On Fri, 4 Jan 2019 at 11:55, Gilles <gil...@harfang.homelinux.org> wrote:
>
> hi.
>
> On Thu, 3 Jan 2019 22:04:24 +0530, Amey Jadiye wrote:
> > Hello Maxim / All
> >
> > I put little more efforts to find out the possible cause of the
> > error.
> > In the clirr code, I found the reason for this error and same is
> > documented
> > in clirr documentation [1].
> >
> > A method declaration has been added to the specified interface. This
> > is
> > always reported as a binary-compatibility error, but in practice, the
> > changed class might be used successfully with code compiled against
> > the old
> > interface depending upon usage patterns.
> >
> > Old code which invokes methods upon code compiled against the new
> > (expanded) interface will continue to work without issues. And old
> > code
> > which implements the old version of the interface will also continue
> > to
> > work correctly as long as no code attempts to invoke any of the
> > newly-added
> > methods against that instance. But the code which (validly) invokes
> > one of
> > the new methods in the interface against an object which implements
> > only
> > the old version of the interface will cause an AbstractMethodError to
> > be
> > thrown at the time the method invocation is attempted.
>
> IIUC, new code that calls the new methods on [Collections] classes
> will crash.  Does not look good.

If that is the case.
I'm not convinced - I think we need to test that assertion.

> On the other hand, Maxim wrote earlier that a method reported by
> Clirr as "new" (cf. below) was in fact present since Java 6...
>
> Are we then sure that it's a Clirr bug?
> If so, why did you not vote "+1" (on the condition that the false
> positive is mentioned in the release notes)?

Note also that Maven Clirr does not distinguish between source and
binary incompatibility.
(I think Clirr does, but this is not shown in the generated report).

> We could also make some definite progress with an actual code
> example that calls the incriminated methods, compiled against the
> current version of the library and then run against the current RC,
> and see whether it crashes.

+1

That would prove whether the assumptions are correct or not.

> Alternatively, we could instate "revapi" in the POM (and disable
> Clirr); and if the report is clean, trust that (though "revapi" is
> still beta).
>
> Opinions?
>
> Gilles
>
> >
> > In 4.2 and 4.3 looks like we have upgraded the java version[2] where
> > for
> > example for Map interface might have added a few more methods causing
> > these
> > errors.
> >
> > For this release, I am voting 0 (Non-Binding) as there is unharmed
> > mess
> > around the clirr.
> >
> > rest of the things are OK with this release, I would encourage to
> > have
> > revapi replacing clirr.
> >
> >
> > [1]  http://clirr.sourceforge.net/clirr-core/exegesis.html
> > [2]
> >
> > https://github.com/apache/commons-collections/commit/482762a13f739631f94d03642b0a55a9b7214c44
> >
> > Regards,
> > Amey
> >
> >
> > On Thu, Jan 3, 2019 at 11:53 AM Amey Jadiye <ameyjad...@gmail.com>
> > wrote:
> >
> >> I spared little time on finding issue however didn't found it in
> >> clirr(may
> >> be hidden somewhere),  today will check if clirr maven plugin have
> >> any
> >> issues. also I saw that few other apache commons modules having same
> >> issue
> >> and are released.
> >>
> >> I also gave try on revapi with commons collection4 4.3RC2.
> >> revapi:check
> >> was clean unlike clirr, looks promising to replace clirr.
> >>
> >> Regards,
> >> Amey
> >>
> >> On Wed, 2 Jan 2019, 12:31 pm Maxim Solodovnik <solomax...@gmail.com
> >> wrote:
> >>
> >>> Hello All,
> >>>
> >>> Just checked clirr report one more time
> >>> This time I took 1 error and perform investigation:
> >>>
> >>> "Method 'public java.util.Collection values()' has been added to an
> >>> interface" org.apache.commons.collections4.BidiMap
> >>>
> >>> In fact I don't understand why this error was reported
> >>> BidiMap extends java.util.Map
> >>> Map has method "public java.util.Collection values()" in all
> >>> versions:
> >>> https://docs.oracle.com/javase/6/docs/api/java/util/Map.html
> >>> https://docs.oracle.com/javase/7/docs/api/java/util/Map.html
> >>> https://docs.oracle.com/javase/8/docs/api/java/util/Map.html
> >>>
> >>> Maybe its clirr issue?
> >>> Would appreciate any help with this investigation
> >>>
> >>> On Mon, 31 Dec 2018 at 16:53, Gilles <gil...@harfang.homelinux.org>
> >>> wrote:
> >>> >
> >>> > On Mon, 31 Dec 2018 07:41:40 +0700, Maxim Solodovnik wrote:
> >>> > > Hello Gilles,
> >>> > >
> >>> > > I already did analysis: [1], all errors are caused by previous
> >>> > > release
> >>> > > 4.3 doesn't introduce any new errors ...
> >>> > >
> >>> > > [1] https://markmail.org/message/l7ftxlvdk4yqxijt
> >>> >
> >>> > I had seen the post but it says:
> >>> > ---
> >>> > these errors are weird. Above classes has no changes comparing to
> >>> 4.2
> >>> > ---
> >>> >
> >>> > But IMHO it was not a conclusion: If the cause of the errors was
> >>> > identified, it could have been mentioned in the release notes
> >>> > and/or the [VOTE] email, in order to avoid further questioning.
> >>> >
> >>> > Is the cause the change of supported JDK?
> >>> >
> >>> > Regards,
> >>> > Gilles
> >>> >
> >>> > >
> >>> > > On Mon, 31 Dec 2018 at 07:26, Rob Tompkins <chtom...@gmail.com>
> >>> > > wrote:
> >>> > >>
> >>> > >> I’ll give it a look tonight or in the morning.
> >>> > >>
> >>> > >> -Rob
> >>> > >>
> >>> > >> > On Dec 30, 2018, at 12:25 AM, Maxim Solodovnik
> >>> > >> <solomax...@gmail.com> wrote:
> >>> > >> >
> >>> > >> > No votes after 3 days :(
> >>> > >> > Is there anything wrong with the RC?
> >>> > >> >
> >>> > >> >> On Thu, 27 Dec 2018 at 05:20, Bruno P. Kinoshita
> >>> > >> <ki...@apache.org> wrote:
> >>> > >> >>
> >>> > >> >> FWIW I had a similar experience, and realized I was doing
> >>> `git
> >>> > >> fetch --all`, but it didn't bring the tags. `git fetch --tags`
> >>> did
> >>> the
> >>> > >> trick. After that I could `git checkout $tag-name`
> >>> > >> >>
> >>> > >> >> Cheers
> >>> > >> >> Bruno
> >>> > >> >>
> >>> > >> >>
> >>> > >> >>
> >>> > >> >>
> >>> > >> >> ________________________________
> >>> > >> >> From: Gilles <gil...@harfang.homelinux.org>
> >>> > >> >> To: dev@commons.apache.org
> >>> > >> >> Sent: Thursday, 27 December 2018 9:26 AM
> >>> > >> >> Subject: Re: [VOTE][RC2] Commons collections 4.3
> >>> > >> >>
> >>> > >> >>
> >>> > >> >>
> >>> > >> >>> On Wed, 26 Dec 2018 21:21:34 +0100, Gilles wrote:
> >>> > >> >>> Hi.
> >>> > >> >>>
> >>> > >> >>>> On Wed, 26 Dec 2018 20:41:59 +0700, Maxim Solodovnik
> >>> wrote:
> >>> > >> >>>> This is a [VOTE] for releasing
> >>> > >> >>>> Apache Commons collections 4.3
> >>> > >> >>>>
> >>> > >> >>>> Tag name:
> >>> > >> >>>>    collections-4.3-RC2 (signature can be checked from git
> >>> using
> >>> > >> >>>> 'git
> >>> > >> >>>> tag -v')
> >>> > >> >>>
> >>> > >> >>> $ git tag -v collections-4.3-RC2
> >>> > >> >>> error: tag 'collections-4.3-RC2' not found.
> >>> > >> >>>
> >>> > >> >>> Although I see it in the link below...
> >>> > >> >>> What is going on?
> >>> > >> >>
> >>> > >> >> Issue vanished with a fresh "clone".
> >>> > >> >> [Sorry for the noise.]
> >>> > >> >>
> >>> > >> >>
> >>> > >> >>>> RC1 was cancelled due to some release steps were not done
> >>> > >> >>>>
> >>> > >> >>>> Tag URL:
> >>> > >> >>>>
> >>> > >> >>>>
> >>> > >> >>>>
> >>> > >>
> >>>
> >>> https://gitbox.apache.org/repos/asf?p=commons-collections.git;a=commit;h=77e37dbf238d26351edb29e95391e3df75095d01
> >>> > >> >>>>
> >>> > >> >>>> Commit ID the tag points at:
> >>> > >> >>>>    77e37dbf238d26351edb29e95391e3df75095d01
> >>> > >> >>>>
> >>> > >> >>>> Site:
> >>> > >> >>>>
> >>> > >> >>>>
> >>> > >> >>>>
> >>> > >>
> >>>
> >>> https://dist.apache.org/repos/dist/dev/commons/collections/4.3-RC2/site/index.html
> >>> > >> >>>
> >>> > >> >>> The Clirr report is still a problem:
> >>> > >> >>>
> >>> > >> >>>
> >>> > >> >>>
> >>> > >>
> >>>
> >>> https://dist.apache.org/repos/dist/dev/commons/collections/4.3-RC2/site/clirr-report.html
> >>> > >> >>>
> >>> > >> >>> [The same errors are reported on my machine, so it's
> >>> > >> >>> not a cache issue...]
> >>> > >> >>>
> >>> > >> >>> Regards,
> >>> > >> >>> Gilles
> >>> > >> >>>
> >>> > >> >>>>
> >>> > >> >>>> Distribution files (committed at revision 31689):
> >>> > >> >>>>
> >>> https://dist.apache.org/repos/dist/dev/commons/collections/
> >>> > >> >>>>
> >>> > >> >>>> Distribution files hashes (SHA256):
> >>> > >> >>>>  commons-collections4-4.3-bin.tar.gz
> >>> > >> >>>>
> >>> > >>
> >>> 214c12fae27403f1a16ca6c108b5a8682be1a885a0dbbfc8eb30941303e1fe94
> >>> > >> >>>>  commons-collections4-4.3-bin.zip
> >>> > >> >>>>
> >>> > >>
> >>> 75b51a98fea6fca3746a3f70c6a0be24c99849e4976c4649214eaa5a009d0aeb
> >>> > >> >>>>  commons-collections4-4.3-src.tar.gz
> >>> > >> >>>>
> >>> > >>
> >>> 399f403feca86dbba7c4162eb90174db45979a2f7db2b3e0ba48240dc43ab434
> >>> > >> >>>>  commons-collections4-4.3-src.zip
> >>> > >> >>>>
> >>> > >>
> >>> 1c637e260b5b9e372d196593c7617ad3adedb6da3ac9196d086f9fc24401f5c3
> >>> > >> >>>>
> >>> > >> >>>> KEYS file to check signatures:
> >>> > >> >>>>   https://www.apache.org/dist/commons/KEYS
> >>> > >> >>>>
> >>> > >> >>>> Maven artifacts:
> >>> > >> >>>>
> >>> > >> >>>>
> >>> > >>
> >>>
> >>> https://repository.apache.org/content/repositories/orgapachecommons-1405/
> >>> > >> >>>>
> >>> > >> >>>> Please select one of the following options:
> >>> > >> >>>>  [ ] +1 Release it.
> >>> > >> >>>>  [ ] +0 Go ahead; I don't care.
> >>> > >> >>>>  [ ] -0 There are a few minor glitches: ...
> >>> > >> >>>>  [ ] -1 No, do not release it because ...
> >>> > >> >>>>
> >>> > >> >>>> This vote will be open for at least 72 hours, i.e. until
> >>> > >> >>>> 2018-12-29T14:00:00Z
> >>> > >> >>>> (this is UTC time).
> >>> > >> >>>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
> For additional commands, e-mail: dev-h...@commons.apache.org
>

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org
For additional commands, e-mail: dev-h...@commons.apache.org

Reply via email to