You are viewing a plain text version of this content. The canonical link for it is here.
Posted to oak-dev@jackrabbit.apache.org by Angela Schreiber <an...@adobe.com> on 2012/03/08 15:57:38 UTC
Integration Tests with TCK
hi
in addition to the proposal by jukka to setup benchmarks
i would like to suggest to soon setup integration testing
that also test jr3 against the TCK.
imo it's important that we identify implementation
details that violate the JCR specification early in the
process.
this will not only allow us to take conscious
decisions regarding compliance violations but will also
be very important to spot possible issues that will arise
with migration of JCR applications to jackrabbit3.
(thinking of apps like sling or cq that relying on quite
some jr2 implementation details that are also reflected
in the TCK :-).
what do you think?
angela
Re: Integration Tests with TCK
Posted by Felix Meschberger <fm...@adobe.com>.
ACK, thanks
Regards
Felix
Am 08.03.2012 um 08:17 schrieb Angela Schreiber:
> hi felix
>
>>> this will not only allow us to take conscious
>>> decisions regarding compliance violations but will also
>>> be very important to spot possible issues that will arise
>>> with migration of JCR applications to jackrabbit3.
>>> (thinking of apps like sling or cq that relying on quite
>>> some jr2 implementation details that are also reflected
>>> in the TCK :-).
>>
>> Good point. Of course, I would be very interested to hear about such implementation dependencies to be able to fix them in Sling.
>
> sure. you should not only hear it but we have to track and
> document such changes as detailed as possible... if ever possible
> i would like to collect it during the development process and come
> up with a comprehensive documentation once we encourage people to
> use jr3.
>
> at least we need to have a public checklist at release time that
> allows everyone to see what were the consequences of moving from jr2
> to jr3 and what changes would be required to run on jr3 (hopefully
> next to none) or take full advantage from improvements (hopefully lots).
>
> angela
>
>
Re: Integration Tests with TCK
Posted by Angela Schreiber <an...@adobe.com>.
hi felix
>> this will not only allow us to take conscious
>> decisions regarding compliance violations but will also
>> be very important to spot possible issues that will arise
>> with migration of JCR applications to jackrabbit3.
>> (thinking of apps like sling or cq that relying on quite
>> some jr2 implementation details that are also reflected
>> in the TCK :-).
>
> Good point. Of course, I would be very interested to hear about such implementation dependencies to be able to fix them in Sling.
sure. you should not only hear it but we have to track and
document such changes as detailed as possible... if ever possible
i would like to collect it during the development process and come
up with a comprehensive documentation once we encourage people to
use jr3.
at least we need to have a public checklist at release time that
allows everyone to see what were the consequences of moving from jr2
to jr3 and what changes would be required to run on jr3 (hopefully
next to none) or take full advantage from improvements (hopefully lots).
angela
Re: Integration Tests with TCK
Posted by Felix Meschberger <fm...@adobe.com>.
Hi,
Am 08.03.2012 um 06:57 schrieb Angela Schreiber:
> hi
>
> in addition to the proposal by jukka to setup benchmarks
> i would like to suggest to soon setup integration testing
> that also test jr3 against the TCK.
>
> imo it's important that we identify implementation
> details that violate the JCR specification early in the
> process.
>
> this will not only allow us to take conscious
> decisions regarding compliance violations but will also
> be very important to spot possible issues that will arise
> with migration of JCR applications to jackrabbit3.
> (thinking of apps like sling or cq that relying on quite
> some jr2 implementation details that are also reflected
> in the TCK :-).
Good point. Of course, I would be very interested to hear about such implementation dependencies to be able to fix them in Sling.
>
> what do you think?
Absolutey !
I even would go a step further: A new persistence abstraction - the Mikrokernel API - is defined. We might even create kind of TCK for this API such that multiple implementation (we expect to come up over time) can be tested against that common TCK and the rest of Oak using the Mikrokernel can depend on the Mikrokernel contracts.
Regards
Felix
>
> angela
Re: Integration Tests with TCK
Posted by Stefan Guggisberg <st...@gmail.com>.
On Thu, Mar 8, 2012 at 3:57 PM, Angela Schreiber <an...@adobe.com> wrote:
> hi
>
> in addition to the proposal by jukka to setup benchmarks
> i would like to suggest to soon setup integration testing
> that also test jr3 against the TCK.
>
> imo it's important that we identify implementation
> details that violate the JCR specification early in the
> process.
>
> this will not only allow us to take conscious
> decisions regarding compliance violations but will also
> be very important to spot possible issues that will arise
> with migration of JCR applications to jackrabbit3.
> (thinking of apps like sling or cq that relying on quite
> some jr2 implementation details that are also reflected
> in the TCK :-).
>
> what do you think?
+1, agreed.
cheers
stefan
>
> angela
Re: Integration Tests with TCK
Posted by Michael Dürig <md...@apache.org>.
On 8.3.12 15:57, Angela Schreiber wrote:
> hi
>
> in addition to the proposal by jukka to setup benchmarks
> i would like to suggest to soon setup integration testing
> that also test jr3 against the TCK.
>
> imo it's important that we identify implementation
> details that violate the JCR specification early in the
> process.
>
> this will not only allow us to take conscious
> decisions regarding compliance violations but will also
> be very important to spot possible issues that will arise
> with migration of JCR applications to jackrabbit3.
> (thinking of apps like sling or cq that relying on quite
> some jr2 implementation details that are also reflected
> in the TCK :-).
>
> what do you think?
Yes definitely. I created OAK-6 [1] for this. Depending on the
resolution of OAK-5 [2] TCK tests might be present already (i.e. for the
SPI based JCR implementation)
Michael
[1] https://issues.apache.org/jira/browse/OAK-6
[2] https://issues.apache.org/jira/browse/OAK-5
>
> angela