IGNITE-3422 - ready for review

[daradurvs]

Hello everyone.

Please, review implemented solution.

https://issues.apache.org/jira/browse/IGNITE-3422 - No way to control
object initialization during deserialization/unmarshalling

ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>

[Vladimir Ozerov]

Hi Vyacheslav,

Thank you for contribution. I reviewed implementation again and now I am in
doubts whether our product would really benefit from it or not. See my
comments in the ticket. I'de prefer Denis Magda to chime in and give his
feedback first.

Vladimir.

On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <[hidden email]>
wrote:

> Hello everyone.
>
> Please, review implemented solution.
>
> https://issues.apache.org/jira/browse/IGNITE-3422 - No way to control
> object initialization during deserialization/unmarshalling
>
> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>
>

[dmagda]

Replied.

—
Denis

> On Feb 20, 2017, at 3:08 AM, Vladimir Ozerov <[hidden email]> wrote:
>
> Hi Vyacheslav,
>
> Thank you for contribution. I reviewed implementation again and now I am in
> doubts whether our product would really benefit from it or not. See my
> comments in the ticket. I'de prefer Denis Magda to chime in and give his
> feedback first.
>
> Vladimir.
>
> On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <[hidden email]>
> wrote:
>
>> Hello everyone.
>>
>> Please, review implemented solution.
>>
>> https://issues.apache.org/jira/browse/IGNITE-3422 - No way to control
>> object initialization during deserialization/unmarshalling
>>
>> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>
>>

[daradurvs]

Guys, let's discuss a goal of this task.

I need the task specification.


2017-02-22 2:00 GMT+03:00 Denis Magda <[hidden email]>:

> Replied.
>
> —
> Denis
>
> > On Feb 20, 2017, at 3:08 AM, Vladimir Ozerov <[hidden email]>
> wrote:
> >
> > Hi Vyacheslav,
> >
> > Thank you for contribution. I reviewed implementation again and now I am
> in
> > doubts whether our product would really benefit from it or not. See my
> > comments in the ticket. I'de prefer Denis Magda to chime in and give his
> > feedback first.
> >
> > Vladimir.
> >
> > On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <[hidden email]
> >
> > wrote:
> >
> >> Hello everyone.
> >>
> >> Please, review implemented solution.
> >>
> >> https://issues.apache.org/jira/browse/IGNITE-3422 - No way to control
> >> object initialization during deserialization/unmarshalling
> >>
> >> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>
> >>
>
>

[dmagda]

In my understanding the goal is well-defined in the ticket. At the same time we have a similar task prepared by Pavel for .NET:
https://issues.apache.org/jira/browse/IGNITE-3102 <https://issues.apache.org/jira/browse/IGNITE-3102>

So, what we need to agree on is how to proceed with the implementation. Let us think this over and propose design.

—
Denis

> On Feb 22, 2017, at 2:41 AM, Vyacheslav Daradur <[hidden email]> wrote:
>
> Guys, let's discuss a goal of this task.
>
> I need the task specification.
>
>
> 2017-02-22 2:00 GMT+03:00 Denis Magda <[hidden email] <mailto:[hidden email]>>:
> Replied.
>
> —
> Denis
>
> > On Feb 20, 2017, at 3:08 AM, Vladimir Ozerov <[hidden email] <mailto:[hidden email]>> wrote:
> >
> > Hi Vyacheslav,
> >
> > Thank you for contribution. I reviewed implementation again and now I am in
> > doubts whether our product would really benefit from it or not. See my
> > comments in the ticket. I'de prefer Denis Magda to chime in and give his
> > feedback first.
> >
> > Vladimir.
> >
> > On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <[hidden email] <mailto:[hidden email]>>
> > wrote:
> >
> >> Hello everyone.
> >>
> >> Please, review implemented solution.
> >>
> >> https://issues.apache.org/jira/browse/IGNITE-3422 <https://issues.apache.org/jira/browse/IGNITE-3422> - No way to control
> >> object initialization during deserialization/unmarshalling
> >>
> >> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054 <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>>
> >>
>
>

[daradurvs]

I think it makes a sense to define factory in IgniteConfiguration, as the
Map<Class, InstanceFactory> which will be ised when deserializing.

Interface example:

public intterface InstanceFactory {
    newInstance(Class clazz, BinaryReader reader);
}

P.S.
About "constructor(BinraryReader reader)" - I think there is no sense,
because it just duplicate readObject method and we can't manage a reference
of new instance.
How about singlton-deserialization?



2017-02-23 10:16 GMT+03:00 Denis Magda <[hidden email]>:

> In my understanding the goal is well-defined in the ticket. At the same
> time we have a similar task prepared by Pavel for .NET:
> https://issues.apache.org/jira/browse/IGNITE-3102 <
> https://issues.apache.org/jira/browse/IGNITE-3102>
>
> So, what we need to agree on is how to proceed with the implementation.
> Let us think this over and propose design.
>
> —
> Denis
>
> > On Feb 22, 2017, at 2:41 AM, Vyacheslav Daradur <[hidden email]>
> wrote:
> >
> > Guys, let's discuss a goal of this task.
> >
> > I need the task specification.
> >
> >
> > 2017-02-22 2:00 GMT+03:00 Denis Magda <[hidden email] <mailto:
> [hidden email]>>:
> > Replied.
> >
> > —
> > Denis
> >
> > > On Feb 20, 2017, at 3:08 AM, Vladimir Ozerov <[hidden email]
> <mailto:[hidden email]>> wrote:
> > >
> > > Hi Vyacheslav,
> > >
> > > Thank you for contribution. I reviewed implementation again and now I
> am in
> > > doubts whether our product would really benefit from it or not. See my
> > > comments in the ticket. I'de prefer Denis Magda to chime in and give
> his
> > > feedback first.
> > >
> > > Vladimir.
> > >
> > > On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <
> [hidden email] <mailto:[hidden email]>>
> > > wrote:
> > >
> > >> Hello everyone.
> > >>
> > >> Please, review implemented solution.
> > >>
> > >> https://issues.apache.org/jira/browse/IGNITE-3422 <
> https://issues.apache.org/jira/browse/IGNITE-3422> - No way to control
> > >> object initialization during deserialization/unmarshalling
> > >>
> > >> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054 <
> http://ci.ignite.apache.org/viewQueued.html?itemId=461054>>
> > >>
> >
> >
>
>

[daradurvs]

Guys, any thoughts?

2017-02-23 13:22 GMT+03:00 Vyacheslav Daradur <[hidden email]>:

> I think it makes a sense to define factory in IgniteConfiguration, as the
> Map<Class, InstanceFactory> which will be ised when deserializing.
>
> Interface example:
>
> public intterface InstanceFactory {
>     newInstance(Class clazz, BinaryReader reader);
> }
>
> P.S.
> About "constructor(BinraryReader reader)" - I think there is no sense,
> because it just duplicate readObject method and we can't manage a reference
> of new instance.
> How about singlton-deserialization?
>
>
>
> 2017-02-23 10:16 GMT+03:00 Denis Magda <[hidden email]>:
>
>> In my understanding the goal is well-defined in the ticket. At the same
>> time we have a similar task prepared by Pavel for .NET:
>> https://issues.apache.org/jira/browse/IGNITE-3102 <
>> https://issues.apache.org/jira/browse/IGNITE-3102>
>>
>> So, what we need to agree on is how to proceed with the implementation.
>> Let us think this over and propose design.
>>
>> —
>> Denis
>>
>> > On Feb 22, 2017, at 2:41 AM, Vyacheslav Daradur <[hidden email]>
>> wrote:
>> >
>> > Guys, let's discuss a goal of this task.
>> >
>> > I need the task specification.
>> >
>> >
>> > 2017-02-22 2:00 GMT+03:00 Denis Magda <[hidden email] <mailto:
>> [hidden email]>>:
>> > Replied.
>> >
>> > —
>> > Denis
>> >
>> > > On Feb 20, 2017, at 3:08 AM, Vladimir Ozerov <[hidden email]
>> <mailto:[hidden email]>> wrote:
>> > >
>> > > Hi Vyacheslav,
>> > >
>> > > Thank you for contribution. I reviewed implementation again and now I
>> am in
>> > > doubts whether our product would really benefit from it or not. See my
>> > > comments in the ticket. I'de prefer Denis Magda to chime in and give
>> his
>> > > feedback first.
>> > >
>> > > Vladimir.
>> > >
>> > > On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <
>> [hidden email] <mailto:[hidden email]>>
>> > > wrote:
>> > >
>> > >> Hello everyone.
>> > >>
>> > >> Please, review implemented solution.
>> > >>
>> > >> https://issues.apache.org/jira/browse/IGNITE-3422 <
>> https://issues.apache.org/jira/browse/IGNITE-3422> - No way to control
>> > >> object initialization during deserialization/unmarshalling
>> > >>
>> > >> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054
>> <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>>
>> > >>
>> >
>> >
>>
>>
>

[dmagda]

Vyacheslav, sorry for the delay with a reply.

Let’s put off this ticket for now because, as I see it’s too controversial for us. If no one emerges and asks that he needs it then we will close it later on.

—
Denis

> On Mar 1, 2017, at 12:11 AM, Vyacheslav Daradur <[hidden email]> wrote:
>
> Guys, any thoughts?
>
> 2017-02-23 13:22 GMT+03:00 Vyacheslav Daradur <[hidden email]>:
>
>> I think it makes a sense to define factory in IgniteConfiguration, as the
>> Map<Class, InstanceFactory> which will be ised when deserializing.
>>
>> Interface example:
>>
>> public intterface InstanceFactory {
>>    newInstance(Class clazz, BinaryReader reader);
>> }
>>
>> P.S.
>> About "constructor(BinraryReader reader)" - I think there is no sense,
>> because it just duplicate readObject method and we can't manage a reference
>> of new instance.
>> How about singlton-deserialization?
>>
>>
>>
>> 2017-02-23 10:16 GMT+03:00 Denis Magda <[hidden email]>:
>>
>>> In my understanding the goal is well-defined in the ticket. At the same
>>> time we have a similar task prepared by Pavel for .NET:
>>> https://issues.apache.org/jira/browse/IGNITE-3102 <
>>> https://issues.apache.org/jira/browse/IGNITE-3102>
>>>
>>> So, what we need to agree on is how to proceed with the implementation.
>>> Let us think this over and propose design.
>>>
>>> —
>>> Denis
>>>
>>>> On Feb 22, 2017, at 2:41 AM, Vyacheslav Daradur <[hidden email]>
>>> wrote:
>>>>
>>>> Guys, let's discuss a goal of this task.
>>>>
>>>> I need the task specification.
>>>>
>>>>
>>>> 2017-02-22 2:00 GMT+03:00 Denis Magda <[hidden email] <mailto:
>>> [hidden email]>>:
>>>> Replied.
>>>>
>>>> —
>>>> Denis
>>>>
>>>>> On Feb 20, 2017, at 3:08 AM, Vladimir Ozerov <[hidden email]
>>> <mailto:[hidden email]>> wrote:
>>>>>
>>>>> Hi Vyacheslav,
>>>>>
>>>>> Thank you for contribution. I reviewed implementation again and now I
>>> am in
>>>>> doubts whether our product would really benefit from it or not. See my
>>>>> comments in the ticket. I'de prefer Denis Magda to chime in and give
>>> his
>>>>> feedback first.
>>>>>
>>>>> Vladimir.
>>>>>
>>>>> On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <
>>> [hidden email] <mailto:[hidden email]>>
>>>>> wrote:
>>>>>
>>>>>> Hello everyone.
>>>>>>
>>>>>> Please, review implemented solution.
>>>>>>
>>>>>> https://issues.apache.org/jira/browse/IGNITE-3422 <
>>> https://issues.apache.org/jira/browse/IGNITE-3422> - No way to control
>>>>>> object initialization during deserialization/unmarshalling
>>>>>>
>>>>>> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054
>>> <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>>
>>>>>>
>>>>
>>>>
>>>
>>>
>>

[daradurvs]

Ok, I put it off.

Feel free to contact me to continue this task.

2017-03-03 21:58 GMT+03:00 Denis Magda <[hidden email]>:

> Vyacheslav, sorry for the delay with a reply.
>
> Let’s put off this ticket for now because, as I see it’s too controversial
> for us. If no one emerges and asks that he needs it then we will close it
> later on.
>
> —
> Denis
>
> > On Mar 1, 2017, at 12:11 AM, Vyacheslav Daradur <[hidden email]>
> wrote:
> >
> > Guys, any thoughts?
> >
> > 2017-02-23 13:22 GMT+03:00 Vyacheslav Daradur <[hidden email]>:
> >
> >> I think it makes a sense to define factory in IgniteConfiguration, as
> the
> >> Map<Class, InstanceFactory> which will be ised when deserializing.
> >>
> >> Interface example:
> >>
> >> public intterface InstanceFactory {
> >>    newInstance(Class clazz, BinaryReader reader);
> >> }
> >>
> >> P.S.
> >> About "constructor(BinraryReader reader)" - I think there is no sense,
> >> because it just duplicate readObject method and we can't manage a
> reference
> >> of new instance.
> >> How about singlton-deserialization?
> >>
> >>
> >>
> >> 2017-02-23 10:16 GMT+03:00 Denis Magda <[hidden email]>:
> >>
> >>> In my understanding the goal is well-defined in the ticket. At the same
> >>> time we have a similar task prepared by Pavel for .NET:
> >>> https://issues.apache.org/jira/browse/IGNITE-3102 <
> >>> https://issues.apache.org/jira/browse/IGNITE-3102>
> >>>
> >>> So, what we need to agree on is how to proceed with the implementation.
> >>> Let us think this over and propose design.
> >>>
> >>> —
> >>> Denis
> >>>
> >>>> On Feb 22, 2017, at 2:41 AM, Vyacheslav Daradur <[hidden email]>
> >>> wrote:
> >>>>
> >>>> Guys, let's discuss a goal of this task.
> >>>>
> >>>> I need the task specification.
> >>>>
> >>>>
> >>>> 2017-02-22 2:00 GMT+03:00 Denis Magda <[hidden email] <mailto:
> >>> [hidden email]>>:
> >>>> Replied.
> >>>>
> >>>> —
> >>>> Denis
> >>>>
> >>>>> On Feb 20, 2017, at 3:08 AM, Vladimir Ozerov <[hidden email]
> >>> <mailto:[hidden email]>> wrote:
> >>>>>
> >>>>> Hi Vyacheslav,
> >>>>>
> >>>>> Thank you for contribution. I reviewed implementation again and now I
> >>> am in
> >>>>> doubts whether our product would really benefit from it or not. See
> my
> >>>>> comments in the ticket. I'de prefer Denis Magda to chime in and give
> >>> his
> >>>>> feedback first.
> >>>>>
> >>>>> Vladimir.
> >>>>>
> >>>>> On Wed, Feb 15, 2017 at 5:00 PM, Vyacheslav Daradur <
> >>> [hidden email] <mailto:[hidden email]>>
> >>>>> wrote:
> >>>>>
> >>>>>> Hello everyone.
> >>>>>>
> >>>>>> Please, review implemented solution.
> >>>>>>
> >>>>>> https://issues.apache.org/jira/browse/IGNITE-3422 <
> >>> https://issues.apache.org/jira/browse/IGNITE-3422> - No way to control
> >>>>>> object initialization during deserialization/unmarshalling
> >>>>>>
> >>>>>> ci.tests <http://ci.ignite.apache.org/viewQueued.html?itemId=461054
> >>> <http://ci.ignite.apache.org/viewQueued.html?itemId=461054>>
> >>>>>>
> >>>>
> >>>>
> >>>
> >>>
> >>
>
>