You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cayenne.apache.org by Andrus Adamchik <an...@objectstyle.org> on 2007/06/22 15:23:49 UTC

How about 3.0 M1?

What's the consensus on making 3.0 M1 release?

I think it is long overdue - we want to show the many things we've  
developed over the last 1+ year [1]. Also I just checked in the  
minimal EJBQL support so that users can play with it (that was my own  
minimal TODO). As before "M" (milestone) means an alpha quality  
release with unstable new features, "unstable" indicating that the  
new API can change over the course of the release.

If nobody objects to going forward with M1, I will switch in the  
release preparation mode, testing the code across different  
databases, writing the docs, etc.

Andrus


[1] http://cayenne.apache.org/doc/guide-to-30-features.html

Re: How about 3.0 M1?

Posted by Andrus Adamchik <an...@objectstyle.org>.
My next step would be updating the docs on Wiki and brining the  
updates to SVN.

Any volunteers to run the RAT tool [1] against the trunk and filter  
out bogus failures?

Thanks
Andrus

[1] http://cayenne.apache.org/release-guide.html


On Jul 1, 2007, at 5:03 PM, Andrus Adamchik wrote:

> I am done with cross-DB testing of the trunk code. I tested Cayenne  
> with the following databases, and the current test suite fully  
> passes against all of them:
>
> * Derby
> * FrontBase
> * HSQLDB
> * MySQL
> * OpenBase
> * PostgreSQL
> * Sybase
> * SQLServer
>
> I discovered a bunch of issues and fixed most of them. Those that  
> didn't have an immediate fix, where documented on Jira and the  
> corresponding test cases were temporarily removed from the test  
> suite, or bypassed using the AccessStackAdapter mechanism. Here is  
> the listing:
>
> CAY-811: all db's that support identity columns (MySQL, Derby,  
> SQLServer)
> CAY-820: OpenBase
> CAY-821: FrontBase
>
> Still can't test against Oracle. I will set up a local Oracle  
> instance someday, but likely after M1.
>
> A pleasant surprise was that Maven 2.0.6 caused no surprises so  
> far. Maybe Maven 2 finally stabilized?? So I'd encourage others to  
> try running the unit tests in their environment and report back.  
> I've been updating the following Developer Guide page with test  
> instructions that worked for me. I hope this will help (note that  
> connection.properties file is mostly compatible with previous  
> releases, except that the key names are required to match the POM  
> profile names):
>
>   http://cayenne.apache.org/running-unit-tests.html
>
> Andrus
>
>
>


Re: How about 3.0 M1?

Posted by Andrus Adamchik <an...@objectstyle.org>.
I am done with cross-DB testing of the trunk code. I tested Cayenne  
with the following databases, and the current test suite fully passes  
against all of them:

* Derby
* FrontBase
* HSQLDB
* MySQL
* OpenBase
* PostgreSQL
* Sybase
* SQLServer

I discovered a bunch of issues and fixed most of them. Those that  
didn't have an immediate fix, where documented on Jira and the  
corresponding test cases were temporarily removed from the test  
suite, or bypassed using the AccessStackAdapter mechanism. Here is  
the listing:

CAY-811: all db's that support identity columns (MySQL, Derby,  
SQLServer)
CAY-820: OpenBase
CAY-821: FrontBase

Still can't test against Oracle. I will set up a local Oracle  
instance someday, but likely after M1.

A pleasant surprise was that Maven 2.0.6 caused no surprises so far.  
Maybe Maven 2 finally stabilized?? So I'd encourage others to try  
running the unit tests in their environment and report back. I've  
been updating the following Developer Guide page with test  
instructions that worked for me. I hope this will help (note that  
connection.properties file is mostly compatible with previous  
releases, except that the key names are required to match the POM  
profile names):

   http://cayenne.apache.org/running-unit-tests.html

Andrus



Re: How about 3.0 M1?

Posted by Andrus Adamchik <an...@objectstyle.org>.
Testing progress update...

Since 3.0 has only been tested with HSQLDB till now, trying other  
databases uncovered a few issues. I fixed all that I found so far  
except for CAY-811 that affects MySQL (in fact this issue exists in  
1.2 and 2.0, we just never had a test case for it). I have 100% test  
success rate with HSQLDB and PostgreSQL, and the single failure above  
with MySQL. Will try other DB's shortly.

I also documented the Maven unit test procedure here:

    http://cayenne.apache.org/running-unit-tests.html

Feel free to try it out. Will add more DB profiles shortly.

Andrus


On Jun 22, 2007, at 4:23 PM, Andrus Adamchik wrote:

> What's the consensus on making 3.0 M1 release?
>
> I think it is long overdue - we want to show the many things we've  
> developed over the last 1+ year [1]. Also I just checked in the  
> minimal EJBQL support so that users can play with it (that was my  
> own minimal TODO). As before "M" (milestone) means an alpha quality  
> release with unstable new features, "unstable" indicating that the  
> new API can change over the course of the release.
>
> If nobody objects to going forward with M1, I will switch in the  
> release preparation mode, testing the code across different  
> databases, writing the docs, etc.
>
> Andrus
>
>
> [1] http://cayenne.apache.org/doc/guide-to-30-features.html
>


Re: ToMany Maps [Was: How about 3.0 M1?]

Posted by Robert Zeigler <ro...@puregumption.com>.
Nice. I'll have to check it out when I get a chance (early September?).

Robert

On Aug 22, 2007, at 8/222:47 AM , Andrus Adamchik wrote:

> BTW, I just checked in the code that supports Map relationships  
> (and Set relationships, although this is not that exciting) per  
> CAY-848:
>
>   https://issues.apache.org/cayenne/browse/CAY-848
>
> Mapping that via the Modeler and runtime execution works. There are  
> still a few loose ends, most notably updating the relationship map  
> after commit to remap the objects whose map key property has changed.
>
> Andrus
>
>
> On Jun 23, 2007, at 12:08 AM, Robert Zeigler wrote:
>
>> Does the "JPA-inspiried classic features" include expanded mapping  
>> of relationships? In particular, does it currently include the  
>> notion of mapping a to-many relationship as a map, instead of a  
>> list? :)
>>
>> Robert
>>
>> On Jun 22, 2007, at 6/224:01 PM , Andrus Adamchik wrote:
>>
>>>
>>> On Jun 22, 2007, at 6:08 PM, Michael Gentry wrote:
>>>
>>>> I'm not really opposed, but I haven't looked much at the JPA stuff.
>>>
>>> This release is definitely alpha. While one can do real  
>>> persistence with Cayenne JPA already, we do not claim any kind of  
>>> completeness. Still good to have new JPA-inspired classic  
>>> features out there (such as lifecycle callbacks), as we'll get  
>>> user feedback and better chance of finding the bugs early.
>>>
>>>> However, do we still have an unresolved issue with the password  
>>>> encoding?  I
>>>> tried to subscribe to the Apache Legal mailing list several  
>>>> times and never
>>>> got added, so I could never take the question over there.
>>>
>>> Hmm... not sure whom we ping to get this resolved... One thing to  
>>> check - have you used your apache.org email? This may speed up  
>>> subscription moderation.
>>>
>>>> PS. I'd advertise the new JPA/etc stuff as alpha-quality, but  
>>>> emphasis the
>>>> Cayenne Classic stuff is stable.
>>>
>>> Yep, that's the plan.
>>>
>>> Andrus
>>>
>>
>>
>


ToMany Maps [Was: How about 3.0 M1?]

Posted by Andrus Adamchik <an...@objectstyle.org>.
BTW, I just checked in the code that supports Map relationships (and  
Set relationships, although this is not that exciting) per CAY-848:

   https://issues.apache.org/cayenne/browse/CAY-848

Mapping that via the Modeler and runtime execution works. There are  
still a few loose ends, most notably updating the relationship map  
after commit to remap the objects whose map key property has changed.

Andrus


On Jun 23, 2007, at 12:08 AM, Robert Zeigler wrote:

> Does the "JPA-inspiried classic features" include expanded mapping  
> of relationships? In particular, does it currently include the  
> notion of mapping a to-many relationship as a map, instead of a  
> list? :)
>
> Robert
>
> On Jun 22, 2007, at 6/224:01 PM , Andrus Adamchik wrote:
>
>>
>> On Jun 22, 2007, at 6:08 PM, Michael Gentry wrote:
>>
>>> I'm not really opposed, but I haven't looked much at the JPA stuff.
>>
>> This release is definitely alpha. While one can do real  
>> persistence with Cayenne JPA already, we do not claim any kind of  
>> completeness. Still good to have new JPA-inspired classic features  
>> out there (such as lifecycle callbacks), as we'll get user  
>> feedback and better chance of finding the bugs early.
>>
>>> However, do we still have an unresolved issue with the password  
>>> encoding?  I
>>> tried to subscribe to the Apache Legal mailing list several times  
>>> and never
>>> got added, so I could never take the question over there.
>>
>> Hmm... not sure whom we ping to get this resolved... One thing to  
>> check - have you used your apache.org email? This may speed up  
>> subscription moderation.
>>
>>> PS. I'd advertise the new JPA/etc stuff as alpha-quality, but  
>>> emphasis the
>>> Cayenne Classic stuff is stable.
>>
>> Yep, that's the plan.
>>
>> Andrus
>>
>
>


Re: How about 3.0 M1?

Posted by Andrus Adamchik <an...@objectstyle.org>.
Not yet.

On Jun 23, 2007, at 12:08 AM, Robert Zeigler wrote:

> Does the "JPA-inspiried classic features" include expanded mapping  
> of relationships? In particular, does it currently include the  
> notion of mapping a to-many relationship as a map, instead of a  
> list? :)
>
> Robert
>
> On Jun 22, 2007, at 6/224:01 PM , Andrus Adamchik wrote:
>
>>
>> On Jun 22, 2007, at 6:08 PM, Michael Gentry wrote:
>>
>>> I'm not really opposed, but I haven't looked much at the JPA stuff.
>>
>> This release is definitely alpha. While one can do real  
>> persistence with Cayenne JPA already, we do not claim any kind of  
>> completeness. Still good to have new JPA-inspired classic features  
>> out there (such as lifecycle callbacks), as we'll get user  
>> feedback and better chance of finding the bugs early.
>>
>>> However, do we still have an unresolved issue with the password  
>>> encoding?  I
>>> tried to subscribe to the Apache Legal mailing list several times  
>>> and never
>>> got added, so I could never take the question over there.
>>
>> Hmm... not sure whom we ping to get this resolved... One thing to  
>> check - have you used your apache.org email? This may speed up  
>> subscription moderation.
>>
>>> PS. I'd advertise the new JPA/etc stuff as alpha-quality, but  
>>> emphasis the
>>> Cayenne Classic stuff is stable.
>>
>> Yep, that's the plan.
>>
>> Andrus
>>
>
>


Re: How about 3.0 M1?

Posted by Robert Zeigler <ro...@puregumption.com>.
Does the "JPA-inspiried classic features" include expanded mapping of  
relationships? In particular, does it currently include the notion of  
mapping a to-many relationship as a map, instead of a list? :)

Robert

On Jun 22, 2007, at 6/224:01 PM , Andrus Adamchik wrote:

>
> On Jun 22, 2007, at 6:08 PM, Michael Gentry wrote:
>
>> I'm not really opposed, but I haven't looked much at the JPA stuff.
>
> This release is definitely alpha. While one can do real persistence  
> with Cayenne JPA already, we do not claim any kind of completeness.  
> Still good to have new JPA-inspired classic features out there  
> (such as lifecycle callbacks), as we'll get user feedback and  
> better chance of finding the bugs early.
>
>> However, do we still have an unresolved issue with the password  
>> encoding?  I
>> tried to subscribe to the Apache Legal mailing list several times  
>> and never
>> got added, so I could never take the question over there.
>
> Hmm... not sure whom we ping to get this resolved... One thing to  
> check - have you used your apache.org email? This may speed up  
> subscription moderation.
>
>> PS. I'd advertise the new JPA/etc stuff as alpha-quality, but  
>> emphasis the
>> Cayenne Classic stuff is stable.
>
> Yep, that's the plan.
>
> Andrus
>


Re: How about 3.0 M1?

Posted by Andrus Adamchik <an...@objectstyle.org>.
On Jun 22, 2007, at 6:08 PM, Michael Gentry wrote:

> I'm not really opposed, but I haven't looked much at the JPA stuff.

This release is definitely alpha. While one can do real persistence  
with Cayenne JPA already, we do not claim any kind of completeness.  
Still good to have new JPA-inspired classic features out there (such  
as lifecycle callbacks), as we'll get user feedback and better chance  
of finding the bugs early.

> However, do we still have an unresolved issue with the password  
> encoding?  I
> tried to subscribe to the Apache Legal mailing list several times  
> and never
> got added, so I could never take the question over there.

Hmm... not sure whom we ping to get this resolved... One thing to  
check - have you used your apache.org email? This may speed up  
subscription moderation.

> PS. I'd advertise the new JPA/etc stuff as alpha-quality, but  
> emphasis the
> Cayenne Classic stuff is stable.

Yep, that's the plan.

Andrus


Re: How about 3.0 M1?

Posted by Michael Gentry <bl...@gmail.com>.
I'm not really opposed, but I haven't looked much at the JPA stuff.
However, do we still have an unresolved issue with the password encoding?  I
tried to subscribe to the Apache Legal mailing list several times and never
got added, so I could never take the question over there.

Thanks,

/dev/mrg

PS. I'd advertise the new JPA/etc stuff as alpha-quality, but emphasis the
Cayenne Classic stuff is stable.


On 6/22/07, Andrus Adamchik <an...@objectstyle.org> wrote:
>
> What's the consensus on making 3.0 M1 release?
>
> I think it is long overdue - we want to show the many things we've
> developed over the last 1+ year [1]. Also I just checked in the
> minimal EJBQL support so that users can play with it (that was my own
> minimal TODO). As before "M" (milestone) means an alpha quality
> release with unstable new features, "unstable" indicating that the
> new API can change over the course of the release.
>
> If nobody objects to going forward with M1, I will switch in the
> release preparation mode, testing the code across different
> databases, writing the docs, etc.
>
> Andrus
>
>
> [1] http://cayenne.apache.org/doc/guide-to-30-features.html
>

Re: How about 3.0 M1?

Posted by Tore Halset <ha...@pvv.ntnu.no>.
On Jun 22, 2007, at 15:23, Andrus Adamchik wrote:

> What's the consensus on making 3.0 M1 release?

It will be good to have a M1 release making it easier for people to  
start testing the new features.

  - Tore.




Re: How about 3.0 M1?

Posted by Aristedes Maniatis <ar...@maniatis.org>.
On 22/06/2007, at 11:23 PM, Andrus Adamchik wrote:

> What's the consensus on making 3.0 M1 release?

As some of you know we've been using 3.0 in production since late  
last year. We have no reliability problems at all and we've been  
giving it a good workout. Derby and mySQL only, but running it on 1.4  
and 1.5 on OSX, Windows, Linux and FreeBSD. We know that it works at  
a basic level on Solaris.

My concern has long been that full JPA compliance is such a big goal,  
that 3.0 might be some time off. So either we reduce the goals for  
3.0 or keep releasing snapshots people could use in production.  
Lachlan and I will try to make time to work on inheritance (now that  
the requirements are clear thanks to everyone on this list) once we  
get some other important work out of the way, but I'm thinking that  
that will take some time to fully implement and test.


Also, we need to be clearer about terminology:

http://cayenne.apache.org/2007/03/08/version-30-milestones-and- 
javadoc.html

That indicates that the snapshots already released are 'milestones'.  
Will it be clear that cayenne-client-3.0-M1.jar is superior to  
cayenne-client-3.0-20070227.124237-1.jar? Or should the naming just  
continue on as date stamps?

Ari Maniatis


-------------------------->
Aristedes Maniatis
phone +61 2 9660 9700
PGP fingerprint 08 57 20 4B 80 69 59 E2  A9 BF 2D 48 C2 20 0C C8



Re: How about 3.0 M1?

Posted by Andrus Adamchik <an...@objectstyle.org>.
Cool.

I am very much into "release often release early" philosophy, and I'd  
like to start following it with 3.0 on both counts.

BTW, I just fixed a few remaining EJBQL join bugs that I found, so  
now I will fully concentrate on cross-DB testing and the docs.

Andrus

On Jun 25, 2007, at 7:41 PM, Kevin Menard wrote:
> I'm for it.  I've been out of the 3.0 game for a little bit, but it
> seemed pretty stable the last time I tried it.  It'd be helpful to get
> other people using it, too.  Early adopters can provide plenty of JIRA
> goodness.
>
> -- 
> Kevin
>
>> -----Original Message-----
>> From: Andrus Adamchik [mailto:andrus@objectstyle.org]
>> Sent: Friday, June 22, 2007 9:24 AM
>> To: dev@cayenne.apache.org
>> Subject: How about 3.0 M1?
>>
>> What's the consensus on making 3.0 M1 release?
>>
>> I think it is long overdue - we want to show the many things
>> we've developed over the last 1+ year [1]. Also I just
>> checked in the minimal EJBQL support so that users can play
>> with it (that was my own minimal TODO). As before "M"
>> (milestone) means an alpha quality release with unstable new
>> features, "unstable" indicating that the new API can change
>> over the course of the release.
>>
>> If nobody objects to going forward with M1, I will switch in
>> the release preparation mode, testing the code across
>> different databases, writing the docs, etc.
>>
>> Andrus
>>
>>
>> [1] http://cayenne.apache.org/doc/guide-to-30-features.html
>>
>


RE: How about 3.0 M1?

Posted by Kevin Menard <km...@servprise.com>.
I'm for it.  I've been out of the 3.0 game for a little bit, but it
seemed pretty stable the last time I tried it.  It'd be helpful to get
other people using it, too.  Early adopters can provide plenty of JIRA
goodness.

-- 
Kevin 

> -----Original Message-----
> From: Andrus Adamchik [mailto:andrus@objectstyle.org] 
> Sent: Friday, June 22, 2007 9:24 AM
> To: dev@cayenne.apache.org
> Subject: How about 3.0 M1?
> 
> What's the consensus on making 3.0 M1 release?
> 
> I think it is long overdue - we want to show the many things 
> we've developed over the last 1+ year [1]. Also I just 
> checked in the minimal EJBQL support so that users can play 
> with it (that was my own minimal TODO). As before "M" 
> (milestone) means an alpha quality release with unstable new 
> features, "unstable" indicating that the new API can change 
> over the course of the release.
> 
> If nobody objects to going forward with M1, I will switch in 
> the release preparation mode, testing the code across 
> different databases, writing the docs, etc.
> 
> Andrus
> 
> 
> [1] http://cayenne.apache.org/doc/guide-to-30-features.html
>