G'day all.
On Fri, Apr 19, 2002 at 12:44:49AM -0400, Dan Sugalski wrote:
> Ah. Hmmm. Well, we're already attaching some metadata to ops in a
> different way--that's what the op and inline keywords are doing. For
> metadata that use parameters I can see a scheme like you're
> proposing, though for others it doesn't work.
The reason I tried a general mechanism is that I didn't know exactly
what the optimizer would need before I started. I still don't. I
probably won't even when the optimizer is working. A general approach
gives people a way to experiment, though admittedly little else.
Personally, I'm not sold on the idea either. (My last attempt to send
out feelers on the issue succumbed to Warnock's dilemma.)
> So things like the potential exception-thrower, dead-end, or
> deprecated probably ought to be done the way we're doing now, with
> things that take parameters done differently. (Though I can see
> parenthesized parameters--branch(3), say)
You could always use a Perl 5 modifier-like syntax:
inline op eq(in PMC, in PMC, inconst INT : branch_target)
: throws(strict_refs), alters_state {
# ... insert code here ...
}
Considered it but didn't want to write the parser given that it was a
quick hack.
Cheers,
Andrew Bromage
