Render:
- New; support for dual CPU render (SDL thread)
Currently only works with alternating scanlines, but gives excellent
performance. For both normal render as unified implemented.
Note the "mutex" locks on z-transp buffer render and imbuf loads.
- This has been made possible by major cleanups in render code, especially
getting rid of globals (example Tin Tr Tg Tb Ta for textures) or struct
OSA or using Materials or Texture data to write to.
- Made normal render fully 4x32 floats too, and removed all old optimizes
with chars or shorts.
- Made normal render and unified render use same code for sky and halo
render, giving equal (and better) results for halo render. Old render
now also uses PostProcess options (brightness, mul, gamma)
- Added option ("FBuf") in F10 Output Panel, this keeps a 4x32 bits buffer
after render. Using PostProcess menu you will note an immediate re-
display of image too (32 bits RGBA)
- Added "Hue" and "Saturation" sliders to PostProcess options
- Render module is still not having a "nice" API, but amount of dependencies
went down a lot. Next todo: remove abusive "previewrender" code.
The last main global in Render (struct Render) now can be re-used for fully
controlling a render, to allow multiple "instances" of render to open.
- Renderwindow now displays a smal bar on top with the stats, and keeps the
stats after render too. Including "spare" page support.
Not only easier visible that way, but also to remove the awkward code that
was drawing stats in the Info header (extreme slow on some ATIs too)
- Cleaned up blendef.h and BKE_utildefines.h, these two had overlapping
defines.
- I might have forgotten stuff... and will write a nice doc on the architecture!
Dxf reader had 2 crucial erros;
- didnt read files correct with \r\n enters (\n and \r go fine)
- polyline triangles were added as quads with identical first and last
vertex, going terrible wrong on mesh import of 3DS files.
With the new rule that allows correct calculation of vertex normals on a
mixed solid/smooth mesh, it is essential that vertex normals get
recalculated when changing smooth settings. Such a facility doesnt exist
in Blender yet, only after leaving editmode.
Hacking in a "enter editmode, leave editmode" event on the "Set Smooth"
button isn't nice... instead I've tweaked the calculus of vertexnormals
that it always sets them OK, apart from where they get mixed with solid
faces. Only in rare occasions this can still go "wrong" and needs a
TAB-TAB to fix. Will add that comment in release notes. For next release
we should definitely solve this smoothing bizz!
Backbuffer selection codes issue on OSX.
On some h/w configs, with "1000s of colors set", the drivers seem to round
colors up or down... whilst until now rounding always was just truncating
bits. This gives errors in writing color codes and reading it back.
The c code only changed for OSX. For other OS's and HW this test could be
done as well.
It appeared that the method as used in unified render for "render all sub-
pixels" isn't very well resistant to having large amounts of faces in a
single pixel. The bug file had about 16x70x2 faces per pixel... causing
tremendous slowdown and even wrong render.
I've disabled the option (was coded by Nzc in NaN days) and made it
sampling in the main render loop. Goes much faster, error free. Only loss
is in that it doesnt use superiour subsamples for gauss anymore. Here the
normal render performs slightly better. Not a real issue though. Hard to
notice.
Slightly altered rules for calculating vertexnormals. By only averaging
face normals from faces actually set 'smooth', the result looks much more
nice (for example on a cylinder with caps solid). Vertex normals not being
used by smooth faces are set to the face normal direction.
Shows both in editor as rendering.
This method is now unsupported. The original intent appears to
return the first BPoint of an Ipo curve. However, BPoint Ipo curves
are not implemented and the first point therefore never existed.
The segfault was from an unchecked input parameter.
Calling this method now always throws a NotImplemented exception.
Separate in Mesh editmode, selectmode 'face', and with a selection that
used to be in vertex mode a full select (like selecting the 2 opposing
faces in a cube) caused crash. There was still no good face-select
awareness there... tsk! :)
Actually a 1 liner fix, added displist re-make event for wave objects
being used to vertexparent objects to. Will of course be much better
solved with the dep graph!
- Add UV Sphere now aligns with view, as do all primitives
- Small drawing error in edges between blender areas (black lines were 2
pixels too short)
- Replaced the heuristics in opening sublevel menus, to be less frustrating.
It now keeps sublevel open while mouse moves within the triangle defined
by original mouse location (before moving) and the closest vertical edge
of the new sub menu. Works nice for toolbox and pulldowns.
Particles added to mesh being deformed with Armature crashes.
Bug introduced with 2.33, particle collisions... leon added some fancy
updates for actions and armature displists there, which are highly
disputable. I left it in, because I didn't see it was actually new.
Note: particles added while updating mesh with armatures is not supported,
it will shoot particles from original location.
What does work, is static particles.
frame-sliding (LMB) in action, nla, sound and ipo window went into very
tight while loop, causing cpu to choke. Added nice idle.
Bug fix, found self while testing:
- relative.blend, editmode
- delete part, exit editmode
- enter editmode, undo twice, exit editmode -> crash
The trick to copy 'old' locations of vertices to keys isnt undo-resistant
yet. This fix at least doesnt crash, but copying doesnt work when undo-ing
back to previous editmode session (indices just differ from current mesh...)
When editing Meshes with vertex keys (RVKs), the undo stack isn't
updated correctly, causing a loss of the situation when entering
editmode after changing active key, and doing undo immediate after.
This due to an 'optimizer' in the undo code that tried to prevent undos
pushed on identical situations (e.g. pressing TAB multiple times).
Decided to remove that convention, the annoyance of having possible
multiple undos in a row isn't worth solving it by potential losses.
Combination of:
- render movie
- with border set
- and "crop" set
Crashed.
Was due to feeding movie initialize code wrong frame sizes. Another oldie!
Also found that crop+border+parts doesnt work. This now is prevented with
warning and return.
When the render output directory doesn't exist (on windows) the BLI_recurdir_fileops() function
tries to create a directory with no name. Added check if a name is specified..
that could popup error() boxes. These dont work then, and need to be
disabled by keeping R.flag on R_RENDERING
This is a half fix for a report Rob H works on. Will wait for his test.
Hiding window (Apple-H) caused the modifier to hang when reopened.
Added 1 line that re-reads modifier on "Inputchange" event, which only
is when you switch focus from one app to another.
Might solve more issues reported with 'hanging' keys.
- Dupli-Frames objects didn't render correct with Motion Blur (they moved!)
- VertexKey on Bezier paths didn't animate the 'twist' rotation. This did
work for Nurbs paths.
Plus; removed Nabla slider from Stucci, it doesn't work for that texture.
- Stucci now accepts it too (for normals and color)
Note; stucci doesn't return a single value, like marble or wood, that's
still so, to disable it coloring the "Col" Map input by default.
- Magic texture now accepts Colorband
Sequence editor crash when you delete Scenes which were used as a strip.
Related to that, file reading code didn't restore pointers for strips
correctly when you save with open "Meta" strip.
First bug 8 year old. Second 1 month. :)
When the Quicktime movie output path lacks any path indication, the movie is created in the
executable directory.
Also new, when the Quicktime output movie isn't created for some reason, the rendering will
stop and give an error message (in the console).
Raw fix for the drive letter issue with Quicktime/Windows filenames.
When a drive letter is missing in a filename, it'll use the drive where the executable is.
Fix might be useful for Yafray too, because the extern char bprogname[]; doesn't return
a full path when Blender is started from the console.
Rotated edges (seemingly on quads only) were not drawn even if
'Draw All Edges' is active.
Finally tracked the bug down to a typo in flag checks! Eeugh, was much
head scratching on this one! :)
- code wasn't edge-select aware yet
- the "Median" button now allows to move a "median" correctly around, also
to 0 or 1 in a single step
- missing undo push here.
