vitalyd at gmail.com
Thu Nov 29 19:38:34 EST 2012
Yes, this would be nice to see in hotspot. Still, a full blown value type
with guaranteed stack allocation is sorely missing. I routinely see code
which would benefit from this (e.g. Hashcode/EqualsBuilder style classes
that are utility wrappers around one or just a few primitives). Relying on
EA (and not having a diagnostic/product flag that lets us see if something
got EA'd or not doesn't help) is too unreliable.
Sent from my phone
On Nov 29, 2012 7:22 PM, "Doug Lea" <dl at cs.oswego.edu> wrote:
> As a side note wrt other discussions, it is comforting to
> see some progress in avoiding the other main source of Unsafe
> hackery -- emulating structs/tuples etc. See the new
> blog post on the IBM PackedObjects proposal at
> that specifically discusses this use case.
> (See also the (independent) Oracle JEP
> http://openjdk.java.net/jeps/**169 <http://openjdk.java.net/jeps/169>)
> Concurrency-interest mailing list
> Concurrency-interest at cs.**oswego.edu <Concurrency-interest at cs.oswego.edu>
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Concurrency-interest