tomee-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Romain Manni-Bucau <>
Subject Re: openejb-core-4.6.1-20140513.040854-160.jar broken?
Date Thu, 15 May 2014 18:19:53 GMT
Yep, just wanted to make it less obvious. Internal/custom releases are
not always possible - one sample is open source releases of EE
products. That said it just means we should continue what you started
to do (thank you for that) ie release more often ;)

Romain Manni-Bucau
Twitter: @rmannibucau

2014-05-14 23:43 GMT+02:00 Andy Gumbrecht <>:
> @Romain, it works as well as the end user deems fit at the end of the day -
> There is such a thing as a stable snapshot, that is one that passes all the
> tests that has ever been written so far, and all the end user tests that
> they have for their own app. There is definitely no black and white rule to
> that, only an opinion. It is in fact what we did with openjpa-2.3.0-nonfinal
> and are doing with openjpa-2.4.0-1591689, taking a snapshot and
> making/declaring it stable/released for 'our' needs.
> I wouldn't include a snapshot reference in a production application, but I
> would create my own stable version off a snapshot if I am happy with it
> passing everything that it needs to pass and has a new feature that I need
> at the time. By doing that I would naturally be committed to keeping an eye
> on changes to the current snapshot that would affect the decision to use it
> in the first place - I am sure we will be keeping an eye on OpenJPA for the
> near future for example.
> To have and to keep a snapshot reference in a production app is asking for
> trouble, hence this user issue. To branch off the snapshot (we're talking
> trunk here) when it's 'stable/green' and test test test is much safer.
> To wait four or more months for a release is also just not workable for some
> people, as is the case for us sometimes - So there we are at the 'something
> in-between'.
> That said, I would also 'never' take or use a branch of trunk that has
> staging repos defined and try and make it stable, as that is also doomed to
> failure.
> Andy.
> On 14/05/2014 20:20, Romain Manni-Bucau wrote:
>> @Andy: it doesn't work neither. Users relying on snapshot needs a new
>> feature and it is important to stick on new snapshots to limit risk of
>> potential regressions we sometimes had cause we fixed only a part of
>> the real issue and by side effect the app was no more working. So I
>> think either you use releases (apache or custom) or snapshots but not
>> something in between
>> Romain Manni-Bucau
>> Twitter: @rmannibucau
>> Blog:
>> LinkedIn:
>> Github:
>> 2014-05-14 20:11 GMT+02:00 agumbrecht <>:
>>> Markus,
>>> You also have to understand that 'depending' on a snapshot is generally
>>> not
>>> a good idea. That doesn't mean snapshot builds are not a good idea,
>>> rather
>>> wait for a green build on the buildbot here:
>>>, then deploy your own
>>> version of it somewhere safe and use that until the next green build.
>>> Andy.
>>> -----
>>>      --
>>>      Andy Gumbrecht
>>>      TomEE treibt Tomitribe ! |
>>> --
>>> View this message in context:
>>> Sent from the OpenEJB User mailing list archive at
> --
>   Andy Gumbrecht
>   TomEE treibt Tomitribe! |

View raw message