[concurrency-interest] Concurrency-interest Digest, Vol 37, Issue 7

Joshua Bloch josh at bloch.us
Thu Feb 7 14:23:20 EST 2008


On Feb 7, 2008 11:10 AM, David J. Biesack <David.Biesack at sas.com> wrote:

> But is seems removing the *WithException ops would be a good option to
> simplify things and instead push
> RuntimeException handing into the frameworks.

Yes.  And they'll go away entirely if we ever fix generics' treament of
exceptions.  That you can currently say "throws X" is an empty promise as X
cannot consists of more than one Exception type, even though methods can
declare themselves to throw more than one exception type.  Gilad knew this
to be broken at the time that generics were added to the platform, but
didn't have the time to fix it.

> Is it necessary to specialize on byte vs short vs int vs long parameters;
> seems like
> implicit widening to int or long should suffice and you can eliminate a
> lot of combinations.
> It seems sufficient for jsr166y.forkjoin.
Yes.  That's the sort of compromise that seems reasonable on the face of
it.  Of course we can't know if it's right or not until we gain some
experience with it.

All of that said, I would still really like to fix this as part of a
generics fix-up, assuming it is feasible to do so.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: /pipermail/attachments/20080207/f104aefd/attachment.html 

More information about the Concurrency-interest mailing list