[VOTE][RC2] Commons collections 4.3

classic Classic list List threaded Threaded
71 messages Options
1234
Reply | Threaded
Open this post in threaded view
|

Re: [VOTE][RC2] Commons collections 4.3

garydgregory
We might as well start with releasing fresh versions of our internal maven
plugins, then commons-parent, then components... ;-) More work for the RM!
:-)

Gary

On Sun, Jan 20, 2019 at 5:06 AM Gilles Sadowski <[hidden email]>
wrote:

> Le dim. 20 janv. 2019 à 04:56, Maxim Solodovnik <[hidden email]> a
> écrit :
> >
> > I can see there is activity with replacing clirr with japicmp, maybe it
> > worth to hold this release until commons-parent v48 will be released?
>
> You are the manager. :-)
>
> Gilles
>
> >
> > On Sun, 20 Jan 2019 at 05:52, Gilles Sadowski <[hidden email]>
> wrote:
> >
> > > Hi.
> > >
> > > Le sam. 19 janv. 2019 à 11:36, Maxim Solodovnik <[hidden email]>
> a
> > > écrit :
> > > >
> > > > I'm afraid I need help with this
> > > > since I don't understand why clirr reports these errors
> > >
> > > Nobody knows around here; but your test demonstrates that one of the
> > > errors is a false positive.
> > > So I suggest that, for good measure, you add calls to all the methods
> > > reported by Clirr; then the release notes can rightly mention something
> > > like: "An integration test (<short description of what you've done
> and/or
> > > a link to this discussion>) has shown that the errors mentioned in the
> > > Clirr report are false positives."
> > >
> > > Best regards,
> > > Gilles
> > >
> > >
> > > > > [...]
> > >
> >
> > --
> > WBR
> > Maxim aka solomax
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>
Reply | Threaded
Open this post in threaded view
|

JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Rob Tompkins
Just an FYI, I hacked this together for the sake of reporting on solomax’s work:

Cheers,
-Rob




> On Jan 20, 2019, at 8:48 AM, Gary Gregory <[hidden email]> wrote:
>
> We might as well start with releasing fresh versions of our internal maven
> plugins, then commons-parent, then components... ;-) More work for the RM!
> :-)
>
> Gary
>
> On Sun, Jan 20, 2019 at 5:06 AM Gilles Sadowski <[hidden email]>
> wrote:
>
>> Le dim. 20 janv. 2019 à 04:56, Maxim Solodovnik <[hidden email]> a
>> écrit :
>>>
>>> I can see there is activity with replacing clirr with japicmp, maybe it
>>> worth to hold this release until commons-parent v48 will be released?
>>
>> You are the manager. :-)
>>
>> Gilles
>>
>>>
>>> On Sun, 20 Jan 2019 at 05:52, Gilles Sadowski <[hidden email]>
>> wrote:
>>>
>>>> Hi.
>>>>
>>>> Le sam. 19 janv. 2019 à 11:36, Maxim Solodovnik <[hidden email]>
>> a
>>>> écrit :
>>>>>
>>>>> I'm afraid I need help with this
>>>>> since I don't understand why clirr reports these errors
>>>>
>>>> Nobody knows around here; but your test demonstrates that one of the
>>>> errors is a false positive.
>>>> So I suggest that, for good measure, you add calls to all the methods
>>>> reported by Clirr; then the release notes can rightly mention something
>>>> like: "An integration test (<short description of what you've done
>> and/or
>>>> a link to this discussion>) has shown that the errors mentioned in the
>>>> Clirr report are false positives."
>>>>
>>>> Best regards,
>>>> Gilles
>>>>
>>>>
>>>>>> [...]
>>>>
>>>
>>> --
>>> WBR
>>> Maxim aka solomax
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>



---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]
Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Maxim Solodovnik
Hello Rob,

can you please share brief how-to, so I can regenerate the report? :)

On Sat, 26 Jan 2019 at 02:04, Rob Tompkins <[hidden email]> wrote:

> Just an FYI, I hacked this together for the sake of reporting on solomax’s
> work:
>
> Cheers,
> -Rob
>
>
>
> > On Jan 20, 2019, at 8:48 AM, Gary Gregory <[hidden email]>
> wrote:
> >
> > We might as well start with releasing fresh versions of our internal
> maven
> > plugins, then commons-parent, then components... ;-) More work for the
> RM!
> > :-)
> >
> > Gary
> >
> > On Sun, Jan 20, 2019 at 5:06 AM Gilles Sadowski <[hidden email]>
> > wrote:
> >
> >> Le dim. 20 janv. 2019 à 04:56, Maxim Solodovnik <[hidden email]>
> a
> >> écrit :
> >>>
> >>> I can see there is activity with replacing clirr with japicmp, maybe it
> >>> worth to hold this release until commons-parent v48 will be released?
> >>
> >> You are the manager. :-)
> >>
> >> Gilles
> >>
> >>>
> >>> On Sun, 20 Jan 2019 at 05:52, Gilles Sadowski <[hidden email]>
> >> wrote:
> >>>
> >>>> Hi.
> >>>>
> >>>> Le sam. 19 janv. 2019 à 11:36, Maxim Solodovnik <[hidden email]
> >
> >> a
> >>>> écrit :
> >>>>>
> >>>>> I'm afraid I need help with this
> >>>>> since I don't understand why clirr reports these errors
> >>>>
> >>>> Nobody knows around here; but your test demonstrates that one of the
> >>>> errors is a false positive.
> >>>> So I suggest that, for good measure, you add calls to all the methods
> >>>> reported by Clirr; then the release notes can rightly mention
> something
> >>>> like: "An integration test (<short description of what you've done
> >> and/or
> >>>> a link to this discussion>) has shown that the errors mentioned in the
> >>>> Clirr report are false positives."
> >>>>
> >>>> Best regards,
> >>>> Gilles
> >>>>
> >>>>
> >>>>>> [...]
> >>>>
> >>>
> >>> --
> >>> WBR
> >>> Maxim aka solomax
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: [hidden email]
> >> For additional commands, e-mail: [hidden email]
> >>
> >>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]



--
WBR
Maxim aka solomax
Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Rob Tompkins
Sure.

I cloned down your tag, added these lines to the pom.xml:

https://github.com/apache/commons-text/blob/master/pom.xml#L59-L63

and

https://github.com/apache/commons-text/blob/master/pom.xml#L132-L138 (except just in the plugins part of the build section)

then you run
mvn japicmp:cmp

I couldn’t get the report to work, but that put something in the target directory that’s an html file. I then opened it in the browser and converted it to a pdf for the sake of sending it to the ML.

-Rob

> On Jan 25, 2019, at 11:33 PM, Maxim Solodovnik <[hidden email]> wrote:
>
> Hello Rob,
>
> can you please share brief how-to, so I can regenerate the report? :)
>
> On Sat, 26 Jan 2019 at 02:04, Rob Tompkins <[hidden email]> wrote:
>
>> Just an FYI, I hacked this together for the sake of reporting on solomax’s
>> work:
>>
>> Cheers,
>> -Rob
>>
>>
>>
>>> On Jan 20, 2019, at 8:48 AM, Gary Gregory <[hidden email]>
>> wrote:
>>>
>>> We might as well start with releasing fresh versions of our internal
>> maven
>>> plugins, then commons-parent, then components... ;-) More work for the
>> RM!
>>> :-)
>>>
>>> Gary
>>>
>>> On Sun, Jan 20, 2019 at 5:06 AM Gilles Sadowski <[hidden email]>
>>> wrote:
>>>
>>>> Le dim. 20 janv. 2019 à 04:56, Maxim Solodovnik <[hidden email]>
>> a
>>>> écrit :
>>>>>
>>>>> I can see there is activity with replacing clirr with japicmp, maybe it
>>>>> worth to hold this release until commons-parent v48 will be released?
>>>>
>>>> You are the manager. :-)
>>>>
>>>> Gilles
>>>>
>>>>>
>>>>> On Sun, 20 Jan 2019 at 05:52, Gilles Sadowski <[hidden email]>
>>>> wrote:
>>>>>
>>>>>> Hi.
>>>>>>
>>>>>> Le sam. 19 janv. 2019 à 11:36, Maxim Solodovnik <[hidden email]
>>>
>>>> a
>>>>>> écrit :
>>>>>>>
>>>>>>> I'm afraid I need help with this
>>>>>>> since I don't understand why clirr reports these errors
>>>>>>
>>>>>> Nobody knows around here; but your test demonstrates that one of the
>>>>>> errors is a false positive.
>>>>>> So I suggest that, for good measure, you add calls to all the methods
>>>>>> reported by Clirr; then the release notes can rightly mention
>> something
>>>>>> like: "An integration test (<short description of what you've done
>>>> and/or
>>>>>> a link to this discussion>) has shown that the errors mentioned in the
>>>>>> Clirr report are false positives."
>>>>>>
>>>>>> Best regards,
>>>>>> Gilles
>>>>>>
>>>>>>
>>>>>>>> [...]
>>>>>>
>>>>>
>>>>> --
>>>>> WBR
>>>>> Maxim aka solomax
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>>
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>
>
>
> --
> WBR
> Maxim aka solomax


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Maxim Solodovnik
Thanks a million!

On Sun, 27 Jan 2019 at 00:07, Rob Tompkins <[hidden email]> wrote:

> Sure.
>
> I cloned down your tag, added these lines to the pom.xml:
>
> https://github.com/apache/commons-text/blob/master/pom.xml#L59-L63
>
> and
>
> https://github.com/apache/commons-text/blob/master/pom.xml#L132-L138
> (except just in the plugins part of the build section)
>
> then you run
> mvn japicmp:cmp
>
> I couldn’t get the report to work, but that put something in the target
> directory that’s an html file. I then opened it in the browser and
> converted it to a pdf for the sake of sending it to the ML.
>
> -Rob
>
> > On Jan 25, 2019, at 11:33 PM, Maxim Solodovnik <[hidden email]>
> wrote:
> >
> > Hello Rob,
> >
> > can you please share brief how-to, so I can regenerate the report? :)
> >
> > On Sat, 26 Jan 2019 at 02:04, Rob Tompkins <[hidden email]> wrote:
> >
> >> Just an FYI, I hacked this together for the sake of reporting on
> solomax’s
> >> work:
> >>
> >> Cheers,
> >> -Rob
> >>
> >>
> >>
> >>> On Jan 20, 2019, at 8:48 AM, Gary Gregory <[hidden email]>
> >> wrote:
> >>>
> >>> We might as well start with releasing fresh versions of our internal
> >> maven
> >>> plugins, then commons-parent, then components... ;-) More work for the
> >> RM!
> >>> :-)
> >>>
> >>> Gary
> >>>
> >>> On Sun, Jan 20, 2019 at 5:06 AM Gilles Sadowski <[hidden email]>
> >>> wrote:
> >>>
> >>>> Le dim. 20 janv. 2019 à 04:56, Maxim Solodovnik <[hidden email]
> >
> >> a
> >>>> écrit :
> >>>>>
> >>>>> I can see there is activity with replacing clirr with japicmp, maybe
> it
> >>>>> worth to hold this release until commons-parent v48 will be released?
> >>>>
> >>>> You are the manager. :-)
> >>>>
> >>>> Gilles
> >>>>
> >>>>>
> >>>>> On Sun, 20 Jan 2019 at 05:52, Gilles Sadowski <[hidden email]>
> >>>> wrote:
> >>>>>
> >>>>>> Hi.
> >>>>>>
> >>>>>> Le sam. 19 janv. 2019 à 11:36, Maxim Solodovnik <
> [hidden email]
> >>>
> >>>> a
> >>>>>> écrit :
> >>>>>>>
> >>>>>>> I'm afraid I need help with this
> >>>>>>> since I don't understand why clirr reports these errors
> >>>>>>
> >>>>>> Nobody knows around here; but your test demonstrates that one of the
> >>>>>> errors is a false positive.
> >>>>>> So I suggest that, for good measure, you add calls to all the
> methods
> >>>>>> reported by Clirr; then the release notes can rightly mention
> >> something
> >>>>>> like: "An integration test (<short description of what you've done
> >>>> and/or
> >>>>>> a link to this discussion>) has shown that the errors mentioned in
> the
> >>>>>> Clirr report are false positives."
> >>>>>>
> >>>>>> Best regards,
> >>>>>> Gilles
> >>>>>>
> >>>>>>
> >>>>>>>> [...]
> >>>>>>
> >>>>>
> >>>>> --
> >>>>> WBR
> >>>>> Maxim aka solomax
> >>>>
> >>>> ---------------------------------------------------------------------
> >>>> To unsubscribe, e-mail: [hidden email]
> >>>> For additional commands, e-mail: [hidden email]
> >>>>
> >>>>
> >>
> >>
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: [hidden email]
> >> For additional commands, e-mail: [hidden email]
> >
> >
> >
> > --
> > WBR
> > Maxim aka solomax
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

--
WBR
Maxim aka solomax
Reply | Threaded
Open this post in threaded view
|

Re: [VOTE][RC2] Commons collections 4.3

Maxim Solodovnik
In reply to this post by garydgregory
Hello All,

Thanks to the Rob, I was able to create "compatibility report" using japicmp
It looks green :)

Would it be OK if I will replace clirr report in generated site with
japicmp report?
Or I need to restart the VOTE?

On Sun, 20 Jan 2019 at 20:48, Gary Gregory <[hidden email]> wrote:

> We might as well start with releasing fresh versions of our internal maven
> plugins, then commons-parent, then components... ;-) More work for the RM!
> :-)
>
> Gary
>
> On Sun, Jan 20, 2019 at 5:06 AM Gilles Sadowski <[hidden email]>
> wrote:
>
> > Le dim. 20 janv. 2019 à 04:56, Maxim Solodovnik <[hidden email]> a
> > écrit :
> > >
> > > I can see there is activity with replacing clirr with japicmp, maybe it
> > > worth to hold this release until commons-parent v48 will be released?
> >
> > You are the manager. :-)
> >
> > Gilles
> >
> > >
> > > On Sun, 20 Jan 2019 at 05:52, Gilles Sadowski <[hidden email]>
> > wrote:
> > >
> > > > Hi.
> > > >
> > > > Le sam. 19 janv. 2019 à 11:36, Maxim Solodovnik <
> [hidden email]>
> > a
> > > > écrit :
> > > > >
> > > > > I'm afraid I need help with this
> > > > > since I don't understand why clirr reports these errors
> > > >
> > > > Nobody knows around here; but your test demonstrates that one of the
> > > > errors is a false positive.
> > > > So I suggest that, for good measure, you add calls to all the methods
> > > > reported by Clirr; then the release notes can rightly mention
> something
> > > > like: "An integration test (<short description of what you've done
> > and/or
> > > > a link to this discussion>) has shown that the errors mentioned in
> the
> > > > Clirr report are false positives."
> > > >
> > > > Best regards,
> > > > Gilles
> > > >
> > > >
> > > > > > [...]
> > > >
> > >
> > > --
> > > WBR
> > > Maxim aka solomax
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
> >
>


--
WBR
Maxim aka solomax
Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Rob Tompkins
In reply to this post by Maxim Solodovnik
Before I vote on the the thread, does adding a method to an interface cause BC to break? I would think not. All of the clirr errors are merely additions. Further the JAPICMP report confirms this.

I’m validating the remainder of the release now, and wanted to ask the above question before adding in my vote.

Cheers,
-Rob

> On Jan 27, 2019, at 7:45 AM, Maxim Solodovnik <[hidden email]> wrote:
>
> Thanks a million!
>
> On Sun, 27 Jan 2019 at 00:07, Rob Tompkins <[hidden email]> wrote:
>
>> Sure.
>>
>> I cloned down your tag, added these lines to the pom.xml:
>>
>> https://github.com/apache/commons-text/blob/master/pom.xml#L59-L63
>>
>> and
>>
>> https://github.com/apache/commons-text/blob/master/pom.xml#L132-L138
>> (except just in the plugins part of the build section)
>>
>> then you run
>> mvn japicmp:cmp
>>
>> I couldn’t get the report to work, but that put something in the target
>> directory that’s an html file. I then opened it in the browser and
>> converted it to a pdf for the sake of sending it to the ML.
>>
>> -Rob
>>
>>> On Jan 25, 2019, at 11:33 PM, Maxim Solodovnik <[hidden email]>
>> wrote:
>>>
>>> Hello Rob,
>>>
>>> can you please share brief how-to, so I can regenerate the report? :)
>>>
>>> On Sat, 26 Jan 2019 at 02:04, Rob Tompkins <[hidden email]> wrote:
>>>
>>>> Just an FYI, I hacked this together for the sake of reporting on
>> solomax’s
>>>> work:
>>>>
>>>> Cheers,
>>>> -Rob
>>>>
>>>>
>>>>
>>>>> On Jan 20, 2019, at 8:48 AM, Gary Gregory <[hidden email]>
>>>> wrote:
>>>>>
>>>>> We might as well start with releasing fresh versions of our internal
>>>> maven
>>>>> plugins, then commons-parent, then components... ;-) More work for the
>>>> RM!
>>>>> :-)
>>>>>
>>>>> Gary
>>>>>
>>>>> On Sun, Jan 20, 2019 at 5:06 AM Gilles Sadowski <[hidden email]>
>>>>> wrote:
>>>>>
>>>>>> Le dim. 20 janv. 2019 à 04:56, Maxim Solodovnik <[hidden email]
>>>
>>>> a
>>>>>> écrit :
>>>>>>>
>>>>>>> I can see there is activity with replacing clirr with japicmp, maybe
>> it
>>>>>>> worth to hold this release until commons-parent v48 will be released?
>>>>>>
>>>>>> You are the manager. :-)
>>>>>>
>>>>>> Gilles
>>>>>>
>>>>>>>
>>>>>>> On Sun, 20 Jan 2019 at 05:52, Gilles Sadowski <[hidden email]>
>>>>>> wrote:
>>>>>>>
>>>>>>>> Hi.
>>>>>>>>
>>>>>>>> Le sam. 19 janv. 2019 à 11:36, Maxim Solodovnik <
>> [hidden email]
>>>>>
>>>>>> a
>>>>>>>> écrit :
>>>>>>>>>
>>>>>>>>> I'm afraid I need help with this
>>>>>>>>> since I don't understand why clirr reports these errors
>>>>>>>>
>>>>>>>> Nobody knows around here; but your test demonstrates that one of the
>>>>>>>> errors is a false positive.
>>>>>>>> So I suggest that, for good measure, you add calls to all the
>> methods
>>>>>>>> reported by Clirr; then the release notes can rightly mention
>>>> something
>>>>>>>> like: "An integration test (<short description of what you've done
>>>>>> and/or
>>>>>>>> a link to this discussion>) has shown that the errors mentioned in
>> the
>>>>>>>> Clirr report are false positives."
>>>>>>>>
>>>>>>>> Best regards,
>>>>>>>> Gilles
>>>>>>>>
>>>>>>>>
>>>>>>>>>> [...]
>>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> WBR
>>>>>>> Maxim aka solomax
>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: [hidden email]
>>>>>> For additional commands, e-mail: [hidden email]
>>>>>>
>>>>>>
>>>>
>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>>
>>> --
>>> WBR
>>> Maxim aka solomax
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>>
>
> --
> WBR
> Maxim aka solomax


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Marcelo Vanzin-2
On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]> wrote:
> Before I vote on the the thread, does adding a method to an interface cause BC to break? I would think not. All of the clirr errors are merely additions. Further the JAPICMP report confirms this.

Existing classes that implement the interface won't have the newly
added method. So if some other code calls that method in one of these
implementations, they'll get an "AbstractMethodError" (or something
similar).

--
Marcelo

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Rob Tompkins
@Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like this release should be a -1 then because we’re breaking BC without a major version change. Right??

-Rob


> On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin <[hidden email]> wrote:
>
> On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]> wrote:
>> Before I vote on the the thread, does adding a method to an interface cause BC to break? I would think not. All of the clirr errors are merely additions. Further the JAPICMP report confirms this.
>
> Existing classes that implement the interface won't have the newly
> added method. So if some other code calls that method in one of these
> implementations, they'll get an "AbstractMethodError" (or something
> similar).
>
> --
> Marcelo
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Pascal Schumacher
In reply to this post by Rob Tompkins
Am 28.01.2019 um 20:01 schrieb Rob Tompkins:
> Before I vote on the the thread, does adding a method to an interface cause BC to break?

"Adding a method to an interface does not break compatibility with
pre-existing binaries."

Source:
https://docs.oracle.com/javase/specs/jls/se7/html/jls-13.html#jls-13.5.1


---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Marcelo Vanzin-2
In reply to this post by Rob Tompkins
Haven't looked at the code, but if it's being compiled for java 8, and
the new methods have a default implementation, then it's fine. clirr
just complains because it's too old to know about default methods.

On Mon, Jan 28, 2019 at 11:18 AM Rob Tompkins <[hidden email]> wrote:

>
> @Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like this release should be a -1 then because we’re breaking BC without a major version change. Right??
>
> -Rob
>
>
> > On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin <[hidden email]> wrote:
> >
> > On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]> wrote:
> >> Before I vote on the the thread, does adding a method to an interface cause BC to break? I would think not. All of the clirr errors are merely additions. Further the JAPICMP report confirms this.
> >
> > Existing classes that implement the interface won't have the newly
> > added method. So if some other code calls that method in one of these
> > implementations, they'll get an "AbstractMethodError" (or something
> > similar).
> >
> > --
> > Marcelo
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>


--
Marcelo

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Rob Tompkins
Ok good enough. Then I’ll finish up looking at the signatures in a bit. Will be spotty today. I’m flying from Florida back to Virginia.

-Rob

> On Jan 28, 2019, at 2:22 PM, Marcelo Vanzin <[hidden email]> wrote:
>
> Haven't looked at the code, but if it's being compiled for java 8, and
> the new methods have a default implementation, then it's fine. clirr
> just complains because it's too old to know about default methods.
>
>> On Mon, Jan 28, 2019 at 11:18 AM Rob Tompkins <[hidden email]> wrote:
>>
>> @Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like this release should be a -1 then because we’re breaking BC without a major version change. Right??
>>
>> -Rob
>>
>>
>>> On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin <[hidden email]> wrote:
>>>
>>> On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]> wrote:
>>>> Before I vote on the the thread, does adding a method to an interface cause BC to break? I would think not. All of the clirr errors are merely additions. Further the JAPICMP report confirms this.
>>>
>>> Existing classes that implement the interface won't have the newly
>>> added method. So if some other code calls that method in one of these
>>> implementations, they'll get an "AbstractMethodError" (or something
>>> similar).
>>>
>>> --
>>> Marcelo
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>
>
> --
> Marcelo
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

sebb-2-2
In reply to this post by Marcelo Vanzin-2
On Mon, 28 Jan 2019 at 19:22, Marcelo Vanzin
<[hidden email]> wrote:
>
> Haven't looked at the code, but if it's being compiled for java 8, and
> the new methods have a default implementation, then it's fine. clirr
> just complains because it's too old to know about default methods.

I don't think so.

I think Clirr is complaining because it affects source compatibility,
and the Maven report does not distinguish source/binary complaints.

> On Mon, Jan 28, 2019 at 11:18 AM Rob Tompkins <[hidden email]> wrote:
> >
> > @Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like this release should be a -1 then because we’re breaking BC without a major version change. Right??
> >
> > -Rob
> >
> >
> > > On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin <[hidden email]> wrote:
> > >
> > > On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]> wrote:
> > >> Before I vote on the the thread, does adding a method to an interface cause BC to break? I would think not. All of the clirr errors are merely additions. Further the JAPICMP report confirms this.
> > >
> > > Existing classes that implement the interface won't have the newly
> > > added method. So if some other code calls that method in one of these
> > > implementations, they'll get an "AbstractMethodError" (or something
> > > similar).
> > >
> > > --
> > > Marcelo
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [hidden email]
> > > For additional commands, e-mail: [hidden email]
> > >
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
>
>
> --
> Marcelo
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Marcelo Vanzin-2
On Mon, Jan 28, 2019 at 11:36 AM sebb <[hidden email]> wrote:
> > Haven't looked at the code, but if it's being compiled for java 8, and
> > the new methods have a default implementation, then it's fine. clirr
> > just complains because it's too old to know about default methods.
>
> I don't think so.
>
> I think Clirr is complaining because it affects source compatibility,
> and the Maven report does not distinguish source/binary complaints.

Correct. If affects source compatibility if the new method doesn't
have a default implementation. With a default implementation, it does
not. But clirr does not know about that, since it was last updated
when Java 5 was released, IIRC...

--
Marcelo

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Rob Tompkins
In reply to this post by sebb-2-2


> On Jan 28, 2019, at 2:35 PM, sebb <[hidden email]> wrote:
>
> On Mon, 28 Jan 2019 at 19:22, Marcelo Vanzin
> <[hidden email]> wrote:
>>
>> Haven't looked at the code, but if it's being compiled for java 8, and
>> the new methods have a default implementation, then it's fine. clirr
>> just complains because it's too old to know about default methods.
>
> I don't think so.
>
> I think Clirr is complaining because it affects source compatibility,
> and the Maven report does not distinguish source/binary complaints.
>

By the way Japicmp says the same thing...that methods were added to interfaces.


>>> On Mon, Jan 28, 2019 at 11:18 AM Rob Tompkins <[hidden email]> wrote:
>>>
>>> @Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like this release should be a -1 then because we’re breaking BC without a major version change. Right??
>>>
>>> -Rob
>>>
>>>
>>>> On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin <[hidden email]> wrote:
>>>>
>>>> On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]> wrote:
>>>>> Before I vote on the the thread, does adding a method to an interface cause BC to break? I would think not. All of the clirr errors are merely additions. Further the JAPICMP report confirms this.
>>>>
>>>> Existing classes that implement the interface won't have the newly
>>>> added method. So if some other code calls that method in one of these
>>>> implementations, they'll get an "AbstractMethodError" (or something
>>>> similar).
>>>>
>>>> --
>>>> Marcelo
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>
>>
>> --
>> Marcelo
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [hidden email]
>> For additional commands, e-mail: [hidden email]
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Maxim Solodovnik
In reply to this post by Marcelo Vanzin-2
Hello All,

I have added japicmp report here:
https://dist.apache.org/repos/dist/dev/commons/collections/4.3-RC2/site/japicmp.html
I see no issues here :(

On Tue, 29 Jan 2019 at 03:03, Marcelo Vanzin <[hidden email]>
wrote:

> On Mon, Jan 28, 2019 at 11:36 AM sebb <[hidden email]> wrote:
> > > Haven't looked at the code, but if it's being compiled for java 8, and
> > > the new methods have a default implementation, then it's fine. clirr
> > > just complains because it's too old to know about default methods.
> >
> > I don't think so.
> >
> > I think Clirr is complaining because it affects source compatibility,
> > and the Maven report does not distinguish source/binary complaints.
>
> Correct. If affects source compatibility if the new method doesn't
> have a default implementation. With a default implementation, it does
> not. But clirr does not know about that, since it was last updated
> when Java 5 was released, IIRC...
>
> --
> Marcelo
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>

--
WBR
Maxim aka solomax
Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

garydgregory
In reply to this post by Pascal Schumacher
On Mon, Jan 28, 2019 at 2:21 PM Pascal Schumacher <[hidden email]>
wrote:

> Am 28.01.2019 um 20:01 schrieb Rob Tompkins:
> > Before I vote on the the thread, does adding a method to an interface
> cause BC to break?
>
> "Adding a method to an interface does not break compatibility with
> pre-existing binaries."
>

Of course it does break: I create a library L with interface A declaring
method M1, L calls M1 internally, and release it as 1.0. You implement A in
your app as ImplA.
I release 2.0 with A declaring M2 in A and also calling M1 and M2
internally. Your app updates to 2.0 and blows up when L tries to call M2 on
ImplA.

With Java 8, I can declare M2 as a _default_ method and nothing blows up,
you just get a default behavior that might be helpful or even be harmful.

Gary


>
> Source:
> https://docs.oracle.com/javase/specs/jls/se7/html/jls-13.html#jls-13.5.1
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>
Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

garydgregory
In reply to this post by Rob Tompkins
The bottom line is that new interface method in Collections 4.3 MUST be
default methods to avoid blowing up code. This is possible since Collection
now requires Java 8.

Gary

On Mon, Jan 28, 2019 at 2:53 PM Rob Tompkins <[hidden email]> wrote:

>
>
> > On Jan 28, 2019, at 2:35 PM, sebb <[hidden email]> wrote:
> >
> > On Mon, 28 Jan 2019 at 19:22, Marcelo Vanzin
> > <[hidden email]> wrote:
> >>
> >> Haven't looked at the code, but if it's being compiled for java 8, and
> >> the new methods have a default implementation, then it's fine. clirr
> >> just complains because it's too old to know about default methods.
> >
> > I don't think so.
> >
> > I think Clirr is complaining because it affects source compatibility,
> > and the Maven report does not distinguish source/binary complaints.
> >
>
> By the way Japicmp says the same thing...that methods were added to
> interfaces.
>
>
> >>> On Mon, Jan 28, 2019 at 11:18 AM Rob Tompkins <[hidden email]>
> wrote:
> >>>
> >>> @Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like
> this release should be a -1 then because we’re breaking BC without a major
> version change. Right??
> >>>
> >>> -Rob
> >>>
> >>>
> >>>> On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin
> <[hidden email]> wrote:
> >>>>
> >>>> On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]>
> wrote:
> >>>>> Before I vote on the the thread, does adding a method to an
> interface cause BC to break? I would think not. All of the clirr errors are
> merely additions. Further the JAPICMP report confirms this.
> >>>>
> >>>> Existing classes that implement the interface won't have the newly
> >>>> added method. So if some other code calls that method in one of these
> >>>> implementations, they'll get an "AbstractMethodError" (or something
> >>>> similar).
> >>>>
> >>>> --
> >>>> Marcelo
> >>>>
> >>>> ---------------------------------------------------------------------
> >>>> To unsubscribe, e-mail: [hidden email]
> >>>> For additional commands, e-mail: [hidden email]
> >>>>
> >>>
> >>>
> >>> ---------------------------------------------------------------------
> >>> To unsubscribe, e-mail: [hidden email]
> >>> For additional commands, e-mail: [hidden email]
> >>>
> >>
> >>
> >> --
> >> Marcelo
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail: [hidden email]
> >> For additional commands, e-mail: [hidden email]
> >>
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [hidden email]
> For additional commands, e-mail: [hidden email]
>
>
Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Maxim Solodovnik
Hello All,

It seems my previous post was missed somehow
so will re-post here:

I have added japicmp report here: [1] is has none issues

Additionally I have updated gist [2] it also show there are no issues

In case this is not enough I propose to cancel RC2 and to release 5.0

[1]
https://dist.apache.org/repos/dist/dev/commons/collections/4.3-RC2/site/japicmp.html
[2] https://gist.github.com/solomax/a6fbec6db71bb28dfe53afc566086505

On Tue, 29 Jan 2019 at 20:35, Gary Gregory <[hidden email]> wrote:

> The bottom line is that new interface method in Collections 4.3 MUST be
> default methods to avoid blowing up code. This is possible since Collection
> now requires Java 8.
>
> Gary
>
> On Mon, Jan 28, 2019 at 2:53 PM Rob Tompkins <[hidden email]> wrote:
>
> >
> >
> > > On Jan 28, 2019, at 2:35 PM, sebb <[hidden email]> wrote:
> > >
> > > On Mon, 28 Jan 2019 at 19:22, Marcelo Vanzin
> > > <[hidden email]> wrote:
> > >>
> > >> Haven't looked at the code, but if it's being compiled for java 8, and
> > >> the new methods have a default implementation, then it's fine. clirr
> > >> just complains because it's too old to know about default methods.
> > >
> > > I don't think so.
> > >
> > > I think Clirr is complaining because it affects source compatibility,
> > > and the Maven report does not distinguish source/binary complaints.
> > >
> >
> > By the way Japicmp says the same thing...that methods were added to
> > interfaces.
> >
> >
> > >>> On Mon, Jan 28, 2019 at 11:18 AM Rob Tompkins <[hidden email]>
> > wrote:
> > >>>
> > >>> @Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like
> > this release should be a -1 then because we’re breaking BC without a
> major
> > version change. Right??
> > >>>
> > >>> -Rob
> > >>>
> > >>>
> > >>>> On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin
> > <[hidden email]> wrote:
> > >>>>
> > >>>> On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]>
> > wrote:
> > >>>>> Before I vote on the the thread, does adding a method to an
> > interface cause BC to break? I would think not. All of the clirr errors
> are
> > merely additions. Further the JAPICMP report confirms this.
> > >>>>
> > >>>> Existing classes that implement the interface won't have the newly
> > >>>> added method. So if some other code calls that method in one of
> these
> > >>>> implementations, they'll get an "AbstractMethodError" (or something
> > >>>> similar).
> > >>>>
> > >>>> --
> > >>>> Marcelo
> > >>>>
> > >>>>
> ---------------------------------------------------------------------
> > >>>> To unsubscribe, e-mail: [hidden email]
> > >>>> For additional commands, e-mail: [hidden email]
> > >>>>
> > >>>
> > >>>
> > >>> ---------------------------------------------------------------------
> > >>> To unsubscribe, e-mail: [hidden email]
> > >>> For additional commands, e-mail: [hidden email]
> > >>>
> > >>
> > >>
> > >> --
> > >> Marcelo
> > >>
> > >> ---------------------------------------------------------------------
> > >> To unsubscribe, e-mail: [hidden email]
> > >> For additional commands, e-mail: [hidden email]
> > >>
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [hidden email]
> > > For additional commands, e-mail: [hidden email]
> > >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [hidden email]
> > For additional commands, e-mail: [hidden email]
> >
> >
>


--
WBR
Maxim aka solomax
Reply | Threaded
Open this post in threaded view
|

Re: JAPICMP Collections-4.3-RC2 report pdf (Was: Re: [VOTE][RC2] Commons collections 4.3)

Rob Tompkins


> On Jan 30, 2019, at 1:12 AM, Maxim Solodovnik <[hidden email]> wrote:
>
> Hello All,
>
> It seems my previous post was missed somehow
> so will re-post here:
>
> I have added japicmp report here: [1] is has none issues
>

What do folks think about [1]? It looks much more promising than the clirr report.

-Rob

> Additionally I have updated gist [2] it also show there are no issues
>
> In case this is not enough I propose to cancel RC2 and to release 5.0
>
> [1]
> https://dist.apache.org/repos/dist/dev/commons/collections/4.3-RC2/site/japicmp.html
> [2] https://gist.github.com/solomax/a6fbec6db71bb28dfe53afc566086505
>
>> On Tue, 29 Jan 2019 at 20:35, Gary Gregory <[hidden email]> wrote:
>>
>> The bottom line is that new interface method in Collections 4.3 MUST be
>> default methods to avoid blowing up code. This is possible since Collection
>> now requires Java 8.
>>
>> Gary
>>
>>> On Mon, Jan 28, 2019 at 2:53 PM Rob Tompkins <[hidden email]> wrote:
>>>
>>>
>>>
>>>> On Jan 28, 2019, at 2:35 PM, sebb <[hidden email]> wrote:
>>>>
>>>> On Mon, 28 Jan 2019 at 19:22, Marcelo Vanzin
>>>> <[hidden email]> wrote:
>>>>>
>>>>> Haven't looked at the code, but if it's being compiled for java 8, and
>>>>> the new methods have a default implementation, then it's fine. clirr
>>>>> just complains because it's too old to know about default methods.
>>>>
>>>> I don't think so.
>>>>
>>>> I think Clirr is complaining because it affects source compatibility,
>>>> and the Maven report does not distinguish source/binary complaints.
>>>>
>>>
>>> By the way Japicmp says the same thing...that methods were added to
>>> interfaces.
>>>
>>>
>>>>>> On Mon, Jan 28, 2019 at 11:18 AM Rob Tompkins <[hidden email]>
>>> wrote:
>>>>>>
>>>>>> @Marcelo - Many thanks...Yes. That makes sense. Thanks. Seems like
>>> this release should be a -1 then because we’re breaking BC without a
>> major
>>> version change. Right??
>>>>>>
>>>>>> -Rob
>>>>>>
>>>>>>
>>>>>>> On Jan 28, 2019, at 2:07 PM, Marcelo Vanzin
>>> <[hidden email]> wrote:
>>>>>>>
>>>>>>> On Mon, Jan 28, 2019 at 11:01 AM Rob Tompkins <[hidden email]>
>>> wrote:
>>>>>>>> Before I vote on the the thread, does adding a method to an
>>> interface cause BC to break? I would think not. All of the clirr errors
>> are
>>> merely additions. Further the JAPICMP report confirms this.
>>>>>>>
>>>>>>> Existing classes that implement the interface won't have the newly
>>>>>>> added method. So if some other code calls that method in one of
>> these
>>>>>>> implementations, they'll get an "AbstractMethodError" (or something
>>>>>>> similar).
>>>>>>>
>>>>>>> --
>>>>>>> Marcelo
>>>>>>>
>>>>>>>
>> ---------------------------------------------------------------------
>>>>>>> To unsubscribe, e-mail: [hidden email]
>>>>>>> For additional commands, e-mail: [hidden email]
>>>>>>>
>>>>>>
>>>>>>
>>>>>> ---------------------------------------------------------------------
>>>>>> To unsubscribe, e-mail: [hidden email]
>>>>>> For additional commands, e-mail: [hidden email]
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Marcelo
>>>>>
>>>>> ---------------------------------------------------------------------
>>>>> To unsubscribe, e-mail: [hidden email]
>>>>> For additional commands, e-mail: [hidden email]
>>>>>
>>>>
>>>> ---------------------------------------------------------------------
>>>> To unsubscribe, e-mail: [hidden email]
>>>> For additional commands, e-mail: [hidden email]
>>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: [hidden email]
>>> For additional commands, e-mail: [hidden email]
>>>
>>>
>>
>
>
> --
> WBR
> Maxim aka solomax

---------------------------------------------------------------------
To unsubscribe, e-mail: [hidden email]
For additional commands, e-mail: [hidden email]

1234