EclipseLink/Development/DevMeetings/20080702

Minutes are preceeded by >

02 July 2008 Minutes
---------------------
Attendees:
 Krogh Peter;
 Twelves David;
 Doughan Blaise;
 O'Brien Michael;
 Yorke Gordon; 
 Ware Tom;
 Meswani Mitesh;
 Striffler Roger;
 Norman Michael;
 Smith Shaun;

Actions:
---------------------
>Action: David/Peter check out remaining 9 bugs for show-stopers
>Action: Peter/Tom: SVN branching structure - we may not be able to branch off <trunk>
>Action: Blaise: Make schema validation flag non-validate until we baseline off of Java6

== Agenda July 2 2008 ==

[http://www.timeanddate.com/worldclock/fixedtime.html?month=1&day=30&year=2008&hour=14&min=0&sec=0&p1=188 2PM EST, Wednesday July 2, 2008]

<b>Dial In Info: </b> 1-888-967-2253, Conf ID/Password: 486546/486546

=== Status updates ===
==== 1.0 Release ====
* Bug Updates:
** [http://wiki.eclipse.org/EclipseLink/Development/Bugs Bug Queries and other useful links]
** Open for 1.0: 9
>All ones we are aware of (8 assigned)
>We are a week away from releasing
>Action: David/Peter check out remaining 9 bugs for show-stopers

* Runtime Project Move
** Done
* Graduation/Release review
** Done: June 26th
** next step?

>Version upgrades will be easier after 1.0


==== 1.0.M10====
* Milestone release Date - June 25 2008 
** Milestone up - some regressions
** Infrastructure issues - related to Ganymede??
** Showstopper Mode
*** All 1.0 assigned bugs are accounted for
**** reviewing unassigned bugs today

==== 1.0.M11====
* Branching July 2nd
>1.01 will be off this branch
>Action: Peter/Tom: SVN branching structure - we may not be able to branch off <trunk>
>We are tagging on 1.0 (not branching) and moving on
>tags under the root should be OK
 

* Foundation and JPA Status
* OSGi - Tom
* Workbench
>All checked in

* MOXy, JAXB, SDO - Blaise D
>2 outstanding 1.0 bugs - one is not blocking
>Namespaces: investigated
>worked out right way to do java5 validation - but there is a well known bug on xsi:type values in xerces
>So the correct fix will fail for most users if validation is enabled - even if the validation is valid
>Oracle's xmlparserv2 parser works, but the jdk parser does not enable
>Result: Validation will be disabled

>Worst case: you think validation is on - but there is no report (false positives/negatives)
>Default: there is a flag today that use to be on, with the recent changes we have no way to turn the flag back on
>>Make this default non-validate

>Future: When we support 1.6 as a baseline we will re-add validation in XMLProjectReader
>Logging warnings? We cannot tell that validation is not on
>Peter: leave as-is, validate for some environments
>Blaise: Option: we could switch to new style validation, under 1.6 we get a true pass/fail
>Gordon: This is really an enhancement since we do not get exceptions in 1.5
>Mike N: Schedule cleanup for 1.0.1



* DBWS - Mike N
>DBWS is not in the build probably until 1.0.1, tree still looks the same

* Build System
** Test failures are not being reported.
>Fixing
>We also have some mirror issues on eclipse post-ganymede

** Spatial tests being run by default.
* Documentation Status
>Roger: Doing regression testing on OX - couple of bug fixes early this week

=== General Issues ===
* [[EclipseLink/Development/Release/1.0#1.0_Release_Criteria | Exit Criteria]] 
** 1.0M8  June 4, 2008  Final functionality freeze  
** 1.0RC1  June 18, 2008  Release Candidate 1  
** 1.0RC2  June 25, 2008  Release Candidate 2  
** 1.0 Release  July 9, 2008  Final 1.0 Release (moved from june 25th to be released after Ganymede)
** Schedule Dependencies
*** Eclipse reviews: graduation, release, and move to RT
*** Ganymede to leverage Equinox 3.4 (June 25)
* Pre-release tasks
** 3rd party works-with approval for Oracle JDBC - Doug
** COMPLETE [[EclipseLink/IPLog | IP Log]] - Peter
** Complete [[EclipseLink/Features | Features Index ]] - Doug with support of all committers
** Complete Examples & How-to materials - Doug/Shaun
** FAQ
** Update wiki Documentation for 1.0 Features - Liza (starting in May) with committer support
** Finalize product packaging and distribution - ZIP(s), Bundle access (update site?)
** Prepare and schedule reviews
** Discussion
*** Packaging convention for dependancies (i.e. the projects we depend on that are not from Orbit)
**** Do we want a src and bin/classes directory
**** Should classes just exist at the base of the project (what about the jar we eventually create)
**** How should we name the jars we create
* WIKI Changes should be done in parallel to any bug fixes to eclipselink.*.test ant targets or other public API changes - mobrien 20080529.
* DONE Bugzilla components - sign-off on final changes below
http://dev.eclipse.org/mhonarc/lists/eclipselink-dev/msg00640.html
* Automated Testing 
** Need SRG test models for JPA, MOXy and SDO
*** DB components of those tests should be simple enough to run on Derby
*** Need to test Core SRG on Derby
* Bug Prioritization and targetting
** Currently more bugs are scheduled for 1.0.1 then we are likely to fix in that time
** We currently do not have a system to ensure a bug has a published priority level
** We should discuss how will we deal with these issues
* Bug DB usage
** If you are working on a bug, assign it to yourself.
*** add dummy inbox to the cc list.
** Set the target milestone to the correct milestone.
*** Confusion/concern on this part
*** Milestones
**** 1.0M9 are bugs that are to be fixed by the the corresponding milestone build.
**** 1.0 is a bug that is to be fixed by 1.0, but has no current target milestone.
**** 1.0.1 new bugs should be filed against this target
* Bug council
** Introduce notion of Bug council
** Wed 2:30pm
*** Untargeted bugs are going to be reviewed by bug council on a weekly basis (current proposed time is directly after weekly status meeting (2:30 ET on Wednesdays).
**** A milestone will be set - or set to future
**** Bug may get assigned
*** When you enter a new bug, leaving it untargeted means that the bug council will review to decide which release it is targeted for.
*** If you know which milestone the bug will be fixed in, assign it accordingly.
*** The new target - future - is for bugs that don't have a target yet.
* Code headers
** Ensure Header is dual license EDL + EPL
*** Found and fxed some build xml files.
** Ensure contributor is correct - new code should have code author (no - intial code from oracle)
*** Found and fxed some build xml files.


=== Upcoming Issues ===

The following items require discussion but are not yet at the stage where a solution can be proposed.

* [[EclipseLink/Development/Examples]]: Strategy for the development and maintenance of examples
* [[EclipseLink/Development/OSGi]]: Strategy for delivery of OSGi/Equinox bundles 
* [[EclipseLink/Development/JPA_2.0]]: Strategy for the concurrent development of JPA 2.0 while maintaining JPA 1.0 usage
* [[EclipseLink/Development/Spring]]: Strategy for using EclipseLink within the Spring Framework
* [[EclipseLink/Development/Java6]]: Strategy for using EclipseLink on a Java SE 6 JRE

>Tom Q: What version of EclipseLink will V3 of glassfish use
>V3 release is JavaOne
>V3 code freeze dec/jan
>V3 lite - oct release
>pick up EclipseLink 1.0.1 unless 1.1 is ready

>Mitesh Q) JPA2?
>Nearly finished
>Mitech Q) Testing/CTS
>a couple weeks away from this schedule


>Bug council triage

OUT 217508: work not ready yet for branch
>move to 1.0.1

IN  219683: low impact - risk-wise
>in for 1.0

OUT 232564: weblogic test failure on rollback
>defer to 1.0.1

OUT 234895: DOC
OUT 234914: skip doc bug in this triage as they are unblocking 1.0 bugs

OUT 236500: SDO mixed text
>end of day friday
>no tck for 2.1 so now compliance impact if it is out
>we can patch 1.0, defer to 1.0.1

IN? 237545: [ORDER BY] in lists being reviewed
>Impact is risky
>check in for tomorrow 

OUT 237853: CONCAT into || instead of +
>Andrei may be checking in, if not then defer to 1.0.1

Regression in weaving in CLONE - set private to public
>Tom: test now passes on build machine
>bug# is missing