[windows] Windows with Jahka particle patch

General Forums Blender and CG Discussions
1

Hello Jahka make a gift for the new year : a new patch for his particle system :

Download :

2

Thanks for jahka and thanks for you :slight_smile:

3

thanks LGuillaume and happy new year ^^

4

Dani, nice to see you, long time I don’t talk with you.

Comme on dit bananier et pomme sautés (<= French joke can’t be translated)

5

Has anyone gotten the patch to compile on Linux? Don’t know if it’s just me being careless and not finding a simple issue, but building it on Linux causes the following error after the patch is applied:

"redefinition of ‘RE_inithalo_new’ "

I got the latest CVS source this morning and tried to build a “regular” version but oddly there is a compilation failure in it as well. When it goes to link the “blender” program it fails with "undefined reference to ‘get_particle_state’ and ‘count_psyskeys’ ". I don’t know if the two problems are related. Does anyone know if they are and I’m just to shallow to see the fix?

6

Compiled 100% fine here. Do you have a recent gcc? I’m compiling fine both
unpatched and patched cvs, so it’s definitely something on your computer
or the config you’re using to compile (I’m not compiling the game module,can’t
speak about that).

7

Mattepiu,
care to share that linux build? On GraphicAll.org for instance? Currently the Windows guys are getting all the fun…

8

64bit linux build is available. :wink:

9

Odd. I compiled the CVS version of Blender fine last night. Updated CVS today and it won’t compile even the plain CVS version. It complains about the “get_particle_state” and such mentioned above. I’m at a loss now, as I’ve not changed anything since last night. Odd?

10

Is this patch going to be included in the 2.43 build? (Hope so!)

11

Nope. No way. Waaaay to late for that. At this moment, this is ‘only’ a patch. It still has to be looked at carefully for integration into the main code base. Too bad of course but totally understandable. Blender is losing a bit of its fabled stability on account of the amount of things being put in. Having said that, I’ll switch to a future 2.43-based build incorporating this patch in an instant. I just LOVE the physic based stuff (fluids, softbody, particles, rigid body,…), probably 'cause I’m such a lazy modeler…

12

OK, FWIW I got the “normal” CVS version compiling again. I ran a “scons --clean” and re-downloaded everything from CVS again (just to be safe) and it compiles fine again. I still can’t compile the patches for the particles though. I still get the “redefinition of RE_inithalo_new” issue I reported above.

Oh, well. I’ll keep plugging away at it!

13

Here’s an obvious question, have you tried emailing Jahka about it?

14

Nah, I’ve not emailed Jahka. I figure he’s a busy fella and don’t want me meddlin’ in. I’ll keep at it until I pretty much crumble, then go see what gives. I’m going to “start clean” from CVS again, apply the patches and go for broke!

Thanks guys!

15

OK, got it! I finally held my mouth right or something! Actually, I think it was the combination of running two “build” directories/trees. I removed everything from the “particle” build folder, got the latest CVS source as of 12/33/2006, copied the three files from the wiki and put them in their directories and applied the code patch. (Note, the FULL patch would not run – never has on my machine. It freaks out when it starts creating the “new” header and particle.c files – line 6393.)

I did a build and it worked OK this time. The secret is doing thing in the RIGHT ORDER (obvious as it is NOW) AND doing a “scons --clean” in between any builds of the trees.

Thanks guys!