On 20 August 2015 at 11:30, Christian Schneider <ch...@die-schneider.net> wrote:
> On 20.08.2015 11:36, Jeremy Hughes wrote:
>>>>
>>>> Sometimes no exported package changes but you still have new
>>>> functionality. So a increasing the minor version instead of the bugfix
>>>> version makes sense.
>>
>> I agree, but this has always been the case. The version reflects the
>> semantics of the package or bundle, not the API. So if new function is
>> added and no externals (interfaces etc) change, then the minor version
>> must still get bumped. This is so that a client of that package can
>> indicate its requirement for the new function.
>>
>> Also, version numbers don't get bumped unnecessarily. Say a bundle has
>> changes that would ordinarily mean a minor version change, then the
>> minor version is bumped and the micro version zero'd out. All the
>> other bundles being co-released (assuming they don't have significant
>> ie major version changes) will also have their minor version bumped
>> (and micro version reset).
>> I would object to a the minor version (say) being bumped when all that
>> has changed across the whole release is micro changes.
>>
>>
> I fully agree. If there are only micro changes then the bundle version
> should also only be increased at the micro level.
> What I wanted to say is that you sometimes have changes that do not reflect
> in the API but still can fail existing clients.
> If you know that you have such changes in the code then an increase in the
> micro version would not correctly reflect this.
>
> So basically what I am saying is that we should not blindly trust the output
> of the version plugin.

Good point.

>
> Christian
>
>
> --
> Christian Schneider
> http://www.liquid-reality.de
>
> Open Source Architect
> http://www.talend.com
>

Reply via email to