend /*ARGSUSED*/.
    * data/glr.cc (YYLLOC_DEFAULT): Omit /*CONSTCOND*/ since this code
    is C++ only.
    * data/lalr1.cc (YYUSE): Just use a cast, since this code is C++ only.
    * data/yacc.c (YYSTACK_FREE) [defined YYSTACK_ALLOC]:
    Use YYID(0) rather than 0, for lint.
    (yystrlen): Rewrite to avoid lint warning about ptrdiff_t overflow.
    (yysyntax_error): Rewrite to avoid lint warnings about parenthesization.

Do you think this is still needed?  In particular we now use
consistently YYUSE to "use" (cast to void) arguments that might
not be used.



Bookmark with:

Delicious   Digg   reddit   Facebook   StumbleUpon

Related Messages

opensubscriber is not affiliated with the authors of this message nor responsible for its content.