Age | Commit message (Collapse) | Author |
|
Forgot to set audio_codec correctly on output. Shame on me.
|
|
|
|
re-allocating stripdata which was causing memory free errors, added proxy options, directort read/write and some more flags.
|
|
|
|
This replaces the hard-coded '%PROFILE%\Application Data\..." path
with the correct language-dependant CSIDL_APPDATA.
PLEASE NOTE!
Checks for the previous location have now been REMOVED. Any data
saved in the old hard-coded location will no longer be found by
Blender (this will only affect users of non-english windows.)
The installer has been updated to help the user to migrate their data
to the new location of their choice, if necessary.
Order of preference for .blender on windows is now the following:
1) %HOME% environment var
2) Installation Directory (location of blender.exe)
3) "Current User" (CSIDL_APPDATA)
4) "All Users" (CSIDL_COMMON_APPDATA)
|
|
to the new system didn't work correct.
|
|
entirely correct after a recent bugfix.
|
|
|
|
|
|
|
|
|
|
light overrides, since it works independent of render layers.
|
|
for values > 1.0, luckily for peach this bug was not present
for other light types.
|
|
|
|
|
|
This would cause a very tiny inconvenience where ndof transform wouldn't work if there wasn't an active object (so, only a problem after you delete something and then select objects without setting active).
|
|
|
|
|
|
Lamps would not update either.
|
|
md2_export script
made some other small fixes while testing. (divide by zero errors)
|
|
enabled state.
bug: [#9437] Moving texture up/down bug
patch: [#8938] Moving texture up/down to also move 'enabled/disabled'-status of texture
|
|
Simple button change
|
|
|
|
[#8067] external texture plugin thread-safe modifications
Submitted By: David Anderson (davywavy)
It makes it so the "result" array is passed in, instead of a global var.
I expanded the patch so it will play nice with older plugins that are not
thread safe as well.
I also updated the existing plugins in the release, so they are thread safe.
--------------- What do people think of this.... ------------------
This should maybe be talked about in the functionality board or something,
but what do people think of adding in default texture/sequence plugins.
or making a seperate tree like lib for plugins.
The reason I ask is we have had a couple of upgrades to the plugin system.
(supporting float buffers for sequencer, and this one for textures)
http://www.cs.umn.edu/~mein/blender/plugins does not store revisions of
plugins I just make sure they work with the latest version. This is
getting messy. I haven't upgraded a lot of them to use floats (I know,
I'm lazy, and now this will also make modifications to the plugins)
It would be nice to have some of the standard ones under revision control.
We also seem to be having an explosion of platforms supported. It would
be nice to have platform maintainers compiling plugins as well for releases.
(Its getting to be more work for me to keep up with things...)
I'll go back to my corner now and be quite. ;)
Kent
|
|
Kent
|
|
other warnings that only occurred with 64bit systems. Wish I'd known about
these macros earlier!
|
|
with 64bit os's
Also use Py_ssize_t which we might need to define for older python's
|
|
problem. Now the fixed length strings are limited in the sscanf so
it should be much better now.
Kent
|
|
completely fixed yet.
|
|
blocks that were previously missed; and b) greatly increase my
ohloh stats!
|
|
empty to refer to self object
|
|
http://www.freebsd.org/cgi/cvsweb.cgi/ports/graphics/blender/files/patch-source_blender_blenlib_intern-storage.c
Kent
|
|
It is now possible for Action Channels to be displayed using the colour of its group. The colour for an Action Group is only set when it is auto-created by keyframing (controls to tweak this may/may-not be added, time permitting).
To enable the display of these colours, the View -> 'Use Group Colors' must be enabled. However, this is disabled by default.
|
|
|
|
using persp as a flag which worked but isnt correct.
|
|
Unfortunately, my previous fixes here to get gears working resulted in this problem. The flip side of that was that gears worked for quite a few cases.
Alas, it is not to be... The bug is now fixed, but gears no longer work any more. Sadly, there is not much I can do about this.
|
|
|
|
computations with the option disabled. Also small memory usage
optimization for vertex groups.
|
|
Secunia Research
SAID: SA29818 (http://secunia.com/advisories/29818/)
Credit: Stefan Cornelius, Secunia Research
The old code trys to do a sscanf %s %d %s %d from a line in the
image file.
Now it copies over that line to a max buffer of size 540 chars before doing
the sscanf.
(I just picked a constant that was siginficatly large)
It also checks to see if it gets all 4 values if not return NULL.
Kent
|
|
|
|
|
|
changing ipo's.
|
|
debugging, also added a name for each dependency relation, and when
a cycle is detected it now prints the full cycle to the console.
|
|
also added some checks for sculpt mixing with other modes (was possible to mix sculpt+posemode)
|
|
|
|
with geometry, and makes halos look more volumetric.
Patch contributed by Markus Ilmola, thanks!
|
|
working correct with translucency. Removed a check that presumably
is an optimization, but can't work correct and also gives no noticable
performance difference in my tests.
|
|
|
|
computers, probably due to slow multithreaded malloc. Now each render
thread keeps a list of qmc samplers that it fills as needed while
rendering (since it is hard to predict the actual amount needed in
advance due to ray recursion).
|
|
with edited hair.
|