Re: [Spirit-general] Combining grammars, ASTs and rules

[<ka...@na...>]

Bryan Ross wrote:

> One thing just real quick:
>=20
> When you want to make an AST, you have to use the ast_parse() function,=
 like
> so:
>=20
> tree_parse_info<iter_t> info =3D ast_parse(
> 	buffer,
> 	str_p("Prologue") >> XML >> str_p("Epilogue"),
> 	space_p
> 	);

Mhm.

I have been fiddling around during the day and ended up with the followin=
g=20
construct:

     parse_tree_match_t hit =3D (static_cast<rule_t>(str_p("Prologue") >>=
 g >>=20
str_p("Epilogue"))).parse(scan);

which seems to do almost what I want, even if I haven't figured out how t=
o get=20
the information I want from hit in the right order.  I'll try your sugges=
tion=20
and see what happens.

PS, is there *any* way of speeding up compilation?   I think five minutes=
 for a=20
100-line program is ridiculous.  My "real" application requires the bette=
r part=20
of an hour to compile - I don't get many compile-debug-rewrite iterations=
 out of=20
that in a day...

--=20
Kai-Mikael J=E4=E4-Aro	email:  ka...@na...	    "Don't win, don't lose=
."
NADA, KTH 		voice:  +46 8 790 61 18		-- Yoshigasaki-sensei
SE-100 44  Stockholm	fax:    +46 8 10 24 77
SWEDEN		=09