patchset 2 report (billh's resignation)

patchset 2 report (billh's resignation)

Post by sh.. » Mon, 10 Feb 2003 14:48:31



Alexey,

Thanks much for your comment, which wiped my doubt out.


> > The original purpose of the funding, which is boosting the TCK work,
> > is inconsistent with the following announce by Greg:


> > | I'd also like to thank the FreeBSD Foundation
> > | who have provided the funding to allow Alexey to continue his work.

> > This announce says that FBSD Foundation has provided the fund for the
> > porting work of JDK 1.4.1, which does not boost the TCK work at all.

> Greg is correctly stated current stage of my work, but he did not said
> anything about target purpose of work for that I was contracted for.  Main
> aim of my contract is to have "JDK 1.4.x passed thru TCK tests". Therefore
> I don't see a ground of your words.
> According to TCK -- next week I am going to setup environment for TCK
> tests.  I.e. first pass of TCK for 1.4 tests will be completed next week.

Now, I see that, in addition to Greg, you are also allowed to access
TCK.  That is a very good news because TCK has been one of the
toughest barriers between the forthcoming binary releases and us.

Can anyone estimate how long does it take to pass TCK for 1.4.X?
(How long time do you and the Foundation expect to have a binary release?)

It is hard to believe for me 1.4.X can pass TCK soon (in a week or a
month) because I have seen the cases of Linux and FreeBSD.
Blackdown Java Linux porting team needed three month or more from the
acquisition of TCK to release a binary of JDK 1.2.
It took a little longer than a year to pass TCK for FreeBSD 1.3.1.
In those cases the JVM was Classic VM.  This time the target is
HotSpot VM, which loads more burdens on OS and core libraries like
threads.

I myself have a little experience on TCK.  In 1999, Someone has
applied JCK (now known as TCK) to the JIT compiler I have been
developping.  A few hundreds of tests failed even though I noticed no
bug of the JIT.  I was very surprised (though one bug can causes much
failure).  This experience gave me an impression of TCK, that is
substantial and tough.



with "unsubscribe freebsd-java" in the body of the message

 
 
 

patchset 2 report (billh's resignation)

Post by Alexey Zelk » Wed, 12 Feb 2003 04:56:23


hi,


> Now, I see that, in addition to Greg, you are also allowed to access
> TCK.  That is a very good news because TCK has been one of the
> toughest barriers between the forthcoming binary releases and us.

> Can anyone estimate how long does it take to pass TCK for 1.4.X?
> (How long time do you and the Foundation expect to have a binary release?)

Sorry, no estimates yet.  I am just today taken TCK package and after
some period of testing I'll publish my expectations.

Quote:> It is hard to believe for me 1.4.X can pass TCK soon (in a week or a
> month) because I have seen the cases of Linux and FreeBSD.
> Blackdown Java Linux porting team needed three month or more from the
> acquisition of TCK to release a binary of JDK 1.2.
> It took a little longer than a year to pass TCK for FreeBSD 1.3.1.
> In those cases the JVM was Classic VM.  This time the target is
> HotSpot VM, which loads more burdens on OS and core libraries like
> threads.

I agree about notes about time and hotspot.  I don't expect to get it
done within a month, but I have hope that it'll take less than a year.


with "unsubscribe freebsd-java" in the body of the message