I just checked LinkMove with the fixed version, everything seems good to me.
I'll start a new voting thread.
On Tue, Apr 14, 2020 at 7:57 PM Andrus Adamchik wrote:
>
> > @Andrus: We'll let you test LM first this time?
>
> Be happy to.
>
>
> > On Apr 14, 2020, at 7:48 PM, Michael Gentry wrote:
> >
> > @John & @Nikita: If easy to fix, let's go ahead and do it just to have a
> > better M1.
> > @Andrus: We'll let you test LM first this time?
> >
> > I guess that makes me more of a -0 now. I'm still not opposed to releasing
> > as-is, but a better M1 is still desirable.
> >
> >
> > On Tue, Apr 14, 2020 at 5:10 AM Nikita Timofeev
> > wrote:
> >
> >> That's not a big effort to prepare a fresh release, it getting faster
> >> each time :).
> >> So I think I could do it one more time.
> >>
> >> As for the issue, the fix seems straightforward, so I'll just go for it.
> >>
> >> On Tue, Apr 14, 2020 at 4:30 AM John Huss wrote:
> >>>
> >>> I would be fine with fixing it up and remaking the release, but I don’t
> >>> know how much of a hassle that is. I can find time to fix AutoAdapter in
> >>> the next two days I think.
> >>>
> >>> On Mon, Apr 13, 2020 at 1:34 PM Andrus Adamchik
> >>> wrote:
> >>>
> Been thinking more about this... This breaks basic commit operations
> >> for
> both Derby and SQLServer, making it a major issue. We've never shipped
> releases with known bugs like this, even milestones. On the other hand
> there is a workaround - use an explicit adapter.
>
> I am voting -0, which means that I wouldn't release it like that, but I
> don't want to veto the release either. If the majority decides we
> >> should
> proceed, so be it.
>
> Andrus
>
>
> > On Apr 13, 2020, at 2:53 PM, Michael Gentry
> >> wrote:
> >
> > Hi Andrus,
> >
> > Given this is a milestone release, I think it is fine to release
> >> with the
> > AutoAdapter issue. These releases aren't required or expected to be
> > production-ready, although that would certainly be nice. They are
> >> for
> > finding issues needing to be addressed, and you have. :-)
> >
> > mrg
> >
> >
> > On Mon, Apr 13, 2020 at 5:24 AM Andrus Adamchik <
> >> and...@objectstyle.org>
> > wrote:
> >
> >> So looks like this will affect users of Derby and SQLServer who are
> >> also
> >> using the AutoAdapter (don't know if anyone still NOT using the
> >> AutoAdapter?). I suppose we can still release the M1, though it will
> not be
> >> usable for me unfortunately until this is fixed.
> >>
> >> Thoughts?
> >>
> >> Andrus
> >>
> >>> On Apr 13, 2020, at 11:40 AM, Nikita Timofeev <
> ntimof...@objectstyle.com>
> >> wrote:
> >>>
> >>> Hi Andrus,
> >>>
> >>> Yes, LinkMove problems related to batch generated keys. The
> >> problem is
> >>> with AutoAdapter, it simply doesn't define
> >>> supportsGeneratedKeysForBatchInserts() method.
> >>> DerbyAdapter itself works fine.
> >>>
> >>> On Sun, Apr 12, 2020 at 11:47 AM Andrus Adamchik <
> and...@objectstyle.org>
> >> wrote:
>
> Still testing the release... As expected 4.2 is no longer a
> >> drop-in
> >> replacement of 4.0 for either Agrest or LinkMove (because of the
> Property
> >> class refactoring). So I made a few tweaks to LM to make it compile
> >> [1].
> >> Now I am getting the following error on commit in a bunch of tests
> >> (e.g.
> >> CreateIT) [2].
>
> Is this related to 57332e865f5dabb9c7adef4bac9a61137f6828c4 (batch
> mode
> >> and PKs on Derby) ?
>
> (Derby version used in tests is 10.14.2.0 - the latest that
> >> supports
> >> Java 8)
>
> Andrus
>
>
> > On Apr 7, 2020, at 2:03 PM, Nikita Timofeev <
> ntimof...@objectstyle.com>
> >> wrote:
> >
> > Hi all,
> >
> > Here is another try for the Cayenne 4.2.M1 release.
> >
> > Release notes:
> >> https://github.com/apache/cayenne/blob/4.2.M1/RELEASE-NOTES.txt
> > Maven repo:
> >>
>
> >> https://repository.apache.org/content/repositories/orgapachecayenne-1038/
> > Assemblies:
> >> https://dist.apache.org/repos/dist/dev/cayenne/4.2.M1/
> >
> > Please evaluate and cast your votes.
> >
> > --
> > Best regards,
> > Nikita Timofeev
>
>
> [1] https://github.com/nhl/link-move/tree/cayenne-4.2 <
> >> https://github.com/nhl/link-move/tree/cayenne-4.2>
>
> [2] org.apache.cayenne.CayenneRuntimeException: [v.4.2.M1 Apr 07
> >> 2020
> >> 09:32:02] Mismatching number of generated PKs: expected 2, instead
> >> got 1
> at
> >>
>
> >> org.apache.cayenne.access.flush.FlushObserver.nextGeneratedRows(FlushObserver.java:77)
> at
>
