blender-staging.git
11 years ago* Added a button to the volume material controls 'Alpha' to generate an
Matt Ebb [Tue, 11 Nov 2008 23:24:10 +0000 (23:24 +0000)]
* Added a button to the volume material controls 'Alpha' to generate an
alpha channel based on the volume's transmission properties, allowing you
to use it in comp etc.

I'd rather not have this button at all, and make it just work properly
by default, however it causes problems with overlapping volumes when
'premul' is on (stoopid thing..) so for the time being, there's the
button. I'll try and fix this up later on when I have more time.

11 years agoPoint Density
Matt Ebb [Mon, 10 Nov 2008 00:14:35 +0000 (00:14 +0000)]
Point Density

* Fixed a stupid crash caused by last commit that worked fine on the mac
(but never should have...)

* Fix for using child particles with the new particle age color options

11 years agoPoint Density texture: colouring
Matt Ebb [Sun, 9 Nov 2008 01:16:12 +0000 (01:16 +0000)]
Point Density texture: colouring

This introduces a few new ways of modifying the intensity and colour output
generated by the Point Density texture. Previously, the texture only output
intensity information, but now you can map it to colours along a gradient
ramp, based on information coming out of a particle system.

This lets you do things like colour a particle system based on the individual
particles' age - the main reason I need it is to fade particles out over time.

The colorband influences both the colour and intensity (using the colorband's
alpha value), which makes it easy to map a single point density texture to
both intensity values in the Map To panel (such as density or emit) and colour
values (such as absorb col or emit col). This is how the below examples are
set up, an example .blend file is available here:

http://mke3.net/blender/devel/rendering/volumetrics/pd_test4.blend

The different modes:

* Constant
No modifications to intensity or colour (pure white)

* Particle Age
Maps the color ramp along the particles' lifetimes:
http://mke3.net/blender/devel/rendering/volumetrics/pd_mod_partage.mov

* Particle Speed
Maps the color ramp to the particles' absolute speed per frame (in Blender
units). There's an additional scale parameter that you can use to bring this
speed into a 0.0 - 1.0 range, if your particles are travelling too faster or
slower than 0-1.
http://mke3.net/blender/devel/rendering/volumetrics/pd_mod_speed.mov

* Velocity -> RGB
Outputs the particle XYZ velocity vector as RGB colours. This may be useful
for comp work, or maybe in the future things like displacement. Again, there's
a scale parameter to control it.
http://mke3.net/blender/devel/rendering/volumetrics/pd_mod_velrgb.mov

11 years ago* Fixed a strange problem with the way textures are interpreted - was causing weird...
Matt Ebb [Tue, 4 Nov 2008 05:17:02 +0000 (05:17 +0000)]
* Fixed a strange problem with the way textures are interpreted - was causing weird things with point density turbulence on

* Reverted the spin field once more..

11 years agoReverted to the previous, more hackish version of the spin field.
Matt Ebb [Sun, 2 Nov 2008 23:25:21 +0000 (23:25 +0000)]
Reverted to the previous, more hackish version of the spin field.
There has to be a better way of doing this though... jahka? :)

11 years ago* Tweaked the spin force field a bit. It worked great for its purpose
Matt Ebb [Sat, 1 Nov 2008 12:45:19 +0000 (12:45 +0000)]
* Tweaked the spin force field a bit. It worked great for its purpose
before, but wasn't playing nice with the other force fields (was
overwriting the velocity rather than acting as a force). This version is
slightly different, but will work a lot better with other force fields too.

11 years ago* Added a new turbulence type: Time. It's not entirely well tested, but so far workin...
Matt Ebb [Fri, 31 Oct 2008 05:29:54 +0000 (05:29 +0000)]
* Added a new turbulence type: Time. It's not entirely well tested, but so far working ok. It's smoother looking than 'velocity' but may need more in depth investigation.

11 years ago* Added a new, slightly experimental force field type: 'Spin'.
Matt Ebb [Thu, 30 Oct 2008 09:26:30 +0000 (09:26 +0000)]
* Added a new, slightly experimental force field type: 'Spin'.
It works similarly to Vortex, but it's a more controllable - it
doesn't send the particles off accelerating into space, but
keeps them spinning around the Z axis of the force field object.

This is safe in the branch, but Jahka if you have any feedback,
I'd be curious to hear :)

11 years agosvn merge -r 17042:17201 https://svn.blender.org/svnroot/bf-blender/trunk/blender
Daniel Genrich [Mon, 27 Oct 2008 20:14:45 +0000 (20:14 +0000)]
svn merge -r 17042:17201 https://svn.blender.org/svnroot/bf-blender/trunk/blender

11 years agoface transp option CLIP wasnt added to the py api.
Campbell Barton [Mon, 27 Oct 2008 15:47:58 +0000 (15:47 +0000)]
face transp option CLIP wasnt added to the py api.
added gameObject.replaceMesh(meshname)  - needed this for an automatically generated scene where 100's of objects would have needed logic bricks automatically added. Quicker to run replace mesh on all of them from 1 script.

11 years agoFix for crash reading a peach file, chubbychestnut.blend. A do_versions
Brecht Van Lommel [Mon, 27 Oct 2008 15:22:55 +0000 (15:22 +0000)]
Fix for crash reading a peach file, chubbychestnut.blend. A do_versions
conversion was reading the mtex array in a library linked material. It
is however not guaranteed that direct_link_* was called  on the material
yet, so the array pointer is not always valid and it can crash.

11 years agoBugfixes:
Joshua Leung [Sun, 26 Oct 2008 09:41:59 +0000 (09:41 +0000)]
Bugfixes:

* #17900 - IK Constraint was not included regardless of what Visual-Keying method was used
* Deleting a Bone Group now corrects indices of those groups that occurred after the one that was deleted
* No more click-a-mania - Delete all vertex groups from a Mesh (Ctrl-Shift-G menu)

11 years agofix for [#17895] Python-generated Curves can't be beveled
Campbell Barton [Sun, 26 Oct 2008 08:06:48 +0000 (08:06 +0000)]
fix for [#17895] Python-generated Curves can't be beveled
radius and weight's values were not initialized for nurbs curves

for 2.48a just use set radius from curve specials menu to work around this.

11 years ago* More improvements for light cache
Matt Ebb [Thu, 23 Oct 2008 03:50:56 +0000 (03:50 +0000)]
* More improvements for light cache

Previously when using light cache, there could be artifacts caused when
voxel points that were sampled outside the volume object's geometry got
interpolated into the rest of the volume. This commit adds a (similar
to a dilate) filter pass after creating the light cache, that fills
these empty areas with the average of their surrounding voxels.

http://mke3.net/blender/devel/rendering/volumetrics/vol_lightcache_filter.jpg

11 years agoDebugging tests paid off, fixed a ridiculously silly bug
Matt Ebb [Thu, 23 Oct 2008 02:15:36 +0000 (02:15 +0000)]
Debugging tests paid off, fixed a ridiculously silly bug
that was preventing light cache from working on some
people's systems (but went just fine on both my windows pc
and mac). I have no idea how the original code even worked
at all, it really shouldn't have.

But fixed now anyway! Thanks a bunch to Zanqdo for patience
in helping me pinpoint this.

11 years agoremoved debug stuff
Matt Ebb [Thu, 23 Oct 2008 02:03:54 +0000 (02:03 +0000)]
removed debug stuff

11 years agoand more debugging
Matt Ebb [Thu, 23 Oct 2008 01:52:04 +0000 (01:52 +0000)]
and more debugging

11 years agomore testing code
Matt Ebb [Thu, 23 Oct 2008 01:40:49 +0000 (01:40 +0000)]
more testing code

11 years agoJust adding some temporary testing code to help try and find a bug.
Matt Ebb [Thu, 23 Oct 2008 01:14:30 +0000 (01:14 +0000)]
Just adding some temporary testing code to help try and find a bug.

11 years agodefault opengl static locations was to try /usr/libGL.a, use the oprngl lib path...
Campbell Barton [Wed, 22 Oct 2008 16:43:38 +0000 (16:43 +0000)]
default opengl static locations was to try /usr/libGL.a, use the oprngl lib path instead - /usr/X11R6/lib/libGL.a

11 years agoMissing NULL check causes crash on Ikey when no active object.
Ton Roosendaal [Wed, 22 Oct 2008 16:13:30 +0000 (16:13 +0000)]
Missing NULL check causes crash on Ikey when no active object.
Error is bad enough to go for a retag... hrmf!

Report from Ernesto (der|kunstler) Mndez in irc. Thanks a lot!

11 years agoNew splash c file (now gimp png) and bumped version to 2.48.1 to make it
Ton Roosendaal [Wed, 22 Oct 2008 11:50:12 +0000 (11:50 +0000)]
New splash c file (now gimp png) and bumped version to 2.48.1 to make it
even a bit more clear.

After this commit we'll tag svn (gasp!) and then go build!

11 years agoPart one of the release update commit
Ton Roosendaal [Wed, 22 Oct 2008 11:48:26 +0000 (11:48 +0000)]
Part one of the release update commit

11 years ago* Minor cleanup of SCons files
Nathan Letwory [Wed, 22 Oct 2008 11:28:10 +0000 (11:28 +0000)]
* Minor cleanup of SCons files

- cleanup of boolean usage - use True and False now instead of 'true'/'false' or 0/1
- changed SConscripts accordingly

11 years ago* Did some small tweaks to how density is used with light
Matt Ebb [Wed, 22 Oct 2008 09:26:18 +0000 (09:26 +0000)]
* Did some small tweaks to how density is used with light
cache - it makes some very good improvements clearing up artifacts:

http://mke3.net/blender/devel/rendering/volumetrics/vol_lc_fixed.jpg

11 years agoadded an option for python Draw.UIBlock(func, mouse_exit) so moving the mouse outside...
Campbell Barton [Wed, 22 Oct 2008 08:21:43 +0000 (08:21 +0000)]
added an option for python Draw.UIBlock(func, mouse_exit) so moving the mouse outside the popup wont close it.
Stops FBX Export and OBJ I/O from flickering a lot.

11 years agomake sure BPY_Err_Handle clears python errors, even if the exception cant be printed...
Campbell Barton [Wed, 22 Oct 2008 07:09:15 +0000 (07:09 +0000)]
make sure BPY_Err_Handle clears python errors, even if the exception cant be printed. Added PyErr_Clear() incase there are other references to exception data (sys.exc_info() from python)

11 years ago* comment fix
Nathan Letwory [Wed, 22 Oct 2008 07:02:30 +0000 (07:02 +0000)]
* comment fix

11 years ago* Fixed a bug in copying (making single user) a point density texture
Matt Ebb [Wed, 22 Oct 2008 06:10:16 +0000 (06:10 +0000)]
* Fixed a bug in copying (making single user) a point density texture

11 years ago(no commit message)
Joilnen Leite [Wed, 22 Oct 2008 05:35:23 +0000 (05:35 +0000)]

11 years ago(no commit message)
Joilnen Leite [Wed, 22 Oct 2008 05:33:45 +0000 (05:33 +0000)]

11 years ago* fix for point density - particle systems weren't being
Matt Ebb [Wed, 22 Oct 2008 05:24:41 +0000 (05:24 +0000)]
* fix for point density - particle systems weren't being
 deformed by lattices

11 years agomissing undo for vgroup operations, countall was running twice from select/deselect...
Campbell Barton [Wed, 22 Oct 2008 04:42:00 +0000 (04:42 +0000)]
missing undo for vgroup operations, countall was running twice from select/deselect and DAG_object_flush_update isnt needed for changed selections.

11 years agofix for [#17878] Scripts operating on blender objects don't clear memory after a...
Campbell Barton [Wed, 22 Oct 2008 03:10:00 +0000 (03:10 +0000)]
fix for [#17878] Scripts operating on blender objects don't clear memory after a crash
This is an interesting bug since it is likely the cause of many other suspicious python crashes in blender.

sys.last_traceback would store references to PyObjects at the point of the crash.
it would only free these when sys.last_traceback was set again or on exit.

This caused many crashes in the BGE while testing since python would end up freeing invalid game objects -
When running scripts with errors, Blender would crash every 2-5 runs - in my test just now it crashed after 4 trys.

It could also segfault blender, when (for eg) you run a script that has objects referenced. then load a new file and run another script that raises an error.
In this case all the invalid Blender-Object's user counts would be decremented, even though none of the pointers were still valid.

11 years agoTweaked rendering inside a volume to always return alpha 1.0.
Matt Ebb [Wed, 22 Oct 2008 02:59:33 +0000 (02:59 +0000)]
Tweaked rendering inside a volume to always return alpha 1.0.
This fixes an issue which darkened the render from inside a
volume with sky or premul on. Still need to find a good way to
get an alpha value back into the shader (for compositing etc)
without getting the render distorted by premul.

11 years agoPoint Density:
Matt Ebb [Wed, 22 Oct 2008 01:31:46 +0000 (01:31 +0000)]
Point Density:

Replaced 'Sharp' falloff with 'Soft'. This falloff type has
a variable softness, and can get some quite smooth results.
It can be useful to get smooth transitions in density when
you're using particles on a large scale:

http://mke3.net/blender/devel/rendering/volumetrics/pd_falloff_soft.jpg

Also removed 'angular velocity' turbulence source - it
wasn't doing anything useful atm

11 years agofix for a compile error with msvc
Matt Ebb [Wed, 22 Oct 2008 00:26:19 +0000 (00:26 +0000)]
fix for a compile error with msvc

11 years ago* Fixed a crash with using light cache on a plane (which doesn't really make sense...
Matt Ebb [Wed, 22 Oct 2008 00:09:12 +0000 (00:09 +0000)]
* Fixed a crash with using light cache on a plane (which doesn't really make sense, but shouldn't crash at least)

11 years agoadd a way for external scripts to call X3D/VRML importer and deal with unknown node...
Campbell Barton [Tue, 21 Oct 2008 23:14:40 +0000 (23:14 +0000)]
add a way for external scripts to call X3D/VRML importer and deal with unknown node types.

11 years agoPatch from Timothy Baldridge
Nathan Letwory [Tue, 21 Oct 2008 23:07:09 +0000 (23:07 +0000)]
Patch from Timothy Baldridge
* add irix6 to GHOST windowing system (same as linux2 et al)
* fix faulty return lines in shrinkwrap.c

11 years agoFix for bug #17881: with SSS disabled, SSS materials didn't
Brecht Van Lommel [Tue, 21 Oct 2008 14:46:00 +0000 (14:46 +0000)]
Fix for bug #17881: with SSS disabled, SSS materials didn't
render textures, related to other recent bugfix for baking.

11 years ago* improvement for CSizeof
Nathan Letwory [Tue, 21 Oct 2008 10:38:24 +0000 (10:38 +0000)]
* improvement for CSizeof
  - instead of fiddling around with strings, just compare pointers. (It's a good thing we have ideasman_42)

11 years agomore python 2.3 compat, should be the last of it.
Campbell Barton [Tue, 21 Oct 2008 09:08:46 +0000 (09:08 +0000)]
more python 2.3 compat, should be the last of it.

11 years agoDuplicating bones in EditMode didn't set the custom-shape and bone-groups too.
Joshua Leung [Tue, 21 Oct 2008 08:53:39 +0000 (08:53 +0000)]
Duplicating bones in EditMode didn't set the custom-shape and bone-groups too.

11 years agoBugfix:
Joshua Leung [Tue, 21 Oct 2008 08:30:02 +0000 (08:30 +0000)]
Bugfix:

"Warning: binarysearch_bezt_index encountered invalid array" errors were being displayed in the console. Was caused by 3d-view show-keyframe for infostring stuff, when an IPO being checked had no keyframes.

11 years ago* Add some missing docu for Blender.Types
Nathan Letwory [Tue, 21 Oct 2008 08:27:38 +0000 (08:27 +0000)]
* Add some missing docu for Blender.Types
* Fix potential crash in CSizeof()

11 years ago* reimplemented some things, hopefully may fix some problems zanqdo was having
Matt Ebb [Tue, 21 Oct 2008 08:21:36 +0000 (08:21 +0000)]
* reimplemented some things, hopefully may fix some problems zanqdo was having

11 years agoBugfixes:
Joshua Leung [Tue, 21 Oct 2008 08:00:19 +0000 (08:00 +0000)]
Bugfixes:

- Adding constraint using button in panel still didn't update Armature Editing buttons properly.
- Minor code tidying of earlier bugfix for armatures
- 'For Transform' option for Limit constraints is now only taken into account for constraints that are enabled.

11 years agofixed a crash in volume shadows
Matt Ebb [Tue, 21 Oct 2008 06:10:36 +0000 (06:10 +0000)]
fixed a crash in volume shadows

11 years ago* fixed a memory leak that was happening during preview render
Matt Ebb [Tue, 21 Oct 2008 02:04:29 +0000 (02:04 +0000)]
* fixed a memory leak that was happening during preview render

11 years ago#17873: "switch direction" for bones can cause infinite loop
Joshua Leung [Tue, 21 Oct 2008 01:40:47 +0000 (01:40 +0000)]
#17873: "switch direction" for bones can cause infinite loop

Second attempt at fixing this bug. Previous fix caused segfault when all bones in a chain are selected. Now it should segments which are selected (i.e. get swapped) will get unparented from segments that aren't (i.e. aren't swapped, so are still in old orientation)

11 years ago[#17873] "switch direction" for bones can cause infinite loop
Martin Poirier [Tue, 21 Oct 2008 01:11:18 +0000 (01:11 +0000)]
[#17873] "switch direction" for bones can cause infinite loop

Stop the infinite loop, but it could be loosing parent relations that it kept before. Aligorith needs to review.

11 years agopy 2.3 compat for lightwave_import and wizard_curve2tree
Campbell Barton [Tue, 21 Oct 2008 00:21:02 +0000 (00:21 +0000)]
py 2.3 compat for lightwave_import and wizard_curve2tree

11 years agorenamed WITH_BF_BPYDOC to WITH_BF_DOCS, added SConscript for building BGE docs too.
Campbell Barton [Mon, 20 Oct 2008 23:21:01 +0000 (23:21 +0000)]
renamed WITH_BF_BPYDOC to WITH_BF_DOCS, added SConscript for building BGE docs too.

11 years ago* fix for crash after latest light cache commit
Matt Ebb [Mon, 20 Oct 2008 23:12:42 +0000 (23:12 +0000)]
* fix for crash after latest light cache commit

11 years agofix for [#17871] PLY file import: blender seems to loose vertex color information
Campbell Barton [Mon, 20 Oct 2008 13:36:18 +0000 (13:36 +0000)]
fix for [#17871] PLY file import: blender seems to loose vertex color information
use Mesh rather then NMesh

11 years ago=== Blender Python API ===
Nathan Letwory [Mon, 20 Oct 2008 12:33:31 +0000 (12:33 +0000)]
=== Blender Python API ===

After some discussion with Campbell, changed the way cstruct sizeof is fetched.

Moved DataSize() to Blender.Types.CSizeof(Blendertype). Supported types return sizeof(data struct), otherwise -1.

To quickly check what types are supported:

import Blender.Types as bt
x = dir(bt)
for t in x:
if t[0] != '_':
s = 'bt.CSizeof(bt.' + t + ')'
print t,"=", eval(s)

11 years agoAnother constraint panel (in Armature Editing Panels) bugfix. This time, when adding...
Joshua Leung [Mon, 20 Oct 2008 09:35:07 +0000 (09:35 +0000)]
Another constraint panel (in Armature Editing Panels) bugfix. This time, when adding a new constraint using vertical layout.

11 years agoA few warnings missed in previous sweep
Joshua Leung [Mon, 20 Oct 2008 09:24:25 +0000 (09:24 +0000)]
A few warnings missed in previous sweep

11 years agoNew volume rendering feature: Light Cache
Matt Ebb [Mon, 20 Oct 2008 07:08:06 +0000 (07:08 +0000)]
New volume rendering feature: Light Cache

This was a bit complicated to do, but is working pretty well now, and can make shading significantly faster to render.

This option pre-calculates self-shading information into a
3d voxel grid before rendering, then uses and interpolates
that data during the main rendering phase, rather than
calculating shading for each sample. It's an approximation
and isn't as accurate as getting the lighting directly,
but in many cases it looks very similar and renders much faster.
The voxel grid covers the object's 3D screen-aligned bounding box
so this may not be that useful for large volume regions like a
big range of cloud cover, since you'll need a lot of resolution.

The render time speaks for itself here:
http://mke3.net/blender/devel/rendering/volumetrics/vol_light_cache_interpolation.jpg

The resolution is set in the volume panel - it's the resolution
of one edge of the voxel grid. Keep in mind that the higher the
resolution, the more memory needed, like in fluid sim. The
memory requirements increase with the cube of the edge
resolution so be careful. I might try and add a little memory
calculator thing like fluid sim has there later.

The voxels are interpolated using trilinear interpolation -
here's a comparison image I made during testing:
http://mke3.net/blender/devel/rendering/volumetrics/vol_light_cache_compare.jpg

There might still be a couple of little tweaks I can do to
improve the visual quality, I'll see.

11 years agoReduced number of MSVC compiler warnings (mostly just casting issues).
Joshua Leung [Mon, 20 Oct 2008 06:39:08 +0000 (06:39 +0000)]
Reduced number of MSVC compiler warnings (mostly just casting issues).

11 years ago- Bugfix: 'Hide' button for PoseMode bones (in Editing Buttons) was setting the ...
Joshua Leung [Mon, 20 Oct 2008 04:53:49 +0000 (04:53 +0000)]
- Bugfix: 'Hide' button for PoseMode bones (in Editing Buttons) was setting the 'Mult' flag instead
- Tidied up comments in Action Editor code a bit

11 years agoFixed overlapping panels in Armature Buttons that resulted after adding a constraint...
Joshua Leung [Sun, 19 Oct 2008 23:52:28 +0000 (23:52 +0000)]
Fixed overlapping panels in Armature Buttons that resulted after adding a constraint/IK using the hotkeys.

11 years ago=== Blender Python API ===
Nathan Letwory [Sun, 19 Oct 2008 23:20:17 +0000 (23:20 +0000)]
=== Blender Python API ===

* add DataSize() to module level
  with this one can get datablock struct size.

11 years agoforgot to add removeRenderLayer() documentation
Nathan Letwory [Sun, 19 Oct 2008 22:09:45 +0000 (22:09 +0000)]
forgot to add removeRenderLayer() documentation

11 years ago=== SCons ===
Nathan Letwory [Sun, 19 Oct 2008 22:01:46 +0000 (22:01 +0000)]
=== SCons ===

[#17867] Adds option to SCONS to generate Python API documentation

Added patch from Brandano with some small improvements (BF_DOCDIR, clean) by yours truly.

To use make sure you have epydoc installed. Enable with WITH_BF_BPYDOC=1.

11 years ago=== Blender Python API ===
Nathan Letwory [Sun, 19 Oct 2008 21:39:50 +0000 (21:39 +0000)]
=== Blender Python API ===
* fix two typos in RenderLayer API (renderosiy -> renderosity in two places. Will break .py's saved with render_save_layers.py, just fix passRadiosiy and passRadiosiyXOR)
* add some docs on RenderLayer API
* fix some copy/paste leftover in render_save_layers.py

11 years agoAdded WITH_BF_NOBLENDER to scons so that blenderplayer can be compiler by itself...
Ken Hughes [Sun, 19 Oct 2008 21:25:17 +0000 (21:25 +0000)]
Added WITH_BF_NOBLENDER to scons so that blenderplayer can be compiler by itself (false by default).  Also added dependency for 'blender' on command line;  seems 'blender' was intended to be a target, but would give the error "Do not know how to make target `blender' ".

11 years agomodified OBJ import and export UI's to be less confusing with nicer layout. removed...
Campbell Barton [Sun, 19 Oct 2008 15:53:22 +0000 (15:53 +0000)]
modified OBJ import and export UI's to be less confusing with nicer layout. removed an unneeded workaround in fbx export.

11 years ago* small code change from Split to env.Glob()
Nathan Letwory [Sun, 19 Oct 2008 13:51:52 +0000 (13:51 +0000)]
* small code change from Split to env.Glob()

11 years ago=== BPy API ===
Nathan Letwory [Sun, 19 Oct 2008 13:48:37 +0000 (13:48 +0000)]
=== BPy API ===
* Add read access for octree resolution

11 years agoAnother WIP commit, nothing to see yet.
Matt Ebb [Sun, 19 Oct 2008 08:25:10 +0000 (08:25 +0000)]
Another WIP commit, nothing to see yet.

11 years agoonly set the SDL audio driver to alsa when not running in background mode and when...
Campbell Barton [Sun, 19 Oct 2008 06:12:11 +0000 (06:12 +0000)]
only set the SDL audio driver to alsa when not running in background mode and when blender is compiled with SDL.

11 years agosource/blender/blenloader/intern/readfile.c - use memmove rather then strncpy for...
Campbell Barton [Sun, 19 Oct 2008 04:02:37 +0000 (04:02 +0000)]
source/blender/blenloader/intern/readfile.c - use memmove rather then strncpy for overlapping strings.
source/blender/blenlib/intern/fileops.c - zero length strings would check for a slash before the strings first char.
source/gameengine/GameLogic/SCA_JoystickSensor.cpp - m_istrig_prev was not initialized
source/blender/src/editmesh.c - active face pointer was not set to NULL in free_editMesh()

11 years agoFew tiny cleanups in Action Editor code (comments only)
Joshua Leung [Sun, 19 Oct 2008 03:41:49 +0000 (03:41 +0000)]
Few tiny cleanups in Action Editor code (comments only)

11 years agoBGE bug #17858 fixed: [SHOWSTOPPER] BGE Memory leak. The memory leak has been identi...
Benoit Bolsee [Sat, 18 Oct 2008 18:37:58 +0000 (18:37 +0000)]
BGE bug #17858 fixed: [SHOWSTOPPER]  BGE Memory leak. The memory leak has been identified and fixed in bullet.

11 years agopatch [#17856] Correction on docs for Ipo module. from Lucas Vella (lvella) + some...
Campbell Barton [Sat, 18 Oct 2008 13:33:27 +0000 (13:33 +0000)]
patch [#17856] Correction on docs for Ipo module. from Lucas Vella (lvella) + some other minor changes.

11 years agoAdded joystick epydocs, only changed incorrect docstrings in SCA_JoystickSensor.cpp
Campbell Barton [Sat, 18 Oct 2008 13:23:30 +0000 (13:23 +0000)]
Added joystick epydocs, only changed incorrect docstrings in SCA_JoystickSensor.cpp
patch [#17857] HotKeys Update by dfelinto - SmoothStroke and Anchored.
Made own edits, removed videoscape and added curve and armature specials.

11 years ago* Fixed eekadoodle problem where faces with a zero index vert in the wrong place...
Campbell Barton [Sat, 18 Oct 2008 09:04:25 +0000 (09:04 +0000)]
* Fixed eekadoodle problem where faces with a zero index vert in the wrong place would reotate the UVs and VCols incorrectly
* Fixed an error that made IndexedLineSet and IndexedPointSet not load for x3d's
* Added support for opening gzip x3d/wrl files

11 years agox3d and vrml importer for loading thes files as static scenes. x3d needs a full pytho...
Campbell Barton [Fri, 17 Oct 2008 19:06:24 +0000 (19:06 +0000)]
x3d and vrml importer for loading thes files as static scenes. x3d needs a full python install. VRML97 files can be imported without python.

for details on what is supported see
http://wiki.blender.org/index.php/Scripts/Manual/Import/X3D_VRML97#Compatibility

11 years agoFix for bug #17829: with ATI drivers, GLSL shadows crashed on physics
Brecht Van Lommel [Fri, 17 Oct 2008 13:36:40 +0000 (13:36 +0000)]
Fix for bug #17829: with ATI drivers, GLSL shadows crashed on physics
'show pivot' drawing in the viewport. Also added checks for some other
cases where this could happen.

11 years agoFix for imagepaint soften tool giving a bit too dark results,
Brecht Van Lommel [Fri, 17 Oct 2008 13:34:20 +0000 (13:34 +0000)]
Fix for imagepaint soften tool giving a bit too dark results,
due to poor float-to-char conversion.

11 years agoBGE bug fix (continuation of previous bug fix):
Benoit Bolsee [Fri, 17 Oct 2008 09:24:35 +0000 (09:24 +0000)]
BGE bug fix (continuation of previous bug fix):
- Forgot to make SCA_ISensor::UnregisterToManager() virtual to intercept active-inactive transition on collision sensor to clear colliders reference.
- Don't record collision on inactive sensor.
  This situation occurs when an object with an inactive collision sensor collides with an object with an active collision sensor: the collision handler triggers both sensors.
  The result of this bug was pending references that eventually cause temporary memory leak (until the sensor is reactivated).

11 years agonon-working WIP commit to continue coding at home.
Matt Ebb [Fri, 17 Oct 2008 05:54:42 +0000 (05:54 +0000)]
non-working WIP commit to continue coding at home.

nothing to see here, move along!

11 years agoBGE showstopper bug fix:
Benoit Bolsee [Thu, 16 Oct 2008 23:33:40 +0000 (23:33 +0000)]
BGE showstopper bug fix:
- Reset hit object pointer at end of frame of touch sensor to avoid returning invalid pointer to getHitObject().
- Clear all references in KX_TouchSensor::m_colliders when the sensor is disabled to avoid loose references.
- Test GetSGNode() systematically for all KX_GameObject functions that can be called from python in case a python controller keeps a reference in GameLogic (bad practice anyway).

11 years ago[#17845] View Snapping causes background grid confusion
Martin Poirier [Thu, 16 Oct 2008 21:43:25 +0000 (21:43 +0000)]
[#17845] View Snapping causes background grid confusion

view needs to be reset while moving to take care of view snapping properly.

11 years agovrml exporting with images was broken.
Campbell Barton [Tue, 14 Oct 2008 23:43:08 +0000 (23:43 +0000)]
vrml exporting with images was broken.

11 years agocleanup epydoc errors
Campbell Barton [Tue, 14 Oct 2008 22:31:10 +0000 (22:31 +0000)]
cleanup epydoc errors

11 years ago== windows installer ==
Andrea Weikert [Tue, 14 Oct 2008 20:27:37 +0000 (20:27 +0000)]
== windows installer ==
small update for release 2.48 - no need to re-tag

11 years ago=== Installer (win32) ===
Nathan Letwory [Tue, 14 Oct 2008 16:24:27 +0000 (16:24 +0000)]
=== Installer (win32) ===

* Fix for [#17651] Silent Install Issue
  - make sure silent install is really silent ( /S on command-line)
* Improve installer to check for msvc90.dll
* Improve installer to check for py 25 install (using registry)

11 years agoresaved in the gimp, now loads with libpng 1.2.30
Campbell Barton [Tue, 14 Oct 2008 14:10:49 +0000 (14:10 +0000)]
resaved in the gimp, now loads with libpng 1.2.30

11 years agoBugfix #17832
Ton Roosendaal [Tue, 14 Oct 2008 12:16:12 +0000 (12:16 +0000)]
Bugfix #17832

Particle system crashed in convertblender, missing NULL check.
This fixes crash, render survives, but I doubt it was meant so...

Will leave it Janne to evaluate later.

11 years agoCollisions: CollisionModifier needs to be resetted for every loop
Daniel Genrich [Tue, 14 Oct 2008 11:54:07 +0000 (11:54 +0000)]
Collisions: CollisionModifier needs to be resetted for every loop

11 years agoBugfix for regression from 2.46 on cloth collisions reported by Daniel Salazar via IRC
Daniel Genrich [Tue, 14 Oct 2008 11:50:30 +0000 (11:50 +0000)]
Bugfix for regression from 2.46 on cloth collisions reported by Daniel Salazar via IRC

11 years agoBugfix on IRC report. Luckily Daniel Salazar *does* carefully check our
Ton Roosendaal [Tue, 14 Oct 2008 11:42:53 +0000 (11:42 +0000)]
Bugfix on IRC report. Luckily Daniel Salazar *does* carefully check our
regression files, it showed small dark outline errors on envmap.

This commit rewinds bugfix #8437, which actually had to be fixed on
another location in code, which was done a few weeks ago. :)

11 years agoBugreport in IRC: Area lamp, "Constant jitterred" had wrong alignment
Ton Roosendaal [Tue, 14 Oct 2008 10:55:18 +0000 (10:55 +0000)]
Bugreport in IRC: Area lamp, "Constant jitterred" had wrong alignment
code, causing the "SamplesY" button to overlap others.

11 years agoBugfix #17830
Ton Roosendaal [Tue, 14 Oct 2008 10:44:22 +0000 (10:44 +0000)]
Bugfix #17830

Index OB pass didn't support FSA for Ztransp.
Also made buttons to set black/white for non-RGBA images hide in Image Window,
the Curves color code only supports 4 channels atm.

11 years agoAction Editor - Grease Pencil Bugfix:
Joshua Leung [Tue, 14 Oct 2008 10:11:01 +0000 (10:11 +0000)]
Action Editor - Grease Pencil Bugfix:

When Action Editor was maximised, Grease Pencil data disappeared. This was due to the screen-swapping that went on. Now, it uses the old-screen that was stored in the maximised Action Editor's screen (should be safe...)

11 years agoGrease Pencil Bugfixes:
Joshua Leung [Tue, 14 Oct 2008 09:36:08 +0000 (09:36 +0000)]
Grease Pencil Bugfixes:
* Onion-skinning with GStep > 0 was not showing enough of a noticable difference between ghosts. Improved method of calculating this.

* Clicking in a Grease-Pencil datablock channel in the Action Editor would crash

11 years agobugfix: copy/paste bug merged from Bullet main repo into Blender, hope it still can...
Erwin Coumans [Tue, 14 Oct 2008 06:14:25 +0000 (06:14 +0000)]
bugfix: copy/paste bug merged from Bullet main repo into Blender, hope it still can make 2.48