leak?

[rogerdpack]

Hi all. Noticed there might be an EM related leak mentioned here:

http://redmine.ruby-lang.org/issues/show/3466
Just pointing it out in case anybody were interested.
Cheers!
-r

[James Tucker]

http://gist.github.com/464518

The above seems to work fine on -head, although I do see the leak with Thin on head. Ask Marc?

>
> --
> You received this message because you are subscribed to the Google Groups "EventMachine" group.
> To post to this group, send email to eventm...@googlegroups.com.
> To unsubscribe from this group, send email to eventmachine...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/eventmachine?hl=en.
>

[Eric Wong]

James Tucker <jftu...@gmail.com> wrote:
> On 5 Jul 2010, at 17:11, rogerdpack wrote:
>
> > Hi all. Noticed there might be an EM related leak mentioned here:
> >
> > http://redmine.ruby-lang.org/issues/show/3466 Just pointing it out
> > in case anybody were interested. Cheers! -r
>
> http://gist.github.com/464518
>
> The above seems to work fine on -head, although I do see the leak with
> Thin on head. Ask Marc?

Another data point, I can reproduce this using EventMachine with
Rainbows! 0.95.1 but not with other concurrency models such as
Rev or FiberSpawn (but I did find something odd with ThreadSpawn
which appears to be a Ruby bug :x)

I've tested with 1.9.2-rc2 on x86_64 and EM 0.12.10

--
Eric Wong

[James Tucker]

Does the bug hide with 1.8?

Would you be able to save me some time and isolate the parser and see if it's occurring there?

>
> --
> Eric Wong

[James Tucker]

On 13 Jul 2010, at 22:50, Eric Wong wrote:

http://gist.github.com/474732

Seems to be stable on 1.9 actually.

Trying to create a minimal case for EM that replicates this, but failing.

[James Tucker]

On 13 Jul 2010, at 22:50, Eric Wong wrote:

Apologies for the spam, reproduced as a standalone test, I will publish details once I've paired this back.

[James Tucker]

On 13 Jul 2010, at 22:50, Eric Wong wrote:

I've now updated the ticket on rubys redmine, this is a leak in ruby 1.9 with Kernel#method. Can you confirm this is the source of your leak too Eric?

http://gist.github.com/474864

Regards,

James

[Eric Wong]

James Tucker <jftu...@gmail.com> wrote:
> I've now updated the ticket on rubys redmine, this is a leak in ruby 1.9 with Kernel#method. Can you confirm this is the source of your leak too Eric?
>
> http://gist.github.com/474864

Yup, I can reproduce this under 1.9.2-rc2 but not 1.9.1-p378.
Thanks for isolating this, James!

--
Eric Wong

[Emmanuel Oga]

  ditto, tested it in both

   ruby-1.9.1-p378 [ x86_64 ]
   ruby-1.9.2-preview3 [ x86_64 ]

  only 1.9.2 leaked (and damn fast, about 2gb in a couple of seconds!)
 

--
Eric Wong

--
You received this message because you are subscribed to the Google Groups "EventMachine" group.
To post to this group, send email to eventm...@googlegroups.com.
To unsubscribe from this group, send email to eventmachine...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/eventmachine?hl=en.

--
--------------------------------------------------------------
EmmanuelOga.com - Software Developer