On Thursday 21 November 2002 12:09 pm, Michael Rueger wrote:
> The NameResolver already throughs exceptions as do the already
> rewritten sockets and the network clients.
>

However, client code still checks for nil from the NetNameResolver 
calls.

> The HTTP stuff will also use exceptions, hopefully I can get a
> first version out over the WE.
> Believe me, don't even try to introduce exceptions in the code as
> it is, there are way to many places you would have to "fix" and I'm
> already working on all this.

Couldn't we have the defaultAction of the new HTTP exceptions return 
the same error Strings that they do now? This will result in no 
client code needing to change, but allowing future code to trap the 
exceptions.

So we'd signal the new exceptions from code that already was returning 
Strings for errors.

We'd have to be careful, of course, to make sure that the exceptions 
made sense in context (that is, that we weren't raising them in a 
pre-existing context that hadn't previously returned Strings on 
errors).

-- 
Ned Konz
http://bike-nomad.com
GPG key ID: BEEA7EFE