Re: Speed decrease between 2.0 and 2.1?

[Marcus G. Daniels]

>>>>> "PJ" == Paul E Johnson <address@hidden> writes:

PJ> I'm building for RH linux as we speak.  When you say "Ok, done" do
PJ> you mean that the createActionForEach reverts to its older, fast
PJ> behavior, or what?

No, createActionForEach still must handle multilanguage heterogeneous
collections, and thus will still be slower.  To get a performance
increment, It is necessary to change usages of
createActionForEach:message: to createActionForEachHomogeneous:message:. 
It is still a change, but is an easy one, and when appropriate,
communicates more.

PJ> Is there any penalty to "nonprofiling" usage of swarm when I
PJ> compile swarm with -pg but don't use it to compile the
PJ> application?  I could just compile rpms that way for everybody, if
PJ> there is no minus.  I've read that -g imposes no penalty in file
PJ> loading or execution, but don't know about -pg.

No, it changes code generation; there ought to be a separate RPM if
you're going to provide that...

PJ> Also, I've been building these against kaffe, and wonder what is
PJ> the recommendation these days concerning jikes (which version,
PJ> still the old 0.47?) and the Blackdown JDK. 

As far as I know, there's no harm in building RPMs with any version of
jikes.  The problem with > 0.47 versions of jikes came when interfaces
actually were *used* in a Java model (when the model was compiled by jikes).

PJ> I got a memo from Borland announcing the SDK 1.3, but I still see
PJ> parallel development on Blackdown.

I guess that Sun takes conservative changes at a slow pace, but
Blackdown stays more up to date with changes as a function of Linux 
and glibc development.  (Guessing.)

PJ> ALso, I want to know why the development kit for Java 2 is using
PJ> these version numbers like 1.2.2 and 1.3.

Marketing gimmick.

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.