[antlr-interest] Can subrules be set to 'n-to-m'?
Scott Stanchfield
scott at javadude.com
Fri Mar 25 19:17:46 PST 2005
No no no no... Silly!
I'm saying that instead of writing
(foo)+[2,3]
It would just be
(foo)[2,3]
As for parens, they should be treated the same way as any other expression
language (other than Lisp ;) -- optional to denote precedence where needed.
I hate typing extra stuff for no good reason. If there were a conflict I'd
agree they should be required, but as Einstein said "make it as simple as
possible, and no simpler".
Everyone deals fine with
x = 3 + 4
Why make them type
x = (3 + 4)
Unless they need the extra precedence indication?
Later,
-- Scott
> -----Original Message-----
> From: John D. Mitchell [mailto:johnm-antlr at non.net]
> Sent: Friday, March 25, 2005 12:14 PM
> To: Scott Stanchfield
> Cc: 'antlr-interest Interest'
> Subject: RE: [antlr-interest] Can subrules be set to 'n-to-m'?
>
> >>>>> "Scott" == Scott Stanchfield <scott at javadude.com> writes:
> [...]
>
> > I think the "+" is redundant here...
>
> So are you also suggesting that (ID)+ should only be written
> as e.g. (ID)[1]?
>
> That would be completely silly. It looks like an indexed
> access operation, it's different than the base case, etc.
>
>
> [...]
>
> > (...)[n,m]
>
> Yes, I agree that it use the square brackets (ala math notation):
>
> (...)+[n,m]
>
> with the ,m part optional.
>
>
> > Should suffice.
>
> > BUT,
>
> > While I'm at it, can (...) be optional for single-ref
> repetition? For
> > example, I'd love to be able to use:
>
> I don't care for that. Why? Because it means that people
> have to scan and understand two different forms for the same
> fundamental meaning. I.e., it's a Good Thing(tm) that the
> parentheses make the loops stand out distinctively.
>
> Take care,
> John
More information about the antlr-interest
mailing list