Wave Modifier defaulted to Y waves, even when both X and Y buttons were
disabled. That case isn't very useful, nevertheless... if buttons allows it
then it should just work!
Another thread render issue, now in Image texture, the 'repeat' value was
still a global... how did I think that would ever work? Just forgot about
that one I guess. :)
Sequence renders, calling scenes with compositing, didn't execute composite
correctly. Confusement caused by the rule that a "Render" handle has same
name as Scene, which gives conflict for the case when a Scene has
sequencing with Scene strips with its own scene in it.
The previous solution for that conflict caused composite not to work. This
commit solves that, but it is still hackish. Main reason is the still
bad global G.scene, in use by compositor.
Render timers were called in the internal render loops (tile processor),
they still had to be moved to the outer loop, so they include fields/blur
or sequencer render timing correctly.
Thread render error: a flag was stored in read-only data to indicate
whether top or bottom hemisphere of sky was rendered. That can't work
with tiles (but did work when scanlines were threads).
Takes a list of polylines and returns a list of face index triplets.
Added this so using mesh.fill() could be avoided since it requires making an object, linking to the scene and cycling editmode for every NGon imported.
Since this is so close to release, It might be good if ken, willian or stivs takes a look at the function to make sure its ok. - Ran 100's of times for importing lightwave models but would be good to doublecheck.
Added extra security for fill faces. This function is being used for
importing now too, and it gets called for quite weird fill cases...
nevertheless, it should not crash!
AO option "Use Distances" does not work for colored AO, only for "Plain".
I've added this info in tooltip, and added event that resets the color
option for AO when "Use Distances" pressed.
In Editmode armature, the autocomplete for bone names used the bones from
the armature itself, not the edit data.
While fixing, also found out autocomplete for vertexgroup didnt work even.
suffered for the entire movie. :)
It only happened when rendering large frames, using a lot of memory and
typically when you also use other software in meantime.
Reason: the main thread does the drawing updating, while rendering is
still continuing. When using Ztransp, there was a free buffer done
when possibly a draw could still be in progress. Only crashed when drawing
is slow... explaining why it only showed up in more complex cases.
The 'edge fill' code failed on filling tiny small polygons. It has a limit
check for double points, which was hardcoded set to 0.0003.
That is (commented in code too) a weak part. Better would be to define a
bounding box first, and then derive the limit from that.
Further, the edge fill code uses blender EditEdge data, which fails when
the filling code removes edges. Certainly a topic to work on once, this
code is from the 80ies!
Weightpaint fix.
Weight painting was recoded in 2.40 to support all features from vertex
paint. That code uses a 'soft' brush by default, which makes it impossible
to assign exact weight values on a single click (which used to be possible)
I've made that an option now, so you can disable it for single-click setting
of values on vertices within the brush (if 'use vertex distances' is
set, of course).
- When saving image from Image Window, the extension was not added (with
the option 'add extensions' set, in Scene buttons output panel)
- Same situation, after a save the data-block Image name was still the old
one, also on 'reload' (in header and browsing menus). Confusing... so
when saving image under new name now also renames the Image block.
A proper check for the error "No Camera" on rendering can only be made
after a renderwindow was initialized. The error menu then shows in the
render window, which appears to not work OK in windows ATI (again!).
So; I've moved the test to before the renderwindow is activated, this is
not a test checking on the entire render pipeline (like Composite nodes or
sequence strips). In case more complex setups cannot render, an error is
printed in the console only.
Small tweak in warning levels for buttons showing Library data.
- Material, panel "Links" did still allow to change some settings
- Material, panel "Texture" did not allow to view different channels
Debugging complex files with drivers is very hard... so I've added the
drivers now to show in Outliner too (under the Ipo).
Note that this works for Actions too, but only when the action is linked
to an Object directly.
(For a real diff you have check previous commit, I forgot to write a log
for that)
In outliner, the icons sometimes were drawing too large or too small.
Same happened in NLA, Action, Image window etc. And it happened
for "International fonts" when set to use 'texture drawing'.
Reason: the API call for setting icon size BIF_icon_set_aspect() was not
used consistantly. Sometimes it was set, sometimes not. And even worse,
for every icon drawn in UI buttons, the icon lookup had to be done twice
because of the aspect function.
Solved it by removing this call, and adding a new function:
BIF_icon_draw_aspect()
The old BIF_icon_draw() call now draws with aspect 1.0 always. The icons
code already had optimal checking for changed sizes, zo a change in aspect
won't result in much cpu overhead. Plus it saves calling icons lookup code,
which will make it all a bit faster.
Andrea: I've added this aspect function a long while ago, I think you also
like it better how it is now? Please check!
Blender.sys.splitext() raises an error if the path is longer then 80 chars. seems a bit short, noted this anyhow end added an example.
(hope its okay to do small doc corrections as I notice them.)
Another outliner + buttonswindow select error:
When clicking on lamp icon, the buttons window shows material, not lamp
buttons. Same happened for F5key btw, when buttons were showing world.
Also fixed: when switching to world buttons with F8key, the preview was
not re-rendered.
In compositor you can mix RGBA and Value buffers freely, but with one
exception... the Composite (output) node!
This solves a crash when connecting a Alpha socket to RGBA input in
Composite node.
Depsgraph fix: the signal DAG_object_flush_update() should flush changes in
'data' to all object users of that data, not only in current scene.
Error was that switching scenes sometimes showed invalid derivedmesh data,
with bezerk drawing.
Boolean modifier allowed to be added after subsurf, but it doesn't support
that. Added "eModifierTypeFlag_RequiresOriginalData" flag to prevent this
from happening in UI.
- Crash on collapse with subsurf modifier. Cause was a missing DAG update
when calling from the 3d view header menu.
There really should be some standard way to place these countall, DAG
update, redraw and undo push calls for editmesh tools, now they're in
random places in the editmesh and UI code.
Not clearing the value for 'parts rendered' caused 3d previewrender to
sometimes fail doing updates (when an escape happens during converting
render data).
Bug #4484:
http://projects.blender.org/tracker/?func=detail&aid=4484&group_id=9&atid=125
A recent update to Draw.c made the button attr setter function not exit
where it should for one of the cases, when new and old strings had
same length (code was missing a "return 0;" line in one of the if
cases). Should work fine now, I don't get the error under linux anymore.
Also saw in the same function that a py object could be created but
was not being decref'ed in two other 'if' cases.
Thanks elbarto for reporting the bug.
Node Editor.
Fix for a fix: cleanup of abuse of uiBlock accidentally disabled the
feature that only makes Node buttons accessible when visible. This to
allow Nodes to overlap and properly used.
#4499 Fix for build of blenderplayer on windows/mingw/scons
fixes some invalid cflags for gcc on windows/mingw/scons
-if sys.platform=='win32':
+if (sys.platform=='win32') & ( Environment().subst('$CC') != 'gcc'):
cflags = ['/GR']
Patch looks ok but no mingw evironment here to test.
So mingw/sconsers to your testing stations.
We can always roll back!
Wire render crashes in Linux (only!), but why? Using a calloc instead of
malloc fixes it, but that doesn't solve the real issue in the code.
I can only imagine that bsearch() in linux has a completely different
implementation...
Anyhoo; let's commit the calloc, at least then we got stable wire render!
A subsequent hotkey press like F5 now cycles through the sub-context in
buttons. However, this should not happen when clicking on Material icon
in outliner!
