t;>>>>>> implementing Logger is serializable, then I am good to go.
>>>>>>>>>
>>>>>>>>> Making Logger Serializable, would make this obvious to users and
>>>>>>>>> might make some developers warm and fuz
er Serializable, would make this obvious to users and
>>>>>>>> might make some developers warm and fuzzy. It just seems like an odd
>>>>>>>> requirement for ALL implementors of Loggers to be serializable. As a
>>>>>>>> third
>
zable. As a
>>>>>>> third
>>>>>>> party Logger implementor, that does not mean it would be done properly.
>>>>>>> Implementing the interface is no guarantee of proper behavior.
>>>>>>>
>>>>>>> Ov
>>>>>
>>>>>> Overall, it seems like it would make Log4j less friendly to 3rd party
>>>>>> implementors. Are there even any?
>>>>>>
>>>>>> Gary
>>>>>>
>>>>>> On Mon, Jul 25, 2016 at
t;>> implementors. Are there even any?
>>>>>
>>>>> Gary
>>>>>
>>>>> On Mon, Jul 25, 2016 at 9:07 AM, Mikael Ståldal <
>>>>> mikael.stal...@magine.com> wrote:
>>>>>
>>>>>> B
>>>>> properly. Implementing the interface is no guarantee of proper
>>>>>>> behavior.
>>>>>>>
>>>>>>> Overall, it seems like it would make Log4j less friendly to 3rd party
>>>>>>> implementors. Are th
proper
>>>>>>> behavior.
>>>>>>>
>>>>>>> Overall, it seems like it would make Log4j less friendly to 3rd party
>>>>>>> implementors. Are there even any?
>>>>>>>
>>>
;>> Gary
>>>>
>>>> On Mon, Jul 25, 2016 at 9:07 AM, Mikael Ståldal <
>>>> mikael.stal...@magine.com> wrote:
>>>>
>>>>> But maybe we actually want to force all implementations to be
>>>>> serializable?
>>
agine.com> wrote:
>>>
>>>> But maybe we actually want to force all implementations to be
>>>> serializable?
>>>>
>>>> On Mon, Jul 25, 2016 at 6:01 PM, Remko Popma
>>>> wrote:
>>>>
>>>>> I think the th
at 6:01 PM, Remko Popma
>>>>>> wrote:
>>>>>> I think the theory is it is preferable to let implementors of an
>>>>>> interface implement java.io.Serializable instead of letting the
>>>>>> interface extend java.io.Serializa
let implementors of an
>>>> interface implement java.io.Serializable instead of letting the interface
>>>> extend java.io.Serializable. That gives future implementors the freedom to
>>>> choose to either implement Serializable or not.
>>>>
>>>&
n, Jul 25, 2016 at 6:01 PM, Remko Popma
>>> wrote:
>>>
>>>> I think the theory is it is preferable to let implementors of an
>>>> interface implement java.io.Serializable instead of letting the interface
>>>> extend java.io.Serializable. That gives
rote:
>>>> I think the theory is it is preferable to let implementors of an interface
>>>> implement java.io.Serializable instead of letting the interface extend
>>>> java.io.Serializable. That gives future implementors the freedom to choose
>>>> to either imple
io.Serializable instead of letting the interface extend
> java.io.Serializable. That gives future implementors the freedom to choose to
> either implement Serializable or not.
>
> On Tue, Jul 26, 2016 at 12:49 AM, Mikael Ståldal <mailto:mikael.stal...@magine.com>> wrote:
&g
t;> On Mon, Jul 25, 2016 at 6:01 PM, Remko Popma
>>>> wrote:
>>>>
>>>>> I think the theory is it is preferable to let implementors of an
>>>>> interface implement java.io.Serializable instead of letting the interface
>>>>> exten
t;>> I think the theory is it is preferable to let implementors of an
>>>> interface implement java.io.Serializable instead of letting the interface
>>>> extend java.io.Serializable. That gives future implementors the freedom to
>>>> choose to either implement Se
mentors of an
>>> interface implement java.io.Serializable instead of letting the interface
>>> extend java.io.Serializable. That gives future implementors the freedom to
>>> choose to either implement Serializable or not.
>>>
>>> On Tue, Jul 26, 2016 at 12:49 AM, Mi
, 2016 at 12:49 AM, Mikael Ståldal <
>> mikael.stal...@magine.com> wrote:
>>
>>> It seems like serializable logger is useful in Apache Spark, which is
>>> popular to use from Scala.
>>>
>>> On Mon, Jul 25, 2016 at 4:20 PM, Matt Sicker wrote:
&g
extend
> java.io.Serializable. That gives future implementors the freedom to choose
> to either implement Serializable or not.
>
> On Tue, Jul 26, 2016 at 12:49 AM, Mikael Ståldal <
> mikael.stal...@magine.com> wrote:
>
>> It seems like serializable logger is useful in Apache Spark, w
, Mikael Ståldal
wrote:
> It seems like serializable logger is useful in Apache Spark, which is
> popular to use from Scala.
>
> On Mon, Jul 25, 2016 at 4:20 PM, Matt Sicker wrote:
>
>> Do Scala programmers use Serializable a lot, or is there a better
>> mechanism there
It seems like serializable logger is useful in Apache Spark, which is
popular to use from Scala.
On Mon, Jul 25, 2016 at 4:20 PM, Matt Sicker wrote:
> Do Scala programmers use Serializable a lot, or is there a better
> mechanism there? Because if there's a more Scala way of doi
It feels like we need an epic to get our overall story straight on
serialization, including docs on what works, doesn't, and what we plan to
have working.
Gary
On Jul 25, 2016 1:06 AM, "Mikael Ståldal" wrote:
> I am asking this since I'm considering making the Scala Logger wrapper
> Serializabl
Do Scala programmers use Serializable a lot, or is there a better mechanism
there? Because if there's a more Scala way of doing it, that would make
more sense. Otherwise, you can always add Serializable later, but you
shouldn't remove it once it's part of the public API.
On 25 July 2016 at 03:06,
I am asking this since I'm considering making the Scala Logger wrapper
Serializable. Would that be a good idea?
On Mon, Jul 25, 2016 at 10:00 AM, Mikael Ståldal
wrote:
> Users can rely on the fact that loggers are always Serializable.
>
> I guess we could also document that fact if we don't want
Users can rely on the fact that loggers are always Serializable.
I guess we could also document that fact if we don't want to mark the
interface Serializable.
On Fri, Jul 22, 2016 at 7:16 PM, Ralph Goers
wrote:
> What advantage is gained by declaring the interface Serializable?
>
> Ralph
>
> On
What advantage is gained by declaring the interface Serializable?
Ralph
> On Jul 22, 2016, at 9:38 AM, Mikael Ståldal wrote:
>
> I don't think we should limit ourselves to what we add to the API just for
> the sake of alternative implementations.
>
> On Fri, Jul 22, 2016 at 6:09 PM, Gary Greg
I don't think we should limit ourselves to what we add to the API just for
the sake of alternative implementations.
On Fri, Jul 22, 2016 at 6:09 PM, Gary Gregory
wrote:
> I do not think you want to _force_ all implementation to be serializable.
> Serialization is not a main feature of loggers, l
I do not think you want to _force_ all implementation to be serializable.
Serialization is not a main feature of loggers, logging is.
Gary
On Fri, Jul 22, 2016 at 12:45 AM, Mikael Ståldal
wrote:
> But if this is a feature that is useful for some users, why not fully
> advertise it by making the
Is that generally a followed pattern? The JDK itself isn't even consistent
in that regard (e.g., Lock isn't serializable, but all its implementations
are).
On 22 July 2016 at 02:45, Mikael Ståldal wrote:
> But if this is a feature that is useful for some users, why not fully
> advertise it by ma
But if this is a feature that is useful for some users, why not fully
advertise it by making the Logger interface Serializable?
On Thu, Jul 21, 2016 at 7:53 PM, Matt Sicker wrote:
> It's Serializable because there was a feature request for it. The jira
> ticket is internal (reported by Gary), bu
It's Serializable because there was a feature request for it. The jira
ticket is internal (reported by Gary), but I recall either another bug or a
mailing list post where the justification was so that serializable classes
can still use a Logger instance field (not everyone uses a static field for
t
I asked myself the same question and I guess the answer is that Logger is
included in lots of classes, many of which are Serializable. It would be easier
for us to make AbstractLogger Serializable then for all the users to mark it as
transient and have to manually instantiate the Logger during d
So then why is AbstractLogger, and our implementation in log4j-core,
Serializable?
On Thu, Jul 21, 2016 at 6:18 PM, Gary Gregory
wrote:
> I do not think there is a way to know what folks are doing out there...
>
> Gary
>
> On Thu, Jul 21, 2016 at 9:13 AM, Mikael Ståldal > wrote:
>
>> Are there
I do not think there is a way to know what folks are doing out there...
Gary
On Thu, Jul 21, 2016 at 9:13 AM, Mikael Ståldal
wrote:
> Are there any providers not using AbstractLogger?
>
> On Thu, Jul 21, 2016 at 6:09 PM, Gary Gregory
> wrote:
>
>> Probably to leave the decision on whether to s
Are there any providers not using AbstractLogger?
On Thu, Jul 21, 2016 at 6:09 PM, Gary Gregory
wrote:
> Probably to leave the decision on whether to support serialization to each
> provider.
>
> Gary
>
> On Jul 21, 2016 3:14 AM, "Mikael Ståldal"
> wrote:
>
>> AbstractLogger is Serializable, bu
Probably to leave the decision on whether to support serialization to each
provider.
Gary
On Jul 21, 2016 3:14 AM, "Mikael Ståldal" wrote:
> AbstractLogger is Serializable, but the Logger interface is not. Why is it
> so?
>
> --
> [image: MagineTV]
>
> *Mikael Ståldal*
> Senior software develop
AbstractLogger is Serializable, but the Logger interface is not. Why is it
so?
--
[image: MagineTV]
*Mikael Ståldal*
Senior software developer
*Magine TV*
mikael.stal...@magine.com
Grev Turegatan 3 | 114 46 Stockholm, Sweden | www.magine.com
Privileged and/or Confidential Information may b
37 matches
Mail list logo