Re: [syncope-dev] ci?

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Re: [syncope-dev] ci?

Geert van der Ploeg


2011/8/30 Francesco Chicchiriccò <[hidden email]>
Onl 30/08/2011 11:59, Geert wrote:
Hi all,

Is there a (public) continuous integration system in place for Syncope somewhere that I can check?
I have issues building the current HEAD version of Syncope, so I wonder whether this version is building correctly on the CI system.

Hi Geert,
Syncope is actually missing a proper CI system - googlecode does not offer this (yet?).

Any hint about where to host this? We actually have in mind to put Jenkins in place at Tirasa, eventually.

Anyway, which problem do you have?


r1135 builds successfully. (although it's very slow at some tests, see attached (a part of console output))
From r1137-1141 there are some errors in poms (re. commons-lang, oracle ojdbc)
From r1142 onwards some integration tests (UserTestITCase, RoleTestITCase) fail. See attached log.

Only the latter is of real interest of course...

With kind regards,

Geert




logging.txt (6K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: [syncope-dev] ci?

ilgrosso
Administrator
On 30/08/2011 13:41, Geert van der Ploeg wrote

2011/8/30 Francesco Chicchiriccò <[hidden email]>
Onl 30/08/2011 11:59, Geert wrote:
Hi all,

Is there a (public) continuous integration system in place for Syncope somewhere that I can check?
I have issues building the current HEAD version of Syncope, so I wonder whether this version is building correctly on the CI system.

Hi Geert,
Syncope is actually missing a proper CI system - googlecode does not offer this (yet?).

Any hint about where to host this? We actually have in mind to put Jenkins in place at Tirasa, eventually.

Anyway, which problem do you have?


r1135 builds successfully. (although it's very slow at some tests, see attached (a part of console output))
From r1137-1141 there are some errors in poms (re. commons-lang, oracle ojdbc)
From r1142 onwards some integration tests (UserTestITCase, RoleTestITCase) fail. See attached log.

Test case results depend especially on:
* core/src/test/resources/content.xml (check that the file is not modified against SVN)
* /var/tmp/bundles content - for safety, remove this directory before starting any test

I have just followed this procedure on a Debian GNU / Linux machine, with Maven 3.0.3

1. removed ~/.m2
2. svn co https://syncope.googlecode.com/svn/trunk syncope
3. cd syncope
4. mvn clean install

Anyway, Fabio did the same on a Mac OS X Lion machine and he hit some of the problems you have, depending on the test execution order, probably due to the latest upgrade on failsafe and surefire maven plugins.
He did some changes on the test data and everything is working there as well. AFAIK, he is waiting to complete some tasks before commiting.

For the sake of curiosity, can you tell me why are you building the sources instead of using the published snapshot maven artifacts?

Cheers.
-- 
Francesco Chicchiriccò

"Computer Science is no more about computers than astronomy
is about telescopes." (E. W. Dijkstra)
Reply | Threaded
Open this post in threaded view
|

[syncope-dev] Re: ci?

Geert van der Ploeg
> Test case results depend especially on:
> * core/src/test/resources/content.xml (check that the file is not
> modified against SVN)

check

> * /var/tmp/bundles content - for safety, remove this directory before
> starting any test

check. Although this is a bit brittle in my opinion. Why not use a
temporary directory in the working directory (./target)? /var/tmp/
bundles is used by default when running syncope (and the archetype)
within in a container, so using /var/tmp/ clashes when doing both mvn
builds and running it within a container.

>
> I have just followed this procedure on a Debian GNU / Linux machine,
> with Maven 3.0.3
>
> 1. removed ~/.m2
> 2. svn cohttps://syncope.googlecode.com/svn/trunksyncope
> 3. cd syncope
> 4. mvn clean install

Did this on Win7 (64bits), Maven 3.0.3.
Fails in the same way I described before.

> Anyway, Fabio did the same on a Mac OS X Lion machine and he hit some of
> the problems you have, depending on the test execution order, probably
> due to the latest upgrade on failsafe and surefire maven plugins.
> He did some changes on the test data and everything is working there as
> well. AFAIK, he is waiting to complete some tasks before commiting.

That sounds promising...

> For the sake of curiosity, can you tell me why are you building the
> sources instead of using the published snapshot maven artifacts?

Because sometimes I hit a weird bug (like that Postgres @Lob one, or
the one you fixed in 1138 (pgsql transactional)) that I can fix easily
locally, before I've had the time to raise an issue and supply a
patch.
And because running an application from source from within an IDE is
just easier to test modifications.

Regards,
Geert
Reply | Threaded
Open this post in threaded view
|

Re: [syncope-dev] Re: ci?

ilgrosso
Administrator
On 30/08/2011 14:30, Geert wrote:
>> Test case results depend especially on:
>> * core/src/test/resources/content.xml (check that the file is not modified against SVN)
> check
>
>> * /var/tmp/bundles content - for safety, remove this directory before starting any test
> check. Although this is a bit brittle in my opinion. Why not use a temporary directory in the working directory (./target)? /var/tmp/bundles is used by default when running syncope (and the archetype) within in a container, so using /var/tmp/ clashes when doing both mvn builds and running it within a container.

You are right: opened
http://code.google.com/p/syncope/issues/detail?id=149 for this.

>> I have just followed this procedure on a Debian GNU / Linux machine, with Maven 3.0.3
>>
>> 1. removed ~/.m2
>> 2. svn cohttps://syncope.googlecode.com/svn/trunksyncope
>> 3. cd syncope
>> 4. mvn clean install
> Did this on Win7 (64bits), Maven 3.0.3.
> Fails in the same way I described before.
>
>> Anyway, Fabio did the same on a Mac OS X Lion machine and he hit some of the problems you have, depending on the test execution order, probably due to the latest upgrade on failsafe and surefire maven plugins.
>> He did some changes on the test data and everything is working there as well. AFAIK, he is waiting to complete some tasks before commiting.
> That sounds promising...

Let's wait for Fabio to commit, then...

>> For the sake of curiosity, can you tell me why are you building the sources instead of using the published snapshot maven artifacts?
> Because sometimes I hit a weird bug (like that Postgres @Lob one, or the one you fixed in 1138 (pgsql transactional)) that I can fix easily locally, before I've had the time to raise an issue and supply a
> patch.
> And because running an application from source from within an IDE is just easier to test modifications.

Understood: are you also evaluating to contribute to project's source
code? ;-)

--
Francesco Chicchiriccò

"Computer Science is no more about computers than astronomy
is about telescopes." (E. W. Dijkstra)

Reply | Threaded
Open this post in threaded view
|

Re: [syncope-dev] Re: ci?

Geert van der Ploeg

For the sake of curiosity, can you tell me why are you building the sources instead of using the published snapshot maven artifacts?
Because sometimes I hit a weird bug (like that Postgres @Lob one, or the one you fixed in 1138 (pgsql transactional)) that I can fix easily locally, before I've had the time to raise an issue and supply a
patch.
And because running an application from source from within an IDE is just easier to test modifications.

Understood: are you also evaluating to contribute to project's source code? ;-)

Sure! It's just that you guys normally fix and commit bugs before I even conclude it's indeed a bug instead of my own, local fault :-)

Regards,
Geert