Re: Exception API

[Neil Jerram]

Kevin Ryde <address@hidden> writes:

> I had trouble understanding the different dynamic environments
> specified for that srfi-34 guard.  It looked worryingly like there
> could be jumps out of the originating continuation and then back in if
> there's a re-throw, which could be pretty slow with guile's
> continuations.

Yes, I had the same worry.  I don't think this is intended, though,
and none of the examples in the SRFI do anything so outlandish.
Perhaps one of us should raise this with the R6RS committee?

(My understanding is that SRFI-34 is regarded as one of the things
that will be added to R6RS, and that it is already in pretty much
final form.)

Assuming that it isn't intended, and that this is clarified in R6RS,
we shouldn't need full continuations to implement guard.

(If anyone's wondering about the catch-based implementation of guard
in srfi/srfi-34.scm, I believe it isn't quite right, but I forgot
exactly why.)

Regards,
     Neil