; Stan
> >
> > From: Nikolay Izhikov
> > Sent: 24 июля 2018 г. 16:27
> > To: dev@ignite.apache.org
> > Subject: Re: ConcurrentLinkedHashMap works incorrectly after clear()
> >
> > Hello, Ilya.
> >
> > May be we need to proceed with ticket [1]
should be OK).
>
> Stan
>
> From: Nikolay Izhikov
> Sent: 24 июля 2018 г. 16:27
> To: dev@ignite.apache.org
> Subject: Re: ConcurrentLinkedHashMap works incorrectly after clear()
>
> Hello, Ilya.
>
> May be we need to proceed with ticket [1] "Get rid of
> org.
On Thu, Aug 9, 2018 at 8:16 AM, Alexey Goncharuk wrote:
> Guys, I think we can definitely change current implementation of CLHM with
> a more stable one, but as a temporal solution I see nothing wrong with
> throwing an UnsupportedOperationException from clear() method. Ilya already
> provided a
lya Lantukh. But I
> can
> > > > > mistake here.
> > > > >
> > > > > In general way, I think it's a good case to start thinking about
> how
> > to
> > > > get
> > > > > rid of CLHM. E.g. we can consider thi
[1].
> > > >
> > > > [1] https://github.com/ben-manes/concurrentlinkedhashmap
> > > >
> > > > вт, 24 июл. 2018 г. в 16:45, Stanislav Lukyanov <
> > stanlukya...@gmail.com
> > > >:
> > > >
> > > > > It see
t; >:
> > >
> > > > It seems that we currently use the CLHM primarily as a FIFO cache.
> > > > I see two ways around that.
> > > >
> > > > First, we could implement such LRU/FIFO cache based on another,
> > properly
> > >
aft implementation of
> > > LruEvictionPolicy based on it that passes functional tests,
> > > but I haven’t benchmarked it yet.
> > >
> > > Second, Guava has a Cache API with a lot of configuration options that
> we
> > > could use (license i
nse is Apache, should be OK).
> >
> > Stan
> >
> > From: Nikolay Izhikov
> > Sent: 24 июля 2018 г. 16:27
> > To: dev@ignite.apache.org
> > Subject: Re: ConcurrentLinkedHashMap works incorrectly after clear()
> >
> > Hello, Ilya.
> >
> > May b
tional tests,
> but I haven’t benchmarked it yet.
>
> Second, Guava has a Cache API with a lot of configuration options that we
> could use (license is Apache, should be OK).
>
> Stan
>
> From: Nikolay Izhikov
> Sent: 24 июля 2018 г. 16:27
> To: dev@ignite.apache.org
> Subj
: ConcurrentLinkedHashMap works incorrectly after clear()
Hello, Ilya.
May be we need to proceed with ticket [1] "Get rid of
org.jsr166.ConcurrentLinkedHashMap"?
Especially, if this class is broken and buggy.
[1] https://issues.apache.org/jira/browse/IGNITE-7516
В Вт, 24/07/2018 в 16:20 +0300, Il
Hello, Ilya.
May be we need to proceed with ticket [1] "Get rid of
org.jsr166.ConcurrentLinkedHashMap"?
Especially, if this class is broken and buggy.
[1] https://issues.apache.org/jira/browse/IGNITE-7516
В Вт, 24/07/2018 в 16:20 +0300, Ilya Lantukh пишет:
> Thanks for revealing this issue!
>
Thanks for revealing this issue!
I don't understand why should we disallow calling clear().
One way how it can be re-implemented is:
1. acquire write locks on all segments;
2. clear them;
3. reset size to 0;
4. release locks.
Another approach is to calculate inside
ConcurrentLinkedHashMap.Segmen
12 matches
Mail list logo
