Re: m4sugar

[Akim Demaille]

Le 10 févr. 06 à 02:35, Bob Rossi a écrit :

[bison patches in CC]

> > The first place to look at is certainly yacc.c.  Rename it push.c for
> > instance, and toy with it.  Try to port the work from Odd into it.
> > Maybe you should try to diff yacc.c and odd's skeleton at that time,
> > and port that diff to the current yacc.c.
>
> Hi Akim,
>
> At this point, I'm simply trying to understand how m4 is invoked to
> create the generated c file from yacc.c. If I run bison with --trace I
> see the command
>   /usr/bin/m4
>   /home/bob/download/bison/bison-2.1/target/share/bison/m4sugar/ 
> m4sugar.m4
>   - /home/bob/download/bison/bison-2.1/target/share/bison/c.m4
>     /home/bob/download/bison/bison-2.1/target/share/bison/yacc.c
>
> Since I'm learning m4 for the first time, I was hoping you could  
> confirm
> some of the understanding I have of what's going on above.

Are you sure you are using the CVS version of Bison?  c.m4 should be
included by yacc.c, not fed to m4 by bison.

> So, it's getting input from stdin I'm assuming. For each of the  
> rest of
> the files, m4 reads them in and remembers all the macro's it has
> learned. Probably m4sugar.m4 and c.m4 only contain macro's and  
> yacc.c is
> the file that has to be expanded.
>
> Is this a correct assumption?

Correct.

> Finally, do you know why the '-' is the second argument to m4  
> instead of
> the first?

Because its stdin uses features defined by m4sugar.m4.

> I know this may be trivial stuff, but it's all new to me. I hate to
> start editing yacc.c until I have at least a minimal understanding
> behind the concepts of m4. After looking at the manual for a while,  
> and
> your examples in m4sugar.m4, I'm finally catching on.
>
> Thanks,
> Bob Rossi