> Does '#ident' really have the same semantics as '#undef'? Or
> would it be more sensible to just ignore it?
That was what I was trying to do. Perhaps I failed. I didn't
really understand what the #undef code did. I just made a
qualified guess.
> If ignoring is the answer I would rather propose the following
> patch:
>
> --- parser.l.orig Mon Mar 20 21:46:06 2000
> +++ parser.l Mon Mar 20 21:46:59 2000
> @@ -548,6 +548,7 @@
>
> /* preprocessor junk */
>
> <INITIAL,pp_strips,pp_stripe,pp_stripp,pp_false>^{ws}*#{ws}*pr
> agma[^\n]*
> ;/* Ignore #pragma */
> +<INITIAL,pp_strips,pp_stripe,pp_stripp,pp_false>^{ws}*#{ws}*i
> dent[^\n]*
> ;/* Ignore #ident */
>
> <INITIAL,pp_strips,pp_stripe,pp_stripp,pp_false>^{ws}*#{ws}*line[^\n]*
> ;/* Ignore #line */
> /* We'll get an error on malformed #xxx statements
> * by not recognising '#' at all. This helps tracking
That is probably much better.
