hi harald,

ok - that's why i said "i guess" - with v1.5.2 there was only one commit in
the data-module which >could< have introduced a regression (for
applications based on osgi).
however, i'm happy to hear that we don't have a new regression (hopefully
the mentioned part still works for applications based on osgi).

+1 for contributing additional tests!

regards,
gerhard



2015-12-09 22:06 GMT+01:00 Harald Wellmann <[email protected]>:

> Hi Gerhard,
>
> sorry for not getting back sooner...
>
> So far, I figured out that the regression I've seen was introduced between
> 1.5.0 and 1.5.1, and not in 1.5.2 - I tested with an application that was
> running on DS 1.5.0 previously.
>
> It seems I missed the end of the vote by a few hours, otherwise I would
> change my vote now, and I apologize for any premature confusion.
>
> By the way, the regressions I've been talking about, now and previously
> for some 1.4.x releases, are not related to Pax CDI or OSGi at all.
>
> They all relate to my current commercial closed source daytime projects
> which are based on WildFly 9. There's no OSGi inside, and nothing out of
> the ordinary, I would think. We just happen to be using DeltaSpike Data for
> our persistence layer, and we've been bitten by incompatible changes after
> upgrading DS a couple of times now.
>
> I'll try to isolate some of my issues and contribute new tests for those
> use cases that have been fragile for us, and there's a couple of ideas I'd
> like to discuss in separate threads.
>
> Best regards,
> Harald
>
>
> Am 08.12.2015 um 01:44 schrieb Gerhard Petracek:
>
>> hi harald,
>>
>> since we have seen just 3 commits in the data-module (for this release),
>> i guess it's due to DELTASPIKE-1009 which fixes a regression caused by
>> DELTASPIKE-594.
>> however, we couldn't test it with an osgi-based application.
>>
>> as mentioned in the past:
>> everybody is very welcome to help with the osgi part (esp. with tests) or
>> we have to drop the unofficial (and still partial) osgi-support.
>>
>> regards,
>> gerhard
>>
>
>

Reply via email to