[Compcomm] A possible forum merge proposal

[Mike Dransfield]

Danny Baumann wrote:
> Hi,
>    
>   
>>>>> Plugins are, libbs is being heavily worked on... It's not because you
>>>>> aren't doing anything in that sense that noone is working.
>>>>>
>>>>>   
>>>>>       
>>>>>           
>>>> libbs and bcop seem to have some useful features which David
>>>> is taking on board.  I do not think he is prepared to accept libbs
>>>> or bcop and he is implementing the features independently.
>>>>     
>>>>         
>>> Libbs and BCOP obviously won't be included into Compiz core - but they
>>> are prepared for usage with Compiz. From my understanding, that's what
>>> all this stuff is about.
>>>   
>>>       
>> As I said before you are only interested in the headline
>> features.
>>     
>
> As I said before this is not true. But headline features are those
> needing the most time, are those needing the most discussion and are the
> most obvious. Small bugs and such can be fixed when they arise and don't
> need a "merge plan".
>   

The problem is that they have already arisen, on the compiz-quinn
or beryl forums.  They will have to re-arise and the users are not
happy that things suddenly do not work again.

>>> This is supposed to be that way. The only remaining "big parts" are
>>> BTF/FTB drawing (for 3D and transparent cube), cube and rotate plugins
>>> (and perhaps some features from scale). You might have noticed that e.g.
>>> resize is close to be done.
>>>   
>>>       
>> Again, just the big parts
>>     
>
> Again, should I list every tiny bit?
>   

Yes please.  Thats what we are looking for, I thought that list already
exists?

>   
>>> What is the code which is not merged that you are talking about? I don't
>>> see any big things pending except the three I mentioned above. Code
>>> merge !== Compiz core changes, you know that as well as I do. Maybe the
>>> term "merge" isn't exactly the best wording, but I'm pretty sure you
>>> already got the idea.
>>>   
>>>       
>> There were > 1500 commits to the beryl svn and hundreds
>> in cvs, and an unknown amount on quinns harddrive.
>>
>> Who has gone through all of them and verified and submitted
>> them to the compiz ML?
>>
>> They will mostly be lost and they will relate to edge cases so
>> will be hard to detect.
>>     
>
> And all those commits were commits to either the forked core or the
> forked plugins? None of these was about new plugins or other added stuff
> (BSM, beryl-manager, ...)? Interesting.
>   

What was what, obviously not all of those 2000-odd patches
were relevant, but how many were?

1?  10?  100?


> If you really think there are lots of small fixes in beryl-core that
> aren't ported over because we are mean people or something like that,
> just take a copy of beryl-core and a copy of compiz core (of January)
> and compare. Easy as that.
> Hint: It happens that we already did exactly that.
>
>   
>>>> What about the actual reality of what this forum will be
>>>> supporting?  Will it be another compiz-quinn 'branch'?
>>>>     
>>>>         
>>> Hmm? What should be branched and why?
>>>   
>>>       
>> Who knows?  Dennis seems to disagree with Davids decision
>> to require FBO's so unless that can be resolved there will be
>> some sort of lesser-hardware branch.
>>     
>
> Ah, so you say that's for sure? Are you creating that branch? If not,
> you should stop pretending people have secret plans to branch anything
> just because there is some discussion on a ML - at the end, MLs were
> made for discussion.
>   

My point was that declaring the work done and then
sorting out the minor details is what got us into this mess
in the first place.  It does not help anyone to spend hours
working on something, only to be told it will have to be done
differently.

Why were the changes not mentioned on the list before work
was done?  This way some people are left upset because they
see their work as not acceptable.

> Regards,
>
> Danny
>
> _______________________________________________
> CompComm mailing list
> CompComm at Rock3d.org
> http://www.ubaight.com/mailman/listinfo/compcomm
>