[Bf-committers] Re: [Bf-blender-cvs] CVS commit: blender/source/blender/blenkernel/intern action.c ipo.c blender/source/blender/include BSE_editipo.h blender/source/blender/makesdna DNA_userdef_types.h blender/source/blender/python SConscript blender/source/blender/python/api2_2x NLA.c Object.c Types.c ...

[Stephane SOPPERA]

Toni Alatalo wrote:

>On Monday 19 December 2005 01:13, Stephane SOPPERA wrote:
>  
>
>>The following commit to orange branch does not compile with VC7:
>>    
>>
>
>orange does not use VC.. ;p
>
>but seriously: i should have committed that stuff only today after cleaning, 
>got a bit too enthusiastic with the baking, and the guys were rushing for the 
>autokey-avail-only patch, so i acted too quickly yesterday (i did read all 
>the code :)
>
>  
>
>>the function bake_all_to_action is used (line 1668) before being
>>declared (line 1847 where it is both declared and defined).
>>So when compiling, at least with VC7, at the line 1668 the compiler
>>assumes it returns an "int". Consequently it does not compile on line
>>1847 where the compiler considers the function is re-declared with a
>>different return type.
>>    
>>
>
>many of these new, partly experimental functions in editnla.c are to be 
>removed even - will try to have the time before christmas to do that and fix 
>the ones that are to actually remain.
>
>in the meantime, am sorry for the trouble.
>  
>
Not that much a trouble. I can still compile by modifying my local image 
of the CVS. I was mentioning this compiling issue so that it can be 
correction before the orange branch is merged back  in the main tree.

By the way, which compiler do you use? I am suprised that a compiler can 
compile line 1668 without complaining. It should not.

Stéphane