[Building Sakai] 2.6.x branch is only building if Maven2 repo is deleted

[Anthony Whyte]

The compilation errors you are seeing are related to authz api/impl  
changes ( kernel trunk r 60869, kernel 1.0.x merge r 62993 on 28  
May).  So as David noted earlier Bamboo is not picking up K1 1.0.9- 
SNAPSHOT changes.

My suspicion is that the kernel 1.0.9-SNAPSHOT artifacts in the Sakai  
maven2-snapshot repo need refreshing in order to sync up with the  
associated 2.6.x changes.  But my dear Holmes, 1.0.9-SNAPSHOT was  
refreshed on 28 May so that cannot be the issue.  Ah but Watson look  
at the time stamps:

IU server timestamps:
1.0.9-SNAPSHOT maven2-snapshot refresh 6.55 am, 28 May
kernel 1.0.x commit r 62993 10:36 am, 26 May

See:

http://source.sakaiproject.org/maven2-snapshots/org/sakaiproject/kernel/sakai-kernel-api/1.0.9-SNAPSHOT/

https://source.sakaiproject.org/viewsvn/kernel/branches/kernel-1.0.x/api/src/?view=log

If my hunch is correct, a K1 1.0.x SNAPSHOT refresh will solve your  
immediate problem.  David can confirm.

IMO, if there were changes merged to 2.6.x that required K1 updates we  
should perform a K1 1.0.9 release to support it (with 2.6.0 continuing  
to bind to 1.0.8) rather than rely on a potentially fragile SNAPSHOT  
binding.

Cheers,

Anth


On Jun 3, 2009, at 7:42 PM, Kevin Chan wrote:

> I think the current Sakai 2.6.x /master/pom.xml file binds against  
> 1.0.9-SNAPSHOT:
>
> <sakai.kernel.version>1.0.9-SNAPSHOT</sakai.kernel.version>
>
> Looks like David Horwitz made this change at rev. 62996, which may  
> be related to work done in rev. 62997, both for the 2.6.x branch.
>
> Anthony,
>
> Can you explain how SNAPSHOT jar might cause the problem that I am  
> experiencing?
>
>  Kevin Chan
>
>  Operations Team
>  Educational Technology Services
>  University of California, Berkeley
>
>
> On 6/3/09 4:28 PM, Anthony Whyte wrote:
>> 2.6.x should be bound to K1 1.0.8 not 1.0.9-SNAPSHOT since our  
>> goal, as I understood it, is to eliminate otherwise unstable  
>> SNAPSHOT dependencies in the maintenance branch given that schools  
>> frequently choose to run 2.6.x in production.
>>
>> If a school wants to run 2.6.x and bind against the latest K1 1.0.x  
>> SNAPSHOT they are free to do so by changing the  
>> <sakai.kernel.version> binding in /master/pom.xml  to 1.0.9- 
>> SNAPSHOT but the default binding should be the latest stable  
>> release (e.g., 1.0.8).
>>
>> Anth
>>
>>
>>
>> On Jun 3, 2009, at 6:57 PM, Kevin Chan wrote:
>>
>>> Hi,
>>>
>>> So I have done more tests, but unfortunately, I have not been able  
>>> to
>>> fully fix this issue.
>>>
>>> As I stated before, our automated Bamboo build of Sakai 2.6.x has
>>> started to fail since May 27 (the date of the last good build),  
>>> which is
>>> also the date of the introduction of the use of 1.0.9-SNAPSHOT  
>>> kernel.
>>> Previously, David Horwitz suggested that I delete the
>>> org/sakaiproject/kernel local maven repo folder and rebuild and this
>>> appear to have fixed the problem. However, the next build (for  
>>> which I
>>> do not delete this directory) will inevitably fail.
>>>
>>> This is not happening to our automated Sakai trunk build nor our  
>>> local
>>> 2.6.x build (which is using a modified version of kernel and is  
>>> not up
>>> to date with the Sakai 2.6.x kernel module). I have never seen this
>>> behavior before, so I am hoping someone can:
>>>
>>> 1. Confirm this behavior?
>>> 2. Think of possible causes of this issue?
>>>
>>> Also, the nightly builds of Sakai 2.6.x are building fine, but I am
>>> guessing part of the scripted process is to delete the local maven
>>> ~/.m2/repository folder.
>>>
>>>  Kevin Chan
>>>
>>>  Operations Team
>>>  Educational Technology Services
>>>  University of California, Berkeley
>>>
>>>
>>>
>>> _______________________________________________
>>> sakai-dev mailing list
>>> sakai-dev at collab.sakaiproject.org
>>> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>>>
>>> TO UNSUBSCRIBE: send email to sakai-dev-unsubscribe at collab.sakaiproject.org 
>>>  with a subject of "unsubscribe"
>>>
>>>
>>
>
>