status on the build system & when to start?

Jaroslaw Staniek js at iidea.pl
Fri Sep 9 11:35:44 BST 2005


Stephan Kulow said the following, On 2005-09-09 11:10:

> Am Freitag, 9. September 2005 11:05 schrieb tnagyemail-ml at yahoo.fr:
> 
>>>what's the status of scons/bksys? I have
>>>questions coming from people 
>>>willing to start compiling kdelibs on win32, so
>>
>>Though several folders compile already: dcop, libltdl,
>>kdefx .., there are several unanswered questions about
>>the new framework:
>>* do we want one huge config.h or several config.h per
>>module ?
>>* are the api for building the programs satisfactory ?
>>* how to make the admin/ folder even more modular ?
>>* how intelligent the tools should be ? what kind of
>>helpers can be added to make things easier ?
>>
> * are people willing to compile on win32 help with the tools?

I am tend to advise against using autotools quite strongly for these folks. 
There are brilliant persons that won't mind using autotools and there are 
folks that want the building env. to work "out of the box", what I believe 
bksys can provide in a clean way, especially thanks to Thomas contributions. 
My point could be also that it's better to require fewer installations steps 
to get a working build system. I believe the same could be on the Mac side.

Related question is: understanding all the planned changes for kdelibs4, what 
could be a good time to start trying kdelibs4/win32 target? There are two 
opposable approaches, each having advantage:

(1) being up to date with the trunk, so the API can be improved were needed on 
the fly in terms of protability and purity; this requires a carefull tracking 
of changes at may side, -nothing new but requires much more more more time 
than (2); being prepared to non-compilable state, etc.

(2) waiting for first alpha stage or something like this; but then we can be 
forced to reconsider some decisions already made before that stage (while I 
was waiting)

Having that said, even while I am staying on the kdelibs3.5 bandwagon for 
stability reasons, looks like non-UNIX people willing to help are looking at 
the possibilities to go with KDE.
Back to bksys topic - yes, it's a first todo for me to acept it for win32 
target, especially that the task is relatively independent of current changes 
made to kdelibs4.

-- 
regards / pozdrawiam,
  Jaroslaw Staniek / OpenOffice Polska / Kexi Team
  http://www.openoffice.com.pl  |  http://www.kexi-project.org
  KDElibs/Windows: http://wiki.kde.org/tiki-index.php?page=KDElibs+for+win32




More information about the kde-core-devel mailing list