In user terms: the motion on screen of the selection follows the motion of the mouse pointer.
Gives some errors when the constraint plane is nearly perpendicular to the view port though.
Added a debug print function for 4D vectors to arithb.c
Optimised the 3D -> view projection functions in view.c (a bit).
It used to be a simple hack, scaling lens with 100, and moving the camera
to the back with an equivalent amount.
Because of the hack, making it 100% compatible with older files I could not
achieve (yet?). To help reminding users, I've added a print when reading
old files with Ortho cameras.
Full description of how it works can be found here;
http://www.blender3d.com/cms/Render_changes.515.0.html
Cleaned code somewhat (half the size!) and applied some changes:
- made it compatible with selection flags for new mesh editing (edge/face)
- renamed 'face angles' to 'edge angles' and made it display angles for
2 selected edges too
- removed the confusing convention that always drawed the info if one
of the vertices of edge/face/angle was selected. now it only displays on
minimum of 1 full edge or face selected.
- made it react to "zbuf occluded selection" option (for zbuffered text)
- made it also colorize text for white theme color (TH_TEXT)
On larger meshes this info still draws very confusing. The idea of displaying
the info on 1 vertex selected was probably to show values while grabbing,
although that didnt really work for angles. One idea could be, for transform,
that it draws all lenghts/angles/areas that get changed during transform
All in all, still useful addition tho!
added a casts to:
blender/blenlib/intern/BLI_memarena.c
gameengine/Rasterizer/RAS_TexVert.cpp
Removed unused vars from:
blender/python/api2_2x/vector.c
gameengine/Ketsji/KX_PyMath.cpp gameengine/Ketsji/KX_Scene.cpp
gameengine/Physics/Sumo/Fuzzics/src/SM_Object.cpp
gameengine/Rasterizer/RAS_FramingManager.cpp
gameengine/Rasterizer/RAS_Polygon.cpp
gameengine/Rasterizer/RAS_OpenGLRasterizer/RAS_VAOpenGLRasterizer.cpp
Kent
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!
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..
This was caused by a very primitive method of interpolating quaternions.
It was converting quats to mat and back to quat, and then just doing
a linear interpolation. That whilst quaternions are renowned for having
good interpolation possible.
I've experimented with 2 quaternion interpolation methods, and can only
get one to work correctly... the "official" version from Watt brothers
I can't get working, both are in arithb.c now.
Will arrange *close* review with experienced NLAers for it! But testing
here gives fully predictable results.
Also changed;
- added pointer check in drawaction
- changed puldown menu for correct hotkeys for move NLA strips up/down
- The function "convex()" in editmesh_lib() actually did not deliver
a proper test for convex at all. It was checking only if a quad could
be subdivided into 2 trias. Code for adding face (FKEY) used this
call in total confusing manner. That code was there in 1.40 already,
cannot find any clue what it was supposed todo... :)
Recoded convex() to deliver a proper test. FKEY will give warning on
attempt to make convex faces now.
- Added undo-free for editmode undo on file load
(well, not on OSX to be noticed :). Discovered thanks to adding filling
of curves in Solid display, and bugreport of intrr that ESC in grabbing
curve gives weird corruption.
- EditVlak -> EditFace
- variables called 'evl' -> 'efa'
- functions with 'vlak' in it now have 'face'
Just thought was nice starter for editmesh recode...
Checks for operating systems should be that, and not checks for hardware.
Linux/sparc is a valid combination, using sparc to check for Solaris thus
results in problems on non-Solaris.
http://www.loria.fr/~levy/Galleries/LSCM/index.htmlhttp://www.loria.fr/~levy/Papers/2002/s2002_lscm.pdf
Implementation Least Squares Conformal Maps parameterization, based on
chapter 2 of:
Bruno Levy, Sylvain Petitjean, Nicolas Ray, Jerome Maillot. Least Squares
Conformal Maps for Automatic Texture Atlas Generation. In Siggraph 2002,
July 2002.
Seams: Stored as a flag (ME_SEAM) in the new MEdge struct, these seams define
where a mesh will be cut when executing LSCM unwrapping. Seams can be marked
and cleared in Edit Mode. Ctrl+EKEY will pop up a menu allowing to Clear or Mark
the selected edges as seams.
Select Linked in Face Select Mode now only selects linked faces if no seams
separate them. So if seams are defined, this will now select the 'face group'
defined by the seams. Hotkey is still LKEY.
LSCM Unwrap: unwrap UV's by calculating a conformal mapping (preserving local
angles). Based on seams, the selected faces will be 'cut'. If multiple
'face groups' are selected, they will be unwrapped separately and packed in
the image rectangle in the UV Editor. Packing uses a simple and fast
algorithm, only designed to avoid having overlapping faces.
LSCM can be found in the Unwrap menu (UKEY), and the UV Calculation panel.
Pinning: UV's can be pinned in the UV Editor. When LSCM Unwrap is then
executed, these UV's will stay in place, allowing to tweak the solution.
PKEY and ALT+PKEY will respectively pin and unpin selected UV's.
Face Select Mode Drawing Changes:
- Draw Seams option to enable disable drawing of seams
- Draw Faces option to enable drawing of selected faces in transparent purple
- Draw Hidden Edges option to enable drawing of edges of hidden faces
- Draw Edges option to enable drawing of edges of visible faces
The colors for these seams, faces and edges are themeable.
With a click on the 'COL' buttons (the ones showing RGB) a menu pops up
with three colorpicking fields and a palette.
The fields are the three planar intersections of a HSV cube, each allowing
choosing in the field without the field changing.
The palette is 'modal' unfortunately (couldn't find a simple working other
method) where the button "paste to color" denotes the state that click in
palette copies to edited color, and "copy to palette" means the active
color is copied into the palette...
Todo:
- saving of palette in file
- decide whether ESC leaves without changes...
big, since a float can go to 10^-37. And, this value is still squared, so
a square root will not frustrate it.
Limit now is 10^-35, fixing disappearing faces in extreme small object
thats extreme close to a camera. (thanks OOPz!)
Mirror Object is assigned to Ctrl-M and I've add Ctrl-M to call Mirror Edit (on top of M alone) for practical reason. We should consider switching both to Ctrl-M for pratical reasons but I kept M for backward compatibility
Also added the menu and toolbox counterparts.
Minor addition to arithb.c: A function to print vectors to the console. Easier for debuging.
Now uses 'Application Data/Blender Foundation/Blender' instead of old
"Not A Number" dir.
- Updated windows installer to make this change transparent for the
users. It copies /.blender to the new location and displays a short
message to advise them of the change
(http://homepages.nildram.co.uk/~aphex/installer_msg.jpg).
- Installer also includes fix for opening blend files from explorer (patch provided by Valentin Ungureanu (vung) - thanks!)
Note to CVS users on Win2k/XP: Although blender will continue to work
without changes, you should ideally copy the /.blender dir to
<app data>/Blender Foundation/Blender for the sake of correctness :)
Basically this provides three new things:
1. Choice of a list of noise-base functions, which can be used by the
current Clouds, Marble, Wood textures as well.
2. Three new texture types: Musgrave, Voronoi and DistortedNoise
3. Python access to noise functions (not for render!)
All of this together makes Blender's builtin procedural textures a LOT
more powerful. Here again, a full webpage should be made to show off all
possibilities, and explain some of the more scientific names for settings.
A good read on Musgrave textures can be found here:
http://www.ypoart.com/Downloads/Musgrave.htm
About Voronoi:
http://www.ypoart.com/Downloads/Worley.htm
I can't find official DistortedNoise docs easily... maybe its something
Eeshlo created himself.
I've spent some time to change the patch Eeshlo provided. Worth noting:
- created main texture "Musgrave" with 5 sub choices (instead of 5 new
main textures)
- added for all new textures the option to scale (zoom in out)
- added patch in do_versions to initialize variables
I hope the Python team will check on the Noise.c API. And include in docs!
Removes the creation of a password table for Non Windows machines
and instead calls getpwuid Was a lot slower before,
on systems with many users.
fix provided by Ryan Hayward (rhayward)
Kent
who have set a %HOME% environment var now works again.
Patch provided by Bill Baxter, plus an additional fix for Win9x.
There is still work to be done with this after the 2.32 release...
The font vectordata is scaled on load, so the character size will fit between the default (1.0) linedist.
Warning: this might change the font size in older blend files. (read: breaks backward compatibility)
Code from Peter O'Gorman <ogorman@users.sourceforge.net which has
been credited in the source.
Basically the dlopen() and dlerror() calls have been recreated wrapping
the standard OSX methods.
Names used are osxdlopen() and osxdlerror(). So no naming conflict will
occur.
colors. This because of the pretty weird (ab)use of load & make editmesh...
For each added undo step, the load_editmesh was fed with an empty mesh
to assign data to, without knowledge of what was in the original mesh.
That way UV and color data got lost.
Solved it in 2 steps:
1. removing the ->tface pointer from EditVlak, and make TFace a builtin
struct inside EditVlak. This didnt cost much extra mem, since it already
stored UV and color. This enabled some pretty cleanup in editmesh.c as
well, storing tface pointers was cumbersome.
2. for each undo step, it then generates always a tface and mcol block to
link to the undo Mesh.
Even when it wasn't in the actual Mesh, at exit editmode the original
Mesh is used as reference anyway, and undo-meshes are freed correctly.
The enormous commit is because I had to change the BLI_editVert.h file, and
found it was included in about every file unnecessary. I removed it there.
ALso found out that subsurf has code ready (unfinished) to make UV coords for
the displaylist in EditMode as well, nice to know for later...
(the first save didn't store paths or font settings!)
This is something I broke back at 2.28a :)
I'm suprised there were no bug reports on this, but it's fixed now! ;)
Axis options for TrackTo
LockTrack
FollowPath
Auto creation of TrackTo constraint from Ctrl-T (old track still an option)
Auto creation of FollowPath when parenting to path (Normal parent still an option)
Backward compatibility stuff to convert the per object axis settings to per constraint when a Track constraint is present.
Function to convert old track to constraint (commented out)
Revamped the constraints interface with Matt's work from tuhopuu and the stuff we were discussing earlier.
--------------------
For coders:
unique_constraint_name and *new_constraint_data moved to the kernel (constraint.c)
new Projf function in arithb gives the projection of a vector on another vector
add_new_constraint now takes a constraint type (int) parameter
add_constraint_to_object(bConstraint *con, Object *ob) to link a constraint to an object
add_constraint_to_client(bConstraint *con) to link constraint to current client (object or bone)
add_influence_key_to_constraint (bConstraint *con) to (eventually) add a keyframe to the influence IPO of a constraint
