Hi Chetan

IMO it also makes sense to properly manage export versions for packages
shared between different internal oak modules as we discussed it during
the Oakathon yesterday.

Maybe Marcel could share the corresponding notes...

Kind regards
Angela

On 15/11/17 17:30, "Chetan Mehrotra" <[email protected]> wrote:

>> And I definitely hope that we after that can stop making incompatible
>changes as that separations allows us to stop exporting things that are
>meant to be internal
>
>There are few packages which I think are exported mostly to allow
>other modules in Oak to work and as such are not like api to be used
>by end users (like spi.mount, spi.gc and spi.stats). Enforcing full
>backward compatibility for all spi package would hinder further
>evolution of design going forward.
>
>So would suggest to reconsider this and only version those which are
>meant to be used by users outside of Oak modules
>Chetan Mehrotra
>
>
>On Tue, Nov 14, 2017 at 11:00 PM, Angela Schreiber
><[email protected]> wrote:
>> Hi Chetan
>>
>> Well... I would have excepted that one goal of the m12n was to get a
>>clear
>> separation between public API and internals and everything that we
>>target
>> as API/SPI should be public IMO.
>>
>> And I definitely hope that we after that can stop making incompatible
>> changes as that separations allows us to stop exporting things that are
>> meant to be internal...
>>
>> Angela
>>
>> On 14/11/17 18:10, "Chetan Mehrotra" <[email protected]> wrote:
>>
>>>Do we want to have explicit version for all packages in oak-core-spi
>>>or should we only do it for packages which we expect code outside of
>>>Oak codebase would be using? As once we version it we cannot change in
>>>backward incompatible way easily
>>>Chetan Mehrotra
>>>
>>>
>>>On Tue, Nov 14, 2017 at 10:05 PM, Angela Schreiber
>>><[email protected]> wrote:
>>>> Hi Robert
>>>>
>>>> Ok... I will add 1.0.0 and go ahead tomorrow unless someone objects.
>>>>
>>>> Kind regards
>>>> Angela
>>>>
>>>> On 14/11/17 17:23, "Robert Munteanu" <[email protected]> wrote:
>>>>
>>>>>On Tue, 2017-11-14 at 13:51 +0000, Angela Schreiber wrote:
>>>>>> Any preference wrt the initial version number?
>>>>>
>>>>>The initial version number should be 1.0.0 IMO.
>>>>>
>>>>>Robert
>>>>
>>

Reply via email to