Re: The rebuilder shutdown bug: I don't think so; Spring handles the
shutdown of the ExecutorService.

Re: Constructors: Your approach is totally reasonable, though the
Factory is more my intent.  I'm honestly not sure that integration
test is doing the right thing instantiating its own RIOTripleIterator,
though: I don't see any other place in the codebase that happens.  I
guess that's testing client code?

- Ben

On Mon, Dec 12, 2011 at 3:27 AM, Stephen Bayliss
<stephen.bayl...@acuityunlimited.net> wrote:
> Hi Ben
>
> Thanks for the clarification on this. It sounds like we are not too far off
> tagging 1.5.4 of trippi from fix_executor; and moving on to 1.5.5 for
> Mulgara 2.1.11.
>
> What I'm not sure about is the rebuilder shutdown bug -
> https://jira.duraspace.org/browse/FCREPO-946 - related to
> https://jira.duraspace.org/browse/FCREPO-873.  Is there still something that
> needs implementing to deal with this for Trippi 1.5.4 (ie in the
> fix_executor branch)?
>
> I don't believe there is a default constructor.  I wasn't perhaps entirely
> clear - the constructor change - addition of the ExecutorService -- was
> affecting compilation of the Fedora integration tests; I modified these at
> https://github.com/fcrepo/fcrepo/commit/41b26a3c6ff26da6738e78481c57de1cf30292b7#diff-1 -
> maybe this code should be using
> TripleIteratorFactory.defaultInstance().fromStream(...) instead?
>
> Steve
>
>
> -----Original Message-----
> From: Benjamin Armintor [mailto:armin...@gmail.com]
> Sent: 11 December 2011 20:37
> To: fedora-commons-developers@lists.sourceforge.net
> Subject: Re: [fcrepo-dev] Status of trippi src?
>
> Steve,
> This is a blunder on my part... that branch should be ready to merge (it was
> the change that was meant to be tagged 1.4). If it doesn't have that default
> constructor it should, provided that an appropriate shutdown behavior can be
> implemented. Otherwise we'll reinstate the bug causing a one minute hang
> before shutdown (eg on rebuild ).
>
> I think there's an argument to be made that there shouldn't be a default
> constructor for that reason... maybe we should discuss on Tuesday.
>
> - Ben
>
> On Dec 11, 2011 12:08 PM, "Stephen Bayliss"
> <stephen.bayl...@acuityunlimited.net> wrote:
>>
>> Trippi master doesn't seem to build and pass all of its tests currently;
>> but I can see there's a branch fix_executor with the latest commits which
>> does build and pass tests.
>>
>> I'm working on FCREPO-702 (update to latest Mulgara version), so it would
>> be good to have a "known good" codebase to work from.
>>
>> Currently I've created branch fcrepo-702 from branch fix_executor, updated
>> to Mulgara 2.1.11 plus dependencies, and this is passing all tests.
>>
>> I've also got a Fedora branch fcrepo-702 with this integrated.
>> So:
>>
>> 1) Is fix_executor fit (or close) to merge to master?  And are there any
>> other outstanding planned changes?  If not it would be good to release a new
>> version of trippi soon with changes from fcrepo-702 merged in also (once the
>> fcrepo tests are all passing).
>>
>> 2) As soon as we can get the new mulgara and trippi jars maven-hosted (I
>> may need some help with that! - though I have a revised pom for 2.1.11) I
>> can also merge fcrepo-702 from fcrepo into master once I've finished
>> testing.
>>
>> Particularly on (1) there is a constructor change on RIOTripleIterator,
>> the constructor now takes an executor service - I wonder if the old
>> constructor with some kind of default/singe thread executor should be
>> reinstated, or whether consumers should be providing their own (this has
>> only affected Fedora integration tests; but anyone else who is also using
>> trippi artefacts would also have to make that change - unless we bring back
>> the existing constructor).  Similarly there are a few other modified
>> constructors elsewhere that now take a TripleIteratorFactory; again I wonder
>> if we should bring back in the old constructors also - to avoid any client
>> software changes in a 3.6 release (although the changes are relatively
>> minor).  All thoughts welcomed on this!
>>
>> Thanks
>> Steve
>>
>>
>> ------------------------------------------------------------------------------
>> Learn Windows Azure Live!  Tuesday, Dec 13, 2011
>> Microsoft is holding a special Learn Windows Azure training event for
>> developers. It will provide a great way to learn Windows Azure and what it
>> provides. You can attend the event by watching it streamed LIVE online.
>> Learn more at http://p.sf.net/sfu/ms-windowsazure
>> _______________________________________________
>> Fedora-commons-developers mailing list
>> Fedora-commons-developers@lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/fedora-commons-developers
>>
>
> ------------------------------------------------------------------------------
> Learn Windows Azure Live!  Tuesday, Dec 13, 2011
> Microsoft is holding a special Learn Windows Azure training event for
> developers. It will provide a great way to learn Windows Azure and what it
> provides. You can attend the event by watching it streamed LIVE online.
> Learn more at http://p.sf.net/sfu/ms-windowsazure
> _______________________________________________
> Fedora-commons-developers mailing list
> Fedora-commons-developers@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/fedora-commons-developers
>

------------------------------------------------------------------------------
Learn Windows Azure Live!  Tuesday, Dec 13, 2011
Microsoft is holding a special Learn Windows Azure training event for 
developers. It will provide a great way to learn Windows Azure and what it 
provides. You can attend the event by watching it streamed LIVE online.  
Learn more at http://p.sf.net/sfu/ms-windowsazure
_______________________________________________
Fedora-commons-developers mailing list
Fedora-commons-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/fedora-commons-developers

Reply via email to