You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@harmony.apache.org by Weldon Washburn <we...@gmail.com> on 2007/05/02 20:44:01 UTC
[drlvm][gc] when will switch to GCV5 happen?
All,
It seems the GCV5 team keeps re-integrating with the latest svn HEAD and
never completely fixes all the regression failures. Now that we have taken
the snapshot for javaone, does it make sense to simply switch to GCV5 and
fix the bugs that result?
The reason I ask is because I am working on threading memory leaks that
assume GCV5. It would be good to get the (minor) redesign and development
of the memory leak problems committed quickly.
--
Weldon Washburn
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Tim Ellison <t....@gmail.com>.
Weldon Washburn wrote:
> It seems the GCV5 team keeps re-integrating with the latest svn HEAD and
> never completely fixes all the regression failures. Now that we have taken
> the snapshot for javaone, does it make sense to simply switch to GCV5 and
> fix the bugs that result?
>
> The reason I ask is because I am working on threading memory leaks that
> assume GCV5. It would be good to get the (minor) redesign and development
> of the memory leak problems committed quickly.
When Xiao-Feng is ready I guess.
Regards,
Tim
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Mikhail Loenko <ml...@gmail.com>.
I believe it should.
modularity is P1 from the beginning of this project
Thanks,
Mikhail
2007/5/3, Pavel Pervov <pm...@gmail.com>:
> >
> > The reason I ask is because I am working on threading memory leaks that
> > assume GCV5.
>
>
> How come? Shouldn't DRLVM's TM be GC independent?
> --
> Pavel Pervov,
> Intel Enterprise Solutions Software Division
>
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Weldon Washburn <we...@gmail.com>.
On 5/3/07, Pavel Pervov <pm...@gmail.com> wrote:
>
> >
> > The reason I ask is because I am working on threading memory leaks that
> > assume GCV5.
How come? Shouldn't DRLVM's TM be GC independent?
Yes, that is the goal. And we will get there over time. And a good way to
do this is to set GCV4.1 aside and complete the integration of MMTk.
--
> Pavel Pervov,
> Intel Enterprise Solutions Software Division
>
--
Weldon Washburn
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Xiao-Feng Li <xi...@gmail.com>.
I've made the switch, thanks. xiaofeng
On 5/4/07, Tim Ellison <t....@gmail.com> wrote:
> Mikhail Fursov wrote:
> > My opinion here is the sooner we get switched to GCv5, the sooner we get it
> > stable.
> > So, I'm +1 to switch to GCv5 this month.
>
> No objections here.
>
> Regards,
> Tim
>
--
http://xiao-feng.blogspot.com
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Tim Ellison <t....@gmail.com>.
Mikhail Fursov wrote:
> My opinion here is the sooner we get switched to GCv5, the sooner we get it
> stable.
> So, I'm +1 to switch to GCv5 this month.
No objections here.
Regards,
Tim
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Mikhail Fursov <mi...@gmail.com>.
My opinion here is the sooner we get switched to GCv5, the sooner we get it
stable.
So, I'm +1 to switch to GCv5 this month.
On 5/3/07, Rana Dasgupta <rd...@gmail.com> wrote:
>
> I think that it should, as well. But maybe the priorities are not the
> same.
>
> My suggestion would be to integrate GCv5 as the default as soon as we
> can. It was switched out primarily because of the > 1.5 GB heap issue.
> This should help it stabilize by regular testing etc.
>
> The TM needs to work correctly and demonstrate elimitation of leaks
> etc. with the default GC first. I think that the fix then needs to be
> ported to 4.1, we want it to stay functional. A similar ( not quite
> the same ) case is the two jits we support.
>
> Thanks,
> Rana
>
>
>
> On 5/3/07, Tim Ellison <t....@gmail.com> wrote:
> > Pavel Pervov wrote:
> > >>
> > >> The reason I ask is because I am working on threading memory leaks
> that
> > >> assume GCV5.
> > >
> > >
> > > How come? Shouldn't DRLVM's TM be GC independent?
> >
> > that's what I was wondering
> >
> >
> > Tim
> >
>
--
Mikhail Fursov
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Rana Dasgupta <rd...@gmail.com>.
I think that it should, as well. But maybe the priorities are not the same.
My suggestion would be to integrate GCv5 as the default as soon as we
can. It was switched out primarily because of the > 1.5 GB heap issue.
This should help it stabilize by regular testing etc.
The TM needs to work correctly and demonstrate elimitation of leaks
etc. with the default GC first. I think that the fix then needs to be
ported to 4.1, we want it to stay functional. A similar ( not quite
the same ) case is the two jits we support.
Thanks,
Rana
On 5/3/07, Tim Ellison <t....@gmail.com> wrote:
> Pavel Pervov wrote:
> >>
> >> The reason I ask is because I am working on threading memory leaks that
> >> assume GCV5.
> >
> >
> > How come? Shouldn't DRLVM's TM be GC independent?
>
> that's what I was wondering
>
>
> Tim
>
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Tim Ellison <t....@gmail.com>.
Pavel Pervov wrote:
>>
>> The reason I ask is because I am working on threading memory leaks that
>> assume GCV5.
>
>
> How come? Shouldn't DRLVM's TM be GC independent?
that's what I was wondering
Tim
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Pavel Pervov <pm...@gmail.com>.
>
> The reason I ask is because I am working on threading memory leaks that
> assume GCV5.
How come? Shouldn't DRLVM's TM be GC independent?
--
Pavel Pervov,
Intel Enterprise Solutions Software Division
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Rana Dasgupta <rd...@gmail.com>.
I have no problems with this. It makes sense to iron out remaining
GCV5 issues as the default, against which tests run regularly.
On 5/2/07, Weldon Washburn <we...@gmail.com> wrote:
> All,
>
> It seems the GCV5 team keeps re-integrating with the latest svn HEAD and
> never completely fixes all the regression failures. Now that we have taken
> the snapshot for javaone, does it make sense to simply switch to GCV5 and
> fix the bugs that result?
>
> The reason I ask is because I am working on threading memory leaks that
> assume GCV5. It would be good to get the (minor) redesign and development
> of the memory leak problems committed quickly.
>
> --
> Weldon Washburn
>
Re: [drlvm][gc] when will switch to GCV5 happen?
Posted by Alexey Petrenko <al...@gmail.com>.
+1 for GCv5 back...
2007/5/2, Weldon Washburn <we...@gmail.com>:
> All,
>
> It seems the GCV5 team keeps re-integrating with the latest svn HEAD and
> never completely fixes all the regression failures. Now that we have taken
> the snapshot for javaone, does it make sense to simply switch to GCV5 and
> fix the bugs that result?
>
> The reason I ask is because I am working on threading memory leaks that
> assume GCV5. It would be good to get the (minor) redesign and development
> of the memory leak problems committed quickly.
>
> --
> Weldon Washburn
>