Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Davide Giannella]

On 12/02/2016 15:00, Davide Giannella wrote:
> Team,
>
> the build is constantly failing on my local with (oak-upgrade)
> ...
Thank you everyone for the effort and collaboration. All the issues are
resolved and I will start producing the cut around 10.30am GMT.

Please refrain any commit around that time up to when you see the voting
email.

Cheers
Davide

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Manfred Baedke]

Hi Tomek,

Thx a lot. I'll patch the relevant branches now.

Best regards,
Manfred

On 2/15/2016 11:41 AM, Tomek Rekawek wrote:

Hello,

The already mentioned JCR-2633 puts jcr:mixinTypes property into 
NodePropBundle#getPropertyEntries(). As a result, the oak-upgrade code 
responsible for replacing mix:simpleVersionable with mix:versionable doesn’t 
work correctly (the results are replaced by the original properties). I 
explained this in OAK-4018 and attached a patch.

Best regards,
Tomek


On 15/02/16 11:17, "Julian Sedding"  wrote:


The test failures in the issue seem to suggest that this may be relates to
simple versionables. IIRC we recently added support for some broken JR2
constructs. Could they have been fixed in the last JR release? If that's
the case it may no longer be possible to populate the source repository for
the tests.

Just pure guesses, but I thought it might help.

Regards
Julian


On Monday, February 15, 2016, Davide Giannella  wrote:


On 12/02/2016 18:36, Manfred Baedke wrote:

Hi,

This is due to change 1721196 (associated with JCR-2633), which
changes the persistent data model. Probably the test has just to be
tweaked accordingly, I'll look into it during WE.

Thank you very much Manfred.

I've filed https://issues.apache.org/jira/browse/OAK-4018 to keep track
and block 1.3.16.

 From here, once it's fixed in JR we have potentially 2 options:

1) unlock 1.3.16 by downgrading to JR 2.11.3
2) release JR 2.12.1, upgrade to Oak, release 1.3.16. Which will bring
the oak relase around 4-5 days late.

I'm for two as it will give us more coverage around the inclusion of the
new stable JR release.

Thoughts?

Davide

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Davide Giannella]

On 15/02/2016 10:47, Julian Reschke wrote:
> ...in the meantime, I think we should change oak-trunk back to use
> 2.11.3, so we have a clean build until this new issue is understood
> and fixed.

Done! :)

Cheers
Davide

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Julian Reschke]

On 2016-02-15 11:47, Julian Reschke wrote:

...in the meantime, I think we should change oak-trunk back to use
2.11.3, so we have a clean build until this new issue is understood and
fixed.

Best regards, Julian


Related to that...: OAK 1.2 currently uses Jackrabbit 2.10.0, although 
2.10.1 has been released a long time ago.


Shouldn't Oak always use the latest stable release from the matching 
Jackrabbit branch?


Best regards, Julian

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Julian Reschke]

On 2016-02-15 10:31, Davide Giannella wrote:

On 12/02/2016 18:36, Manfred Baedke wrote:

Hi,

This is due to change 1721196 (associated with JCR-2633), which
changes the persistent data model. Probably the test has just to be
tweaked accordingly, I'll look into it during WE.

Thank you very much Manfred.

I've filed https://issues.apache.org/jira/browse/OAK-4018 to keep track
and block 1.3.16.


From here, once it's fixed in JR we have potentially 2 options:


1) unlock 1.3.16 by downgrading to JR 2.11.3
2) release JR 2.12.1, upgrade to Oak, release 1.3.16. Which will bring
the oak relase around 4-5 days late.

I'm for two as it will give us more coverage around the inclusion of the
new stable JR release.

Thoughts?

Davide


...in the meantime, I think we should change oak-trunk back to use 
2.11.3, so we have a clean build until this new issue is understood and 
fixed.


Best regards, Julian

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Tomek Rekawek]

Hello,

The already mentioned JCR-2633 puts jcr:mixinTypes property into 
NodePropBundle#getPropertyEntries(). As a result, the oak-upgrade code 
responsible for replacing mix:simpleVersionable with mix:versionable doesn’t 
work correctly (the results are replaced by the original properties). I 
explained this in OAK-4018 and attached a patch.

Best regards,
Tomek


On 15/02/16 11:17, "Julian Sedding"  wrote:

>The test failures in the issue seem to suggest that this may be relates to
>simple versionables. IIRC we recently added support for some broken JR2
>constructs. Could they have been fixed in the last JR release? If that's
>the case it may no longer be possible to populate the source repository for
>the tests.
>
>Just pure guesses, but I thought it might help.
>
>Regards
>Julian
>
>
>On Monday, February 15, 2016, Davide Giannella  wrote:
>
>> On 12/02/2016 18:36, Manfred Baedke wrote:
>> > Hi,
>> >
>> > This is due to change 1721196 (associated with JCR-2633), which
>> > changes the persistent data model. Probably the test has just to be
>> > tweaked accordingly, I'll look into it during WE.
>> Thank you very much Manfred.
>>
>> I've filed https://issues.apache.org/jira/browse/OAK-4018 to keep track
>> and block 1.3.16.
>>
>> From here, once it's fixed in JR we have potentially 2 options:
>>
>> 1) unlock 1.3.16 by downgrading to JR 2.11.3
>> 2) release JR 2.12.1, upgrade to Oak, release 1.3.16. Which will bring
>> the oak relase around 4-5 days late.
>>
>> I'm for two as it will give us more coverage around the inclusion of the
>> new stable JR release.
>>
>> Thoughts?
>>
>> Davide
>>
>>
>>

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Julian Sedding]

The test failures in the issue seem to suggest that this may be relates to
simple versionables. IIRC we recently added support for some broken JR2
constructs. Could they have been fixed in the last JR release? If that's
the case it may no longer be possible to populate the source repository for
the tests.

Just pure guesses, but I thought it might help.

Regards
Julian


On Monday, February 15, 2016, Davide Giannella  wrote:

> On 12/02/2016 18:36, Manfred Baedke wrote:
> > Hi,
> >
> > This is due to change 1721196 (associated with JCR-2633), which
> > changes the persistent data model. Probably the test has just to be
> > tweaked accordingly, I'll look into it during WE.
> Thank you very much Manfred.
>
> I've filed https://issues.apache.org/jira/browse/OAK-4018 to keep track
> and block 1.3.16.
>
> From here, once it's fixed in JR we have potentially 2 options:
>
> 1) unlock 1.3.16 by downgrading to JR 2.11.3
> 2) release JR 2.12.1, upgrade to Oak, release 1.3.16. Which will bring
> the oak relase around 4-5 days late.
>
> I'm for two as it will give us more coverage around the inclusion of the
> new stable JR release.
>
> Thoughts?
>
> Davide
>
>
>

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Davide Giannella]

On 12/02/2016 18:36, Manfred Baedke wrote:
> Hi,
>
> This is due to change 1721196 (associated with JCR-2633), which
> changes the persistent data model. Probably the test has just to be
> tweaked accordingly, I'll look into it during WE.
Thank you very much Manfred.

I've filed https://issues.apache.org/jira/browse/OAK-4018 to keep track
and block 1.3.16.

>From here, once it's fixed in JR we have potentially 2 options:

1) unlock 1.3.16 by downgrading to JR 2.11.3
2) release JR 2.12.1, upgrade to Oak, release 1.3.16. Which will bring
the oak relase around 4-5 days late.

I'm for two as it will give us more coverage around the inclusion of the
new stable JR release.

Thoughts?

Davide

Re: Oak 1.3.16 release plan

[Stefan Egli]

Hi Davide,

As mentioned on the list OAK-4006 is in discussion and in the works. So,
depending on the outcome it might require a small delay.

Cheers,
Stefan

On 11/02/16 11:45, "Davide Giannella"  wrote:

>Hello team,
>
>I'm planning to cut Oak 1.3.16 on Monday 15th February more or less 10am
>GMT.
>
>If there are any objections please let me know. Otherwise I will
>re-schedule any non-resolved issue for the next iteration.
>
>Thanks
>Davide
>
>

Re: Oak 1.3.16 release plan

[Davide Giannella]

On 12/02/2016 12:03, Stefan Egli wrote:
> Hi Davide,
>
> As mentioned on the list OAK-4006 is in discussion and in the works. So,
> depending on the outcome it might require a small delay.
>
Sure not a problem. Ping me directly when done so that we can short-cut
any delays.

Cheers
Davide

oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Davide Giannella]

Team,

the build is constantly failing on my local with (oak-upgrade)

Failed tests:  
referencedOlderThanOrphaned(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
No history found for /versionables/mix:simpleVersionable/old
 
copyAllVersions(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
Node /versionables/mix:simpleVersionable/old should have mix:versionable
mixin
 
onlyReferencedAfterDate(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
No history found for /versionables/mix:simpleVersionable/young
 
onlyReferenced(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest): No
history found for /versionables/mix:simpleVersionable/old
 
referencedSinceDate(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
Node /versionables/mix:simpleVersionable/young should have
mix:versionable mixin

Can anyone give a look as it will be a blocker for next Monday release?

Cheers
Davide

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Davide Giannella]

On 12/02/2016 16:39, Julian Reschke wrote:
> It seems this was caused by switching to Jackrabbit 2.12.0.
>
> Reverting http://svn.apache.org/viewvc?rev=1730031=rev might
> unblock the release, but of course we'll still need to find out what
> Jackrabbit change caused the regression, why that wasn't catched by
> Jackrabbit tests, fix that, and make a new Jackrabbit release...

ok. We'll need a stable JR into Oak for 1.4. Honestly I'd rather delay
the 1.3.16 a couple of days for a proper fix so that we'll have longer
coverage in oak with the JR 2.12 branch.

The bug came indeed out after my inclusion of JR 2.12 which highlights
how coupled JR and Oak are; and that with the current approach we need a
reliable continuous build of JR SNAPSHOT and Oak SNAPSHOT to early catch
regressions.

Davide

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Julian Reschke]

On 2016-02-12 16:00, Davide Giannella wrote:

Team,

the build is constantly failing on my local with (oak-upgrade)

Failed tests:
referencedOlderThanOrphaned(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
No history found for /versionables/mix:simpleVersionable/old

copyAllVersions(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
Node /versionables/mix:simpleVersionable/old should have mix:versionable
mixin

onlyReferencedAfterDate(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
No history found for /versionables/mix:simpleVersionable/young

onlyReferenced(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest): No
history found for /versionables/mix:simpleVersionable/old

referencedSinceDate(org.apache.jackrabbit.oak.upgrade.CopyVersionHistoryTest):
Node /versionables/mix:simpleVersionable/young should have
mix:versionable mixin

Can anyone give a look as it will be a blocker for next Monday release?

Cheers
Davide


It seems this was caused by switching to Jackrabbit 2.12.0.

Reverting http://svn.apache.org/viewvc?rev=1730031=rev might 
unblock the release, but of course we'll still need to find out what 
Jackrabbit change caused the regression, why that wasn't catched by 
Jackrabbit tests, fix that, and make a new Jackrabbit release...


Best regards, Julian

Re: oak-upgrade test failures (was Re: Oak 1.3.16 release plan)

[Manfred Baedke]

Hi,

This is due to change 1721196 (associated with JCR-2633), which changes 
the persistent data model. Probably the test has just to be tweaked 
accordingly, I'll look into it during WE.


Best regards,
Manfred

On 2/12/2016 5:50 PM, Davide Giannella wrote:

On 12/02/2016 16:39, Julian Reschke wrote:

It seems this was caused by switching to Jackrabbit 2.12.0.

Reverting http://svn.apache.org/viewvc?rev=1730031=rev might
unblock the release, but of course we'll still need to find out what
Jackrabbit change caused the regression, why that wasn't catched by
Jackrabbit tests, fix that, and make a new Jackrabbit release...

ok. We'll need a stable JR into Oak for 1.4. Honestly I'd rather delay
the 1.3.16 a couple of days for a proper fix so that we'll have longer
coverage in oak with the JR 2.12 branch.

The bug came indeed out after my inclusion of JR 2.12 which highlights
how coupled JR and Oak are; and that with the current approach we need a
reliable continuous build of JR SNAPSHOT and Oak SNAPSHOT to early catch
regressions.

Davide

Oak 1.3.16 release plan

[Davide Giannella]

Hello team,

I'm planning to cut Oak 1.3.16 on Monday 15th February more or less 10am
GMT.

If there are any objections please let me know. Otherwise I will
re-schedule any non-resolved issue for the next iteration.

Thanks
Davide