Getting the Bovinator off the ground

Getting the Bovinator off the ground

[Eric Abrahamsen]

[I was trying to send this to emacs.cedet, but haven't figured out how
to post there. Hoping someone here can help.]

I'm running emacs from git, and have installed CEDET from git as well.

I'm trying to figure out how to use language grammars, and am obviously
misunderstanding something fairly basic. I'm just trying to get a super
barebones Bovine grammar set up, and then trying to use
`semantic-parse-region' to see what it returns, and start playing with
it.

I'm probably misunderstanding something fairly basic here.
`semantic-parse-region' is supposed to return some kind of structure,
right?

Here's my basic grammar:

%package helloworld
%start         test
%keyword BEGINTEST "THISISATEST"
%token COLON     "\\`[:]\\'"

%%
test : bol BEGINTEST COLON value
       ( $1 )
     ;
value : string
        ( $1 )
      ;
%%

I compile and load the el file, then go into a buffer containing this
text:

THISISATEST:hello world

Run (helloworld--install-parser), and then (semantic-parse-region
(point-min) (point-max)).

First of all it tells me that `semantic-lex' is an obsolete function,
then returns nil.

What's it supposed to return? How do I get at the parsed results?

Thanks,
Eric

Re: Getting the Bovinator off the ground

[Eric Abrahamsen]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> [I was trying to send this to emacs.cedet, but haven't figured out how
> to post there. Hoping someone here can help.]
>
> I'm running emacs from git, and have installed CEDET from git as well.
>
> I'm trying to figure out how to use language grammars, and am obviously
> misunderstanding something fairly basic. I'm just trying to get a super
> barebones Bovine grammar set up, and then trying to use
> `semantic-parse-region' to see what it returns, and start playing with
> it.

Actually, what I'm really after is just a way of parsing regular textual
input into a regular lisp structure. I feel like I'm continually
shooting myself in the foot with regular expressions, and wanted
something a little "stricter". I really like how grammars are defined in
the Bovinator's *.by files -- that feels like the right approach -- but
it may also be bringing in a bunch of of unneeded complexity: I don't
need tags or completion or anything.

Are there any lighter-weight approaches to this problem?

Thanks,
Eric