[Bf-committers] Compiling under Windows with MinGW

[joeedh]

sh484090 at 12move.nl wrote:

>>From: Jeremy Wall <zaphar at gmail.com>
>>I had this exact same problem myself on mingw but was never able to
>>resolve it. I'd be interested in whether someone else manages it.
>>    
>>
>
>Hi,
>
>I've been able to compile under MinGW for some time now. I currently use
>mingw with a standard msys1.0.10 setup and GCC 3.4 under Win98SE. I apply
>the following to get things to work, starting from a clear source tree.
>
>1) Add "user-def.mk" file to bf-blender/blender with following lines:
>    export FREE_WINDOWS=true
>    export NAN_NO_KETSJI=true
>    export OS=windows
>Obviously, I don't use the game engine, hence line 2. See point 2) below
>for the last line.
>
>2) In bf-blender/source/blender/nan-definitions.mk replace
>    export OS := $(shell echo ${CONFIG_GUESS} | sed -e 's/-.*//')
>by
>    export OS ?= $(shell echo ${CONFIG_GUESS} | sed -e 's/-.*//')
>The question mark allows this line to be pre-empted by the "export
>OS=windows" line in user-def.mk
>Reason for this patch is that guessconfig on msys returns "mingw-32-i386"
>and
>we need "OS=windows".
>Note: I haven't gone down a path to attempt to replace all "mingw-32-i386"
>
>  
>
Tuhopuu's guesconfig returns "windows".  Get that.

>by "windows". Also see point 7) below on the library path.
>
>3) In bf-blender/source/blender/makesdna/intern/Makefile: under target DNA.c
>ensure that "makesdna" is called and not "cl_wrapper.pl" e.g.:
>
>$(DIR)/$(DEBUG_DIR)DNA.c: $(DIR)/$(DEBUG_DIR)makesdna
>     $(DIR)/$(DEBUG_DIR)makesdna $(DIR)/$(DEBUG_DIR)DNA.c
>
>The cl_wrapper perl script generates a VC++ style compiler call 
>(which we don't need) to build DNA.o  (and I don't have perl anyway).
>"makesdna" works fine for me.
>  
>
I have to manually run makesdna myself. . .

*yawn*.   Well, that's enough for now.

joeedh