Re: push parser

[Bob Rossi]

On Fri, Dec 08, 2006 at 04:35:19PM -0500, Joel E. Denny wrote:
> On Fri, 8 Dec 2006, Bob Rossi wrote:
> 
> > Yup. I've got a patch coming. Is it ok to do this?
> >   struct yypstate *yypstate = ...;
> > and use yypstate as the type and variable name? If not, any suggestions?
> 
> If we do this:
> 
>   typedef struct yypstate yypstate;
> 
> which seems to be the norm (for example, YYSTYPE and YYLTYPE), then that 
> might get us in trouble.
> 
> In non-Bison code, I try to name types in UpperCamelCase.  The above 
> situation is one of the reasons why.  glr.c is good about this internally.  
> However, it doesn't seem to be the precedent anywhere else in Bison, so I 
> guess we need something like this:
> 
>   yypstate *yyps;

OK, here you go.

2006-12-08  Bob Rossi  <address@hidden>                                         
                        
                                                                                
                        
        * data/push.c: (yypvarsinit, yypvars, struct yypvars): Rename to...
        (yypstateinit, yypstate, struct yypstate): ... these and update 
        all uses.
        (b4_yyssa, b4_yyerror_range, yypstate_init, yypushparse): Rename pv to
        yyps.                                                                   
                      
        (yypstateinit, yypushparse): Rename to ...
        (yypstate_init, yypush_parse): ... these and update all uses.
        (yypstate_init, yypush_parse): Declare in Bison-generated header file.  
                          
        * tests/calc.at (_AT_DATA_CALC_Y): Use newly named functions.           
                        

Thanks,
Bob Rossi

push.c.diff
Description: Text document