You are viewing a plain text version of this content. The canonical link for it is here.
Posted to notifications@logging.apache.org by Apache Jenkins Server <je...@builds.apache.org> on 2019/08/31 18:50:30 UTC
Jenkins build became unstable: Log4j 2 3.x #476
See <https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes>
Re: Jenkins build became unstable: Log4j 2 3.x #476
Posted by Matt Sicker <bo...@gmail.com>.
Just a small inconsistency I found while refactoring. I fixed the issue I
caused already.
On Sat, Aug 31, 2019 at 23:15, Ralph Goers <ra...@dslextreme.com>
wrote:
> I don’t understand. What is the problem? As you noted it has always been
> this way.
>
> Ralph
>
> > On Aug 31, 2019, at 9:03 PM, Matt Sicker <bo...@gmail.com> wrote:
> >
> > The plugin category is used to distinguish that in this API, not the
> class.
> > It seems as though both use the same interface except for the one class.
> > Generic type erasure avoided this from cropping up before.
> >
> > On Sat, Aug 31, 2019 at 20:50, Ralph Goers <ra...@dslextreme.com>
> > wrote:
> >
> >>
> >> Because it isn’t a PatternConverter? Look at what it does. The
> >> newInstance method creates a DatePatternConverter. The
> >> FileDatePatternConverter and IntegerPatternConverter are both used for
> the
> >> filePattern in the rolling file appender, not for a pattern layout.
> >>
> >> Ralph
> >>
> >>> On Aug 31, 2019, at 12:35 PM, Matt Sicker <bo...@gmail.com> wrote:
> >>>
> >>> FileDatePatternConverter, not DatePatternConverter.
> >>>
> >>> On Sat, 31 Aug 2019 at 14:35, Matt Sicker <bo...@gmail.com> wrote:
> >>>>
> >>>> Ah, this is an interesting historical design decision dating back to
> >>>> the very first commit! Why exactly does DatePatternConverter _not_
> >>>> implement PatternConverter like literally every other converter class
> >>>> does?
> >>>>
> >>>> On Sat, 31 Aug 2019 at 14:29, Matt Sicker <bo...@gmail.com> wrote:
> >>>>>
> >>>>> Shit, I'm on it.
> >>>>>
> >>>>> On Sat, 31 Aug 2019 at 13:50, Apache Jenkins Server
> >>>>> <je...@builds.apache.org> wrote:
> >>>>>>
> >>>>>> See <
> >>
> https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes
> >>>
> >>>>>>
> >>>>>
> >>>>>
> >>>>> --
> >>>>> Matt Sicker <bo...@gmail.com>
> >>>>
> >>>>
> >>>>
> >>>> --
> >>>> Matt Sicker <bo...@gmail.com>
> >>>
> >>>
> >>>
> >>> --
> >>> Matt Sicker <bo...@gmail.com>
> >>>
> >>
> >>
> >> --
> > Matt Sicker <bo...@gmail.com>
>
>
> --
Matt Sicker <bo...@gmail.com>
Re: Jenkins build became unstable: Log4j 2 3.x #476
Posted by Ralph Goers <ra...@dslextreme.com>.
I don’t understand. What is the problem? As you noted it has always been this way.
Ralph
> On Aug 31, 2019, at 9:03 PM, Matt Sicker <bo...@gmail.com> wrote:
>
> The plugin category is used to distinguish that in this API, not the class.
> It seems as though both use the same interface except for the one class.
> Generic type erasure avoided this from cropping up before.
>
> On Sat, Aug 31, 2019 at 20:50, Ralph Goers <ra...@dslextreme.com>
> wrote:
>
>>
>> Because it isn’t a PatternConverter? Look at what it does. The
>> newInstance method creates a DatePatternConverter. The
>> FileDatePatternConverter and IntegerPatternConverter are both used for the
>> filePattern in the rolling file appender, not for a pattern layout.
>>
>> Ralph
>>
>>> On Aug 31, 2019, at 12:35 PM, Matt Sicker <bo...@gmail.com> wrote:
>>>
>>> FileDatePatternConverter, not DatePatternConverter.
>>>
>>> On Sat, 31 Aug 2019 at 14:35, Matt Sicker <bo...@gmail.com> wrote:
>>>>
>>>> Ah, this is an interesting historical design decision dating back to
>>>> the very first commit! Why exactly does DatePatternConverter _not_
>>>> implement PatternConverter like literally every other converter class
>>>> does?
>>>>
>>>> On Sat, 31 Aug 2019 at 14:29, Matt Sicker <bo...@gmail.com> wrote:
>>>>>
>>>>> Shit, I'm on it.
>>>>>
>>>>> On Sat, 31 Aug 2019 at 13:50, Apache Jenkins Server
>>>>> <je...@builds.apache.org> wrote:
>>>>>>
>>>>>> See <
>> https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes
>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Matt Sicker <bo...@gmail.com>
>>>>
>>>>
>>>>
>>>> --
>>>> Matt Sicker <bo...@gmail.com>
>>>
>>>
>>>
>>> --
>>> Matt Sicker <bo...@gmail.com>
>>>
>>
>>
>> --
> Matt Sicker <bo...@gmail.com>
Re: Jenkins build became unstable: Log4j 2 3.x #476
Posted by Matt Sicker <bo...@gmail.com>.
The plugin category is used to distinguish that in this API, not the class.
It seems as though both use the same interface except for the one class.
Generic type erasure avoided this from cropping up before.
On Sat, Aug 31, 2019 at 20:50, Ralph Goers <ra...@dslextreme.com>
wrote:
>
> Because it isn’t a PatternConverter? Look at what it does. The
> newInstance method creates a DatePatternConverter. The
> FileDatePatternConverter and IntegerPatternConverter are both used for the
> filePattern in the rolling file appender, not for a pattern layout.
>
> Ralph
>
> > On Aug 31, 2019, at 12:35 PM, Matt Sicker <bo...@gmail.com> wrote:
> >
> > FileDatePatternConverter, not DatePatternConverter.
> >
> > On Sat, 31 Aug 2019 at 14:35, Matt Sicker <bo...@gmail.com> wrote:
> >>
> >> Ah, this is an interesting historical design decision dating back to
> >> the very first commit! Why exactly does DatePatternConverter _not_
> >> implement PatternConverter like literally every other converter class
> >> does?
> >>
> >> On Sat, 31 Aug 2019 at 14:29, Matt Sicker <bo...@gmail.com> wrote:
> >>>
> >>> Shit, I'm on it.
> >>>
> >>> On Sat, 31 Aug 2019 at 13:50, Apache Jenkins Server
> >>> <je...@builds.apache.org> wrote:
> >>>>
> >>>> See <
> https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes
> >
> >>>>
> >>>
> >>>
> >>> --
> >>> Matt Sicker <bo...@gmail.com>
> >>
> >>
> >>
> >> --
> >> Matt Sicker <bo...@gmail.com>
> >
> >
> >
> > --
> > Matt Sicker <bo...@gmail.com>
> >
>
>
> --
Matt Sicker <bo...@gmail.com>
Re: Jenkins build became unstable: Log4j 2 3.x #476
Posted by Ralph Goers <ra...@dslextreme.com>.
Because it isn’t a PatternConverter? Look at what it does. The newInstance method creates a DatePatternConverter. The FileDatePatternConverter and IntegerPatternConverter are both used for the filePattern in the rolling file appender, not for a pattern layout.
Ralph
> On Aug 31, 2019, at 12:35 PM, Matt Sicker <bo...@gmail.com> wrote:
>
> FileDatePatternConverter, not DatePatternConverter.
>
> On Sat, 31 Aug 2019 at 14:35, Matt Sicker <bo...@gmail.com> wrote:
>>
>> Ah, this is an interesting historical design decision dating back to
>> the very first commit! Why exactly does DatePatternConverter _not_
>> implement PatternConverter like literally every other converter class
>> does?
>>
>> On Sat, 31 Aug 2019 at 14:29, Matt Sicker <bo...@gmail.com> wrote:
>>>
>>> Shit, I'm on it.
>>>
>>> On Sat, 31 Aug 2019 at 13:50, Apache Jenkins Server
>>> <je...@builds.apache.org> wrote:
>>>>
>>>> See <https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes>
>>>>
>>>
>>>
>>> --
>>> Matt Sicker <bo...@gmail.com>
>>
>>
>>
>> --
>> Matt Sicker <bo...@gmail.com>
>
>
>
> --
> Matt Sicker <bo...@gmail.com>
>
Re: Jenkins build became unstable: Log4j 2 3.x #476
Posted by Matt Sicker <bo...@gmail.com>.
FileDatePatternConverter, not DatePatternConverter.
On Sat, 31 Aug 2019 at 14:35, Matt Sicker <bo...@gmail.com> wrote:
>
> Ah, this is an interesting historical design decision dating back to
> the very first commit! Why exactly does DatePatternConverter _not_
> implement PatternConverter like literally every other converter class
> does?
>
> On Sat, 31 Aug 2019 at 14:29, Matt Sicker <bo...@gmail.com> wrote:
> >
> > Shit, I'm on it.
> >
> > On Sat, 31 Aug 2019 at 13:50, Apache Jenkins Server
> > <je...@builds.apache.org> wrote:
> > >
> > > See <https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes>
> > >
> >
> >
> > --
> > Matt Sicker <bo...@gmail.com>
>
>
>
> --
> Matt Sicker <bo...@gmail.com>
--
Matt Sicker <bo...@gmail.com>
Re: Jenkins build became unstable: Log4j 2 3.x #476
Posted by Matt Sicker <bo...@gmail.com>.
Ah, this is an interesting historical design decision dating back to
the very first commit! Why exactly does DatePatternConverter _not_
implement PatternConverter like literally every other converter class
does?
On Sat, 31 Aug 2019 at 14:29, Matt Sicker <bo...@gmail.com> wrote:
>
> Shit, I'm on it.
>
> On Sat, 31 Aug 2019 at 13:50, Apache Jenkins Server
> <je...@builds.apache.org> wrote:
> >
> > See <https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes>
> >
>
>
> --
> Matt Sicker <bo...@gmail.com>
--
Matt Sicker <bo...@gmail.com>
Re: Jenkins build became unstable: Log4j 2 3.x #476
Posted by Matt Sicker <bo...@gmail.com>.
Shit, I'm on it.
On Sat, 31 Aug 2019 at 13:50, Apache Jenkins Server
<je...@builds.apache.org> wrote:
>
> See <https://builds.apache.org/job/Log4j%202%203.x/476/display/redirect?page=changes>
>
--
Matt Sicker <bo...@gmail.com>