You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by Rob Tompkins <ch...@gmail.com> on 2018/05/01 21:05:23 UTC
[rng] japicmp failure
I was starting to fiddle with [parent] version 46 with [rng] and stumbled across the seemingly legitimate japicmp failure. @Gilles - any thoughts here? If it is indeed correct I would think those would be BC incompatible changes between 1.0 and 1.1.
-Rob
Comparing source compatibility of commons-rng-core-1.1-SNAPSHOT.jar against commons-rng-core-1.0.jar
===! UNCHANGED CLASS: PUBLIC ABSTRACT org.apache.commons.rng.core.source32.AbstractWell (not serializable)
=== CLASS FILE FORMAT VERSION: 50.0 <- 50.0
*** MODIFIED CLASS: PUBLIC ABSTRACT org.apache.commons.rng.core.source32.IntProvider (not serializable)
=== CLASS FILE FORMAT VERSION: 50.0 <- 50.0
--- REMOVED METHOD: PUBLIC(-) ABSTRACT(-) int next()
*** MODIFIED CLASS: PUBLIC ABSTRACT org.apache.commons.rng.core.source64.LongProvider (not serializable)
=== CLASS FILE FORMAT VERSION: 50.0 <- 50.0
--- REMOVED METHOD: PUBLIC(-) ABSTRACT(-) long next()
Re: [RNG] Release v1.1 without the examples? (Was: [rng] japicmp failure)
Posted by Gilles <gi...@harfang.homelinux.org>.
On Thu, 7 Jun 2018 11:47:33 -0400, Rob Tompkins wrote:
>> On Jun 7, 2018, at 11:26 AM, Gilles <gi...@harfang.homelinux.org>
>> wrote:
>>
>>> On Thu, 7 Jun 2018 10:49:28 -0400, Rob Tompkins wrote:
>>> I’m a +0 for not using japicmp because of its issues. We do need
>>> another [parent] release because of the japicmp skip availability
>>> in
>>> 0.12.0. I’d like to get the [release-plugin] up to snuff before
>>> going
>>> out though.
>>>
>>> Is [rng] ready to go for 1.1?
>>
>> It has been ready for more than 6 months!
>> [Even more when taking into account that the last work on it was
>> "just" to ensure that the JPMS modules were correct.]
>>
>>> If so, I can try to roll an RC for you.
>>
>> You are welcome, of course. But how will you get around the problem?
>
> Use the old version of commons-parent that works?
I'm afraid that old CP versions will turn up other problems (one
I recall would be "FindBugs" vs "SpotBugs") so that the cure would
be the same: release without the problematic module; and this is a
non-issue since it's mainly an illustration of the functionality
provided by the component, and some tools only useful during
development (for testing new RNG implementations).
Note also that only "mvn site" is affected. Would it be possible
to separate the "site" goal from the rest of the goals used by the
release plugin? Instead of
$ mvn -Prelease -Pcommons-rng-examples clean test site deploy
(which fails), we'd have
$ mvn -Prelease clean test site
$ mvn -Prelease -Pcommons-rng-examples deploy
The site would be missing info about the "examples" module, but
that's much of a problem than no release (and no site).
Regards,
Gilles
> -Rob
>>
>> Regards,
>> Gilles
>>
>>>
>>> -Rob
>>>
>>>> On Jun 7, 2018, at 10:29 AM, Gilles <gi...@harfang.homelinux.org>
>>>> wrote:
>>>>
>>>> Hello.
>>>>
>>>> The failure is triggered when including the "examples" module.
>>>> [Reminder: "japicmp" seems to complain about the absence of an
>>>> earlier version.]
>>>>
>>>> Thus, I'm contemplating to release v1.1 without the offending
>>>> module. Any objection to my preparing a RC on that basis?
>>>>
>>>> Regards,
>>>> Gilles
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [RNG] Release v1.1 without the examples? (Was: [rng] japicmp failure)
Posted by Rob Tompkins <ch...@gmail.com>.
> On Jun 7, 2018, at 11:26 AM, Gilles <gi...@harfang.homelinux.org> wrote:
>
>> On Thu, 7 Jun 2018 10:49:28 -0400, Rob Tompkins wrote:
>> I’m a +0 for not using japicmp because of its issues. We do need
>> another [parent] release because of the japicmp skip availability in
>> 0.12.0. I’d like to get the [release-plugin] up to snuff before going
>> out though.
>>
>> Is [rng] ready to go for 1.1?
>
> It has been ready for more than 6 months!
> [Even more when taking into account that the last work on it was
> "just" to ensure that the JPMS modules were correct.]
>
>> If so, I can try to roll an RC for you.
>
> You are welcome, of course. But how will you get around the problem?
Use the old version of commons-parent that works?
-Rob
>
> Regards,
> Gilles
>
>>
>> -Rob
>>
>>> On Jun 7, 2018, at 10:29 AM, Gilles <gi...@harfang.homelinux.org> wrote:
>>>
>>> Hello.
>>>
>>> The failure is triggered when including the "examples" module.
>>> [Reminder: "japicmp" seems to complain about the absence of an
>>> earlier version.]
>>>
>>> Thus, I'm contemplating to release v1.1 without the offending
>>> module. Any objection to my preparing a RC on that basis?
>>>
>>> Regards,
>>> Gilles
>>>
>>>
>>> ---------------------------------------------------------------------
>>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>>> For additional commands, e-mail: dev-help@commons.apache.org
>>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>> For additional commands, e-mail: dev-help@commons.apache.org
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [RNG] Release v1.1 without the examples? (Was: [rng] japicmp failure)
Posted by Gilles <gi...@harfang.homelinux.org>.
On Thu, 7 Jun 2018 10:49:28 -0400, Rob Tompkins wrote:
> I’m a +0 for not using japicmp because of its issues. We do need
> another [parent] release because of the japicmp skip availability in
> 0.12.0. I’d like to get the [release-plugin] up to snuff before going
> out though.
>
> Is [rng] ready to go for 1.1?
It has been ready for more than 6 months!
[Even more when taking into account that the last work on it was
"just" to ensure that the JPMS modules were correct.]
> If so, I can try to roll an RC for you.
You are welcome, of course. But how will you get around the problem?
Regards,
Gilles
>
> -Rob
>
>> On Jun 7, 2018, at 10:29 AM, Gilles <gi...@harfang.homelinux.org>
>> wrote:
>>
>> Hello.
>>
>> The failure is triggered when including the "examples" module.
>> [Reminder: "japicmp" seems to complain about the absence of an
>> earlier version.]
>>
>> Thus, I'm contemplating to release v1.1 without the offending
>> module. Any objection to my preparing a RC on that basis?
>>
>> Regards,
>> Gilles
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
>> For additional commands, e-mail: dev-help@commons.apache.org
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [RNG] Release v1.1 without the examples? (Was: [rng] japicmp failure)
Posted by Rob Tompkins <ch...@gmail.com>.
I’m a +0 for not using japicmp because of its issues. We do need another [parent] release because of the japicmp skip availability in 0.12.0. I’d like to get the [release-plugin] up to snuff before going out though.
Is [rng] ready to go for 1.1? If so, I can try to roll an RC for you.
-Rob
> On Jun 7, 2018, at 10:29 AM, Gilles <gi...@harfang.homelinux.org> wrote:
>
> Hello.
>
> The failure is triggered when including the "examples" module.
> [Reminder: "japicmp" seems to complain about the absence of an
> earlier version.]
>
> Thus, I'm contemplating to release v1.1 without the offending
> module. Any objection to my preparing a RC on that basis?
>
> Regards,
> Gilles
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
> For additional commands, e-mail: dev-help@commons.apache.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
[RNG] Release v1.1 without the examples? (Was: [rng] japicmp failure)
Posted by Gilles <gi...@harfang.homelinux.org>.
Hello.
The failure is triggered when including the "examples" module.
[Reminder: "japicmp" seems to complain about the absence of an
earlier version.]
Thus, I'm contemplating to release v1.1 without the offending
module. Any objection to my preparing a RC on that basis?
Regards,
Gilles
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [rng] japicmp failure
Posted by Gilles <gi...@harfang.homelinux.org>.
On Tue, 1 May 2018 17:05:23 -0400, Rob Tompkins wrote:
> I was starting to fiddle with [parent] version 46 with [rng] and
> stumbled across the seemingly legitimate japicmp failure. @Gilles -
> any thoughts here? If it is indeed correct I would think those would
> be BC incompatible changes between 1.0 and 1.1.
>
> -Rob
>
> Comparing source compatibility of commons-rng-core-1.1-SNAPSHOT.jar
> against commons-rng-core-1.0.jar
> ===! UNCHANGED CLASS: PUBLIC ABSTRACT
> org.apache.commons.rng.core.source32.AbstractWell (not serializable)
> === CLASS FILE FORMAT VERSION: 50.0 <- 50.0
What is meant by the above?
> *** MODIFIED CLASS: PUBLIC ABSTRACT
> org.apache.commons.rng.core.source32.IntProvider (not serializable)
> === CLASS FILE FORMAT VERSION: 50.0 <- 50.0
> --- REMOVED METHOD: PUBLIC(-) ABSTRACT(-) int next()
> *** MODIFIED CLASS: PUBLIC ABSTRACT
> org.apache.commons.rng.core.source64.LongProvider (not serializable)
> === CLASS FILE FORMAT VERSION: 50.0 <- 50.0
> --- REMOVED METHOD: PUBLIC(-) ABSTRACT(-) long next()
According to Clirr, those are not BC errors:
http://commons.apache.org/proper/commons-rng/commons-rng-core/clirr-report.html
Best,
Gilles
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org