[Bf-committers] Modifiers design

[Campbell Barton]

change the order of linking r28406. so modifiers can use anything
defined in blenderkernel.

On Tue, Apr 20, 2010 at 4:34 PM, Campbell Barton <ideasman42 at gmail.com> wrote:
> correction, modifiers can call BKE functions, however I was getting
> linking errors with some BKE functions, I moved these to files,
>  source/blender/modifiers/intern/MOD_boolean_util.c
>  source/blender/modifiers/intern/MOD_fluidsim_util.c
>
> When linking the libraries in the reverse order the exten's from
> MOD_modifiertypes.h couldnt be found by BKE
>
> On Tue, Apr 20, 2010 at 4:00 PM, Campbell Barton <ideasman42 at gmail.com> wrote:
>> This week I'll look into having modifiers be able to call BKE func's
>> again, there was a circular dependency between the modifiers and
>> blenderkernel but there are a few ways this can be resolved.
>>
>> On Tue, Apr 20, 2010 at 11:42 AM, Brecht Van Lommel <brecht at blender.org> wrote:
>>> Hi,
>>>
>>> This is not good software architecture, but it would need much deeper
>>> changes to fix. Blenlib is intended to be a library with more generic
>>> data structures, like lists, hashes, not something specific like
>>> DerivedMesh. And you can't pull DerivedMesh out of blenkernel without
>>> creating a two way dependency between these then. It may be good to
>>> split up blenkernel some day but this is better done if there is a
>>> bigger design behind it, for now you can assume the modifiers to be
>>> part of blenkernel, they were just moved into a separate directory for
>>> convenience.
>>>
>>> Brecht.
>>>
>>> On Tue, Apr 20, 2010 at 10:30 AM, j.bakker at atmind.nl <j.bakker at atmind.nl> wrote:
>>>>
>>>>
>>>> In April 2010 the modifiers are moved out of the kernel and into MOD_*
>>>> libraries.
>>>> http://www.pasteall.org/pic/2818 (old situation)
>>>>
>>>> Modifiers are called from within the kernel and modifiers uses (and
>>>> creates) DerivedMeshes as result. DerivedMeshes are also located in the
>>>> kernel.
>>>> http://www.pasteall.org/pic/2819 (current situation)
>>>>
>>>> Software architecture at the moment in this area is not well formalized.
>>>> The Blender kernel regulates modifiers. Modifiers initiates DerivedMeshes.
>>>> DerivedMeshes are at the moment located in the Blender kernel. In software
>>>> architecture point of view it is better to place DerivedMeshes in Blender
>>>> library. As DerivedMeshes are used in more places in the kernel this task
>>>> can be complicated.
>>>>
>>>> I want to suggest to move DerivedMeshes to Blender library. It will make
>>>> the dependancies and responsibilities between the different libraries
>>>> clearer.
>>>> http://www.pasteall.org/pic/2820 (proposed situation)
>>>>
>>>> What do you all think?
>>>>
>>>> Jeroen Bakker.
>>>>
>>>> --------------------------------------------------------------------
>>>> mail2web LIVE – Free email based on Microsoft® Exchange technology -
>>>> http://link.mail2web.com/LIVE
>>>>
>>>>
>>>> _______________________________________________
>>>> Bf-committers mailing list
>>>> Bf-committers at blender.org
>>>> http://lists.blender.org/mailman/listinfo/bf-committers
>>>>
>>>>
>>> _______________________________________________
>>> Bf-committers mailing list
>>> Bf-committers at blender.org
>>> http://lists.blender.org/mailman/listinfo/bf-committers
>>>
>>
>>
>>
>> --
>> - Campbell
>>
>
>
>
> --
> - Campbell
>



-- 
- Campbell