[Pharo-project] 12186 image quit problem

Pavel Krivanek pavel.krivanek at gmail.com
Sun Oct 10 04:09:10 EDT 2010


On Fri, Oct 8, 2010 at 7:40 PM, Igor Stasenko <siguctua at gmail.com> wrote:

> 2010/10/8 Pavel Krivanek <pavel.krivanek at gmail.com>:
> >
> >
> > On Fri, Oct 8, 2010 at 7:15 PM, Pavel Krivanek <pavel.krivanek at gmail.com
> >
> > wrote:
> >>
> >>
> >> On Fri, Oct 8, 2010 at 5:06 PM, Igor Stasenko <siguctua at gmail.com>
> wrote:
> >>>
> >>> 2010/10/8 Pavel Krivanek <pavel.krivanek at gmail.com>:
> >>> > Hi Marcus,
> >>> > is that really all what happened between 12172 and 12186?
> >>> >
> >>> > update12186
> >>> > Issue 3071: fixing timestamp tests
> >>> > Issue 3068: Non portable SmalltalkImage>>platformName
> >>> > deprecationWarning
> >>> > Issue 3063: removing explain
> >>> > Issue 3039: There should not be any configuration on Core
> >>> > It doesn't explain the problems with VM versions and in 12186 I
> cannot
> >>> > build
> >>> > the PharoKernel (it hangs without error in unload phase)
> >>> > -- Pavel
> >>> >
> >>> please , tell me if it keep hanging after you doing:
> >>>
> >>> WeakArray restartFinalizationProcess.
> >>>
> >>> it seems like a deadlock with WeakArray's FinalizationLock or
> >>> FinalizationSemaphore.
> >>
> >> it hangs for me...
> >> -- Pavel
> >
> > with the MCMethodDefinition>>shutdown hack it work well
>
> so, at least we know what causing it.
>
> Now i wonder, what may cause that
>   WeakArray classPool at: #FinalizationLock
> get locked by some other process and not released anymore?
>
>
if I replace critical:ifError: by critical:ifLocked: in WeakArray >>
removeWeakDependent:, it shows that it is really locked. I can reproduce it
everytime. Any idea how can I determine who locked it?

-- Pavel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.pharo.org/pipermail/pharo-dev_lists.pharo.org/attachments/20101010/4f66973d/attachment-0001.html>


More information about the Pharo-dev mailing list