Skip to main content
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [emf-dev] Maintenance Release Target Milestones 
Christian,

Comments below.

Christian W. Damus wrote:

Hi, Ed,
That implies more bug administration: at the "end" of the maintenance branch (when is that?), we change all of the SRx milestone targets to indicate that the bug was actually targeted for the GA release? That doesn't make sense. 
Sounds like it could be automated though.
As I understand it, the goal for the EMF and EMFT projects is to promote all of their components to projecthood by Galileo, right? 
When the foundation's infrastructures supports it, yes.
Then, we will all have a rational line-up of milestones that we can maintain separately and according to the version numbers that are current for us. 
Yes more more shuffling...
So, how about, until then, we accept a small number of additional shared milestones so that those of us who care about being able to accurately reflect a bug's history can do that? The choices won't grow at all after Galileo.
I'm not thrilled with that approach. I suppose you'll argue that setting the version in which you've fixed it isn't good enough either... (The MDT targets are still a mess. It would be nice if it worked the same way.) 

Thanks,

Christian


On 26-Nov-08, at 2:41 AM, Ed Merks wrote:
One might argue that SR1 always applies to the current service stream. Who would ever fix bugs that target releases more than a year old? :-P Once that service stream is released, the target could be changed to the name of the release.
I certainly don't see value in filling up the target list with a rapidly growing number of choices. 


Nick Boldt wrote:
I'm down with the Ganymede x.y.1 idea, but I prefer Ganymede SR1 and SR2 to x.y.1 and x.y.2, since it better aligns with the train. True too, the SR2 build may include x.y.3, not .2, if an interim extra build was done (eg., GMF).
That said I'll wait for Ed to weigh in when he gets back from his travels abroad. 

Nick

Christian W. Damus wrote:

Hi, Eike,
I agree. I would be happiest if Bugzilla had separate "Target Version" and "Target Milestone" fields, but alas, it does not. Of course, in that case, a "Ganymede x.y.2" target would be a target version, and not a milestone, but that would be fine.
I thought that the problem was in the proliferation of version *numbers* in the target milestones, so that we had, for example, 2.4.1/1.2.1/1.0.1 all referring to the same maintenance release of different EMF components, where a single "Ganymede x.y.1" would suffice for the lot.
However, now I am stuck with a bunch of bugs that were targeted for maintenance releases and are now targeted to "Past," and I have no sensible alternative in the current list of milestones. I just want to get back to conveying the same information in my bugs as they did last week.
"Ganymede x.y.1" etc. is not a great solution, but it fits the current database schema and it would actually work. 

If anyone has a cleaner solution, please let me know!

Thanks,

Christian

On 24-Nov-08, at 2:04 PM, Eike Stepper wrote:


Christian,
I always thought a clean solution would include separate target versions and generic target milestones. But I believe that this would require a broader discussion, if possible at all... 

Cheers
/Eike

----
http://thegordian.blogspot.com


Christian W. Damus schrieb:

Hi, all,

I have a follow-up problem on the subject of target milestones.
I see that we have SR1 and SR2 milestones, now, which I suppose are meant for the targeting of bugs into the maintenance releases. However, this raises some questions: 

   * how do I indicate which release stream
     (Europa/Ganymede/Galileo) the maintenance fix is targeted at?
      It doesn't make sense to use the flags because they are for
     planning, and these fixes aren't planned.  (besides, only
     Galileo has a flag)
   * my components have maintenance releases that line up with the
     train SR1 and SR2, but also more releases in between.  For
     example, I had a 1.2.1 release before 1.2.2 which corresponded
     with SR2.  We could add, say, SR3 and SR4 milestones, but the
     "SR" terminology suggests a correspondence with the train
     timetable
Can we, perhaps, add generic milestones as follows, to solve both of these issues? 

   * Ganymede x.y.1
   * Ganymede x.y.2
   * Ganymede x.y.3
   * Galileo x.y.1
   * Galileo x.y.2
   * Galileo x.y.3
I'm not sure that the SR1 and SR2 milestones will be useful. Perhaps they could then be deleted. 

What do other EMF committers think of this?

Thanks,

Christian

--
Christian W. Damus
Senior Software Developer, Zeligsoft Inc.
Component Lead, Eclipse MDT OCL and EMF-QTV
E-mail: cdamus@xxxxxxxxxxxxx <mailto:cdamus@xxxxxxxxxxxxx>
------------------------------------------------------------------------ 

_______________________________________________
emf-dev mailing list
emf-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/emf-dev



_______________________________________________
emf-dev mailing list
emf-dev@xxxxxxxxxxx <mailto:emf-dev@xxxxxxxxxxx>
https://dev.eclipse.org/mailman/listinfo/emf-dev


--
Christian W. Damus
Senior Software Developer, Zeligsoft Inc.
Component Lead, Eclipse MDT OCL and EMF-QTV
E-mail: cdamus@xxxxxxxxxxxxx <mailto:cdamus@xxxxxxxxxxxxx>
------------------------------------------------------------------------ 

_______________________________________________
emf-dev mailing list
emf-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/emf-dev




_______________________________________________
emf-dev mailing list
emf-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/emf-dev


--
Christian W. Damus
Senior Software Developer, Zeligsoft Inc.
Component Lead, Eclipse MDT OCL and EMF-QTV
E-mail: cdamus@xxxxxxxxxxxxx



_______________________________________________
emf-dev mailing list
emf-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/emf-dev

Back to the top