In Object mode, Solid drawing, selection was just always not what you
wanted, the rules for wireframe (cycling) can't be used then.
New rule is;
- it always selects the frontmost visible item (wires in solid too), unless
- frontmost is active already, then it picks the 2nd
- mouse didnt move 2 pixels while selecting, then it starts cycling
Works nice :)
For composting sky behind alpha, a gamma corrected alpha_under was used,
but Blender didnt make the gamma tables unless OSA was set.
Old bug!
Changed; while rendering with "Backbuf" that doesn't exist, blender still
renders, but without backfbuf now. It used to stop render and return, with
only a print in console... confusing.
Gaussian sampling/rendering now supported too!
Also corrected gamma corrected adding of colors, which gives better alpha
and blending with sky than normal render does. The latter I could check
once too...
Added more support for lasso select; it now does:
- UV FaceSelect
- Editmode curve, surface, lattice
Missing: MBall and Armature; these make use of opengl selection still.
Thanks to Brecht & Valgrind, found 2 cases for unitialized variables in
the render code. Both were for 2D texture input (Window and Sticky) which
didn't initalize a the third coordinate. Goes fine for 2D textures, but not
for 3d ones :)
of 'envmap', 'magic', and 'plugin' modes.
The stucci texture also is not exact match, since it cannot be fully
emulated in yafray because of implementation issues. It will work best
for low turbulence values (which is actually not taken into account
in the export code).
Also, since Blender's static noise is basically just direct random number
output, don't expect the exact same result when rendered in yafray, but in this
case that probably shouldn't be that much of a problem...
(needs yafray from cvs)
(btw, on a side note, I put this in the comments once when working on the noise
stuff for Blender, but noise is affected by the 'depth' parameter, and there
is no way to control this directly from the GUI, can only be done by temporarily
switching to 'clouds' for instance.)
Crash in this situation;
- one blender file with more scenes
- render image
- goto another Scene with no camera and larger output size for image
- render
Render buffer has to be freed then :)
- Ztransp material didn't raytrace at all (now just traces it entirely,
remember too set the transp depth for it)
- tramsparent material reflected wrong in mirror material, due to
specular being added without alpha.
- Cleaned up some code to improve raytrace speed some. The old conventions
from before the AA recode were still there, this allowed coherence for
octree traversal. Current AA doesn't allow this anymore.
Added is improved check for 'first hit' on shadow render, per lamp this
now is stored
All in all, render with ray trace improved about 10-15%.
Option to switch left/right mouse didn't work for floating panels. It then
still selected stuff behind the button.
Was due to using wrong variable for events.
- fgets() doesn't recognize DOS strings in unices, made my own fgets
- reading polyline didnt support 3d lines
- reading polyline didn't create edges
All in all, this code isn't very nice and badly structured... a code
review or test (yes sirdude!) could have been useful. :)
- Bug #1890. While render subsurf in editmode, orco coordinates were
incorrect (subsurf vertices in editmode have different ordering)
Bug was oldie.
- IRC reported by Stefano; oops (not outliner) didnt show lamp textures
nor material ipos correctly.
Found error in oops curved lines too (the ones starting at right side)
for our interactive 3d artists!
- AL+Z (potato) texture view now draws objects without Texture Faces
compliant with (new) default in engine. Meaning lit faces, using
Material RGB for diffuse and specular. It used to draw these as 'shaded'
which is quite useless.
Per definition, Potato mode is supposed to show what engine draws.
(Note; this only for Mesh objects)
- Added a couple of more calls to ensure Shaded drawmode updates on
changing RGB sliders, moving lamps to layers, or switch layers
This based on feedback from Mal, who'll also give it all good tests.
View3d windows have 'unlock' option and 'localview', which was also
supported on render (F12 with mouse in window) to render only what is
visible there.
This didn't work very nice, with even code in render module to read from
interface variable. Removed that from render, and coded the exception
where it belongs, in renderwin.c
Reporter found omission in editmode undo. When you convert objects from Font
to Curve and/or to Mesh, the stack wasn't cleared (same object pointer!)
giving crashes on undo-restore calls. Nice discovery. :)
Old code in displist.c caused crash on OGL "render this view" option for
3D windows. No idea why this exception was in displist.c... must be for
old engine or so?
Two accumulating errors, causing 'scanline' errors too, but now based on
using different filtering values for transparent shadows.
Was another 2 cases of unused variable render as well. :)
Nkey panel in IpoWindow, the "Xmin" and "Xmax" buttons had a much too
narrow range, causing weird results when trying to fill in larger values.
NOTE; the button range method in Blender is still totally insufficient...
This patch should make gettext return unicode, which is what freetype
expects (also explicitly set in the same file). It makes translations
that support non-ascii *and* non-unicode encodings, such as latin 1 for
western european languages, draw correctly. The problem only occurs for
platforms where those encodings are default, i.e. in Linux if
/etc/locale.gen looks like this:
sv_SE ISO-8859-15
If it were to read UTF-8 this patch doesn't change the behaviour at all.
This is probably the case for Mac OS X and perhaps Windows.
orco coordinates were assumed to be valid when pointers were non-null,
this seems to have been a wrong assumption, pointers were probably unitialized
(but only for some objects like surfaces?), so when the export code tried to
access them, it crashed.
Now the rendermaterial texco flag is tested instead, which probably is how
it should have been done in the first place...
On large changes of the Lattice resolution button, the undopush crashed.
Reason was that the push happened for buttons before the actual event for
buttons was executed.
Solved by creating new event UNDOPUSH that's being added to the queue by
by buttons now.
- Made button undo texts for number buttons more clear
- Added undo push for missing Add lamp/empty/lattice/camera
Static-animated particles didn't recalculate on render
when used for duplicators (did work for halos). Allows material/texture ipo
for animated motion.
Move to layer buttons, using SHIFT+hotkeys to set button values and use
mouse to press OK undid the last change, this due to active flag hanging
on that button. It now clears hotkey active flags on mouseclick first.
On appending in posemoding, Blender could crash.
Found out the new 'select appended' also sets the 'active base/object'.
This isn't OK on that level of the code, 'active' exists (and needs to be
set) on UI code level in src/, not in the middle of file reading function.
Also removed weird negative bitflag that enforced 'select appended' to
be default. If you want that, set the flag itself in do_versions or so.
In my opinion it is best saved in .B.blend instead. Another time...
Appending or Linking data from other file always cleared scene scriptlinks,
due to having 1 line of code just on wrong location... is bug since 2.25,
so finally something again I didn't recently break in working code :)
- Deformed mesh didnt give correct 'orco' anymore on render
- Mesh editmode, "Separate loose parts", crash on empty mesh (no verts)
- Global undo; buttons sending redraw events gave extra undo push called
"Make single user or local". That latter was wrong event handling for
REDRAW event, harmless, until now :)
First one is a real showstopper... bah!
Found out the conversion functions for indices <-> framebuffer only went
up to 18 bits, making selection in zbuffer-select (or vpaint/faceselect
even) not work on large datasets. Provided you have 24 bits display, of
course.
This limit was just dangling in the code since long ago... modern times eh!
with the zblur plugin for faster dof (or other seq.plugins that need the
zbuffer).
I don't think the conversion to a blender zbuffer value is quite correct,
but at least it does produce usable results for zblur without too much
differences with the blender render (at least not for the short tests that I
could do in this short time...)
- Nurbs objects didnt render ray-transp. Was caused by 1) wrong normal calc
and 2) quads with vertices on same locations. Solved it with nice code that
checks illegal quads and turns them into triangles
- Made 'physics engine' variable SUMO default
- Typo in toolbox fix; cyclic isn't F but C
- loopselect doesnt work proper with vertices outside window
- outliner crash on script select (potential showstopper...)
- make curve parent menu, press ESC, crashed blender
All minimal changes in code. no new development, promised!
