trinity-devel@lists.pearsoncomputing.net

Message: previous - next
Month: October 2012

Re: more discussion

From: Nick Leverton <nick@...>
Date: Tue, 16 Oct 2012 18:55:01 +0000 (UTC)
In article <201210160530.31317.slavek.banko@...>,
Slávek Banko  <trinity-devel@...> wrote:
>On Tuesday 16 of October 2012 00:17:11 Jekyll Wu wrote:
>> On 2012年10月16日 02:39, Darrell Anderson wrote:
>> > In practical terms, Trinity is a continuation of KDE 3.5.10
>>
>> I'm sorry, but I have to say you(Trinity) are not. KDE promises binary
>> compatibility and you(Trinity) breaks it. Is that called continuation ?
>>
>> Pretending and disclaiming you are a continuation of KDE3 would
>> eventually only bring more bad names to you(Trinity) and probably KDE.
>>
>> Face and acknowledge the truth: You are a fork, and everybody already
>> knows.
>>
>> Seriously, is that so hard and shameful?
>>
>> Regards
>> Jekyll
>>
>
>Are you sure it's broken binary compatibility? I use Twinkle compiled with KDE 
>support. Binary is compiled against KDE 3.5.x libraries from Squeeze and QT3 
>from Squeeze. And same unchanged binary working without any problems with the 
>TDE 3.5.x libraries and QT3 from Trinity.

I suspect KDE long term binary compatibility has as much to do with
the C++ ABI as the code's API.  During the lifetime of KDE 3, Debian
progressed from kdelibs4, to (I think) kdelibs4c, to kdelibs4c2, to
kdelibs4c2a, all due to g++ ABI changes.  That's leaving aside any Qt
soname changes, which I don't remember so well and can't be fussed to
dig out of the archives.

Nick
-- 
"The Internet, a sort of ersatz counterfeit of real life"
	-- Janet Street-Porter, BBC2, 19th March 1996