Re: [Chicken-users] catching exceptions

[F. Wittenberger]

Am Dienstag, den 29.07.2008, 05:17 -0700 schrieb Elf:
> srfi-34 is meaningless without srfi-35 and srfi-36.  nothing in srfi-34 
> details the actual format of exceptions/conditions.

Maybe I'm the only one, but I consider this separation of concern an
advantage of srfi-34 over srfi-12.

While it's true that I'm using parts of srfi-35 to encode some of the
exception I'm using, this is a mere accident and might change at any
time.

However I do sometimes simply raise numbers or symbols or other objects
(e.g. srfi-9 records), when that appears appropriate to me.  Simpler
than property conditions.


> all of this is self-contained in srfi-12.  the reason for srfi-12's withdrawl 
> was not 
> because of any flaws inherent in srfi-12, but because william clinger, the 
> author, disappeared apparently for a bit and therefore there was no 
> discussion.  srfi-34 and related srfis are brittle and encode things in a 
> nonschemelike way, with a lot of extra parsing and ridiculousness involved.

That's true: I view srfi-35 more like an interesting way to encode some
types akin to multiple inheritance - though not exactly.  Not too
interesting.  srfi-36 feels somewhat uninteresting to me.

But as I said: I have still little interest in that discussion until my
code runs.  It can't be hard to allow me to raise arbitrary objects and
re-establish the old handler during the exception handling.  I just
can't do it for some mythical reason.

/Jörg