[Building Sakai] Cafe trunk checkout is broken

Wed Oct 14 10:44:55 PDT 2009

Ok with r67777 the cafe trunk checkout is working. I get 2 errors on
startup. Something is looking for dav & podcasts (I suspect something in
archive).  These are non fatal and I'll be looking at those next

D

David Horwitz wrote:
> Hi Aaron,
>
> Site-manage definatly depends on the cm-api, whether this is runtime
> or only compilation I'm checking. I don't need to tell you that there
> are many bad dependencies in Sakai :-) I'm hoping that this
> refactoring will a) alert tool developers to what their depending on
> b) if there are bad dependencies at least you can easily put a
> combination together with only the services and not the tool
> implementations too.
>
> Unfortunately its way past our capacity to fix every tool in the build :-)
>
> D
>
> Aaron Zeckoski wrote:
>> There used to be stuff in site and sections which depended on the CM
>> api. That may no longer be the case in trunk but if it still is it
>> would be a little weird for the kernel to fail if "edu-services" is
>> not there.
>> Someone doing the refactoring may want to check on this.
>> -AZ
>>
>>
>> On Wed, Oct 14, 2009 at 11:49 AM, Steve Swinsburg
>> <steve.swinsburg at gmail.com> wrote:
>>   
>>> The CM API is now part of the edu-services bundle. However these new bundles
>>> come down from the repo at deploy time just like the kernel does, so aren't
>>> included in a source checkout, so you couldn't just add it as a module in
>>> the pom and build it.
>>>
>>> I'm not sure how difficult it would be to pull down only a small part out of
>>> these binary bundles? Or could we just list the inner dependency (ie cm-api
>>> vs the entire edu-services) and maven will figure it out?
>>>
>>> cheers,
>>> Steve
>>>
>>>
>>>
>>> On 14/10/2009, at 9:12 PM, Aaron Zeckoski wrote:
>>>
>>>     
>>>> I don't follow. The kernel should already be the smallest it is
>>>> possible to be. If it is possible to have a "cafe" kernel then
>>>> something is wrong.
>>>>
>>>> I wouldn't think there is a need for that. Just adjust the profile so
>>>> it only builds and deploys the CM api and it should be fine right?
>>>> It's not like the CM impl is part of the kernel (I hope).
>>>>
>>>> -AZ
>>>>
>>>>
>>>> On Wed, Oct 14, 2009 at 11:07 AM, David Horwitz <david.horwitz at uct.ac.za>
>>>> wrote:
>>>>       
>>>>> Cool was thinking -
>>>>>
>>>>> We have a cafe profile in the kernel-deploy (and maybe others) That will
>>>>> deploy a couple of chery picked components from the bundles that are
>>>>> needed.
>>>>>
>>>>> What do you think?
>>>>>
>>>>> D
>>>>>
>>>>>
>>>>> Aaron Zeckoski wrote:
>>>>>
>>>>> Both sadly. It is one of those things I complained about when it went
>>>>> it but there was no time to really do much about it.
>>>>> However, ideally we should not force the impl on people who do not
>>>>> need any course management code.
>>>>> -AZ
>>>>>
>>>>> On Wed, Oct 14, 2009 at 10:56 AM, David Horwitz <david.horwitz at uct.ac.za>
>>>>> wrote:
>>>>>
>>>>>
>>>>> Is the CM-api a runtime or compile time dependency?
>>>>>
>>>>> D
>>>>>
>>>>> Aaron Zeckoski wrote:
>>>>>
>>>>> It looks like things are getting out of sync with all the shifting
>>>>> around that is going on but we should deprecate cafe-trunk if the
>>>>> profile is being maintained. That said, they are not currently the
>>>>> same (cafe should only have the CM api in it and not all of CM) so
>>>>> some more work remains to be done here before that can happen.
>>>>> -AZ
>>>>>
>>>>> On Wed, Oct 14, 2009 at 2:13 AM, Steve Swinsburg
>>>>> <steve.swinsburg at gmail.com> wrote:
>>>>>
>>>>>
>>>>> It looks like we are going to need to add core-deploy to the cafe POMs
>>>>> (ie in the cafe source and the build profile in trunk) and the
>>>>> cafe .externals otherwise it will be missing the archive-api
>>>>> altogether (which is part of common), and the cm-api (which is part of
>>>>> edu-services). This will cause failures on startup in a fresh Tomcat.
>>>>>
>>>>> Adding core-deploy should fix both.
>>>>>
>>>>> The cafe .externals also checks out it's own version of an archive
>>>>> module and pom. It can be removed from the cafe .externals. It's
>>>>> already out of the pom so it won't build, its just redundant.
>>>>>
>>>>> I'll fix the trunk cafe build profile.
>>>>>
>>>>> cheers,
>>>>> Steve
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> On 14/10/2009, at 9:41 AM, Steve Swinsburg wrote:
>>>>>
>>>>>
>>>>>
>>>>> Just a note that you can also run a cafe build from a full trunk
>>>>> source via the -Pcafe profile as Anthony says below. That profile
>>>>> has been in for about a week and works just like the cafe source
>>>>> does, if you don't want to do another source checkout. But the cafe
>>>>> source is great!
>>>>>
>>>>> cheers,
>>>>> Steve
>>>>>
>>>>>
>>>>> On 14/10/2009, at 3:16 AM, Anthony Whyte wrote:
>>>>>
>>>>>
>>>>>
>>>>> For trunk, do an svn update and then:
>>>>>
>>>>> mvn -Pcafe clean install -Dmaven.test.skip=true sakai:deploy
>>>>>
>>>>>
>>>>>
>>>>> Anth
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> On Oct 13, 2009, at 12:08 PM, Jim Eng wrote:
>>>>>
>>>>>
>>>>>
>>>>> It looks like revision 67589 has broken the Cafe build:
>>>>>
>>>>>    http://source.sakaiproject.org/viewsvn?view=rev&revision=67589
>>>>>
>>>>> Cafe trunk can no longer be checked out because archive-api
>>>>> (https://source.sakaiproject.org/svn/archive/trunk/archive-api
>>>>> ) no longer exists in trunk even though it is specified in the
>>>>> externals:
>>>>>
>>>>>    https://source.sakaiproject.org/svn/cafe/trunk/.externals
>>>>>
>>>>> It's possible some of the other projects have moved as well, but
>>>>> attempting to checkout Cafe trunk fails at that point.
>>>>>
>>>>> I'm just wondering whether Cafe is being abandoned, or if it's
>>>>> just a
>>>>> lag in maintenance to deal with recent upheaval in subversion?
>>>>>
>>>>> Thanks.
>>>>>
>>>>> Jim
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> 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"
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> 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"
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> 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"
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>         
>>>> --
>>>> Aaron Zeckoski (azeckoski (at) vt.edu)
>>>> Senior Research Engineer - CARET - University of Cambridge
>>>> https://twitter.com/azeckoski - http://www.linkedin.com/in/azeckoski
>>>> http://aaronz-sakai.blogspot.com/ - http://tinyurl.com/azprofile
>>>>       
>>>     
>>
>>
>>
>>   
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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"
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/sakai-dev/attachments/20091014/2d0731b8/attachment.html 