On 2014-12-03 12:03, Patrick Ohly wrote:
> On Tue, 2014-12-02 at 18:19 +0100, Karol Lewandowski wrote:
>> On 2014-11-27 08:09, Kévin THIERRY wrote:
>>>
>>> On 11/27/2014 12:18 AM, Carsten Haitzler (The Rasterman) wrote:
>>>> On Wed, 26 Nov 2014 15:21:07 +0100 "Dominig ar Foll (Intel OTC)"
>>>> <[email protected]> said:
>>>>
>>>>> Hello,
>>>>>
>>>>> as you know we are aligning as much as we can Tizen packages version
>>>>> with Yocto 1.7 packages version.
>>
>>>> nope. i didn't know. when/where was this discussed? anyway...
>>
>>> In this mailing list (dev):
>>> https://www.mail-archive.com/[email protected]/msg04451.html
>>
>> I would say that this was announced, discussed - not really.
>
> I think we all agree that further discussion is needed before this
> really can be rolled out, and I am saying that as someone who has been
> involved with evaluating the possibility.
>
> Publishing the new workflow proposal was an invitation to join that
> discussion and a step towards doing it in public.
>
>>> http://permalink.gmane.org/gmane.comp.handhelds.tizen.devel/4706
>>> http://comments.gmane.org/gmane.comp.handhelds.tizen.devel/4912
>>>>
>>>> does this mean we are dropping GBS then as we'll build with OE?
>>> Yes but OBS build will still be supported as an alternative. I'm not
>>> sure if GBS will still be supported but it can make sense since OBS is
>>> still supported...
>>
>> This is not a small thing and this is *not* what we were told when
>> Tizen-Yocto
>> convergence was announced. It was about "adding the capability of building
>> Tizen with Yocto tools" - not about *replacing* these.
>
> The project has morphed over time. It's not done yet either.
>
>> A month ago package updates have been announced - and the need for the
>> versions to be the same in both Tizen and Yocto. This itself is a pain
>> and has
>> some serious consequences - basically it blocks Tizen development as
>> freezes it in some arbitrary point in time (for us Tizen-only devs, that
>> is) .
>
> Let me add my perspective on this.
>
> First, the current alignment activity is meant to ease the migration
> from Tizen to Yocto meta data. It'll be less disruptive when only the
> packaging changes and not also the base upstream version of the code.
>
> Second, updating packages in Tizen has been overdue in many cases. As
> Chanho Park mentioned, some of them haven't been touched since Anas
> packaged them initially for Tizen 3.x. We can't have Tizen being stuck
> on source code that is no longer maintained upstream.
Upgrade is surely needed and I can only applaud and thank for this
coordinated effort.
> And last but not least, this alignment is *not* permanent. There will
> continue to be per-project git trees which will allow using different
> versions in Tizen than used in Yocto. But how often that is done is not
> just a technical question (which we can discuss here) but also a
> business decision (which needs to be decided elsewhere) because it
> introduces additional maintenance costs.
Is Yocto release schedule going to affect us in future? Who is going
to make these binding decisions? What's Tizen's benefit in all of
this?
Tizen.Org was supposed to place where we can do "next-get" platform
devlopment, distant product from product development schedules.
Bringing in non-transparent ("decided elsewhere") business decisions
doesn't seem like recipe for success for me.
Not that everything is 100% transparent right now, question is if
this is step in right direction.
>> Now, we are reading that our development tools has to change too - our
>> tools 'might' work but it would be better if we would switch to Yocto way.
>
> This clearly is a reason to be concerned. I can assure you, as someone
> who, as a software developer, has been at the receiving end of tool
> changes multiple times (Moblin -> MeeGo -> Tizen) I'll do my best to
> ensure that this will be handled well this time.
I have seen all these transitions, heck, I have been involved in design
and implementation of pre-Tizen - scratchbox2-based build environment
and while I see technical value - now I would like to hear business
rationale. Making it easier for OEMs/ISVs to apply Tizen is certainly
good business case, but I would not consider any solution that would
have *slighest* chance of not being 100% compatible with current one.
Please consider that Samsung is so far biggest OEM with a few products
based on Tizen already released (Gear, Camera) and other in development.
This is all already established and I really wonder how this all is
going to affect current users, developers and infrastructure setups.
Breaking compatibility - now - or in future - is no go for me.
Cheers,
--
Karol Lewandowski, Samsung R&D Institute Poland
_______________________________________________
Dev mailing list
[email protected]
https://lists.tizen.org/listinfo/dev
