- Static variables can be initialized with constants only.
- Removed bunifont.ttf.c from datafiles -- it's not actually a
data file. Unicode font loading stuff is not in blenkernel/font.c
- Allocate as much memory for unzipped data as it's needed.
Default read chunk is 512Kb.
- Fixed regression (or just a typo) in setting utf locale.
- Default locale set to en_US:en works fine now.
- Commented put Nepali language in user preferences -- it's
not supported by current droid font and imo it's better to
have nice font for languages we actually have translation for
rather than allowing to choose more languages in user preferences.
- Make gettext stuff draw-time. so switching between languages
can happens without restart now.
- Added option to translate visible interface (menus, buttons, labels)
and tooltips. Now it's possible to have english UI and localized tooltips.
- Clean-up sources, do not use gettext stuff for things which can be
collected with RNA.
- Fix issues with windows 64bit and ru_RU locale on my desktop
(it was codepage issue).
- Added operator "Get Messages" which generates new text block with
with all strings collected from RNA.
- Changed script for updating blender.pot so now it appends
messages collected from rna to automatically gathered messages.
To update .pot you have to re-generate messages.txt using "Get Messages"
operator and then run update_pot script.
- Clean up old translation stuff which wasn't used and most probably
wouldn't be used.
- Return back "International Fonts" option, so if it's disabled, no
gettext lookups happens on draw.
- Merged read_homefile function back. No need in splitting it.
TODO:
- Custom fonts and font size.
Current font isn't nice at least for russian locale, it's
difficult to read it.
- Put references to messages.txt so gettext can merge translation when
name/description of some property changes.
Use this to raise errors when assigning a string property fails even though the value to assign *is* a string.
Before:
TypeError: bpy_struct: item.attr= val: Object.name expected a string type, not str
After:
TypeError: bpy_struct: item.attr= val: Object.name error assigning string, UnicodeEncodeError('utf-8' codec can't encode character '\udce9' in position 23: surrogates not allowed)
- FindXML2 we had copied from another project was always running and not using cached value, rewrote based on template used for most of our other find modules which makes use of 'FindPackageHandleStandardArgs'
- mark statuc collada libs as advanced.
Volume pre-caching altered shared data simultaneously in multiple threads, causing invalid scattering results when "Asymmetry" value was used. The view vector is now passed as a function argument.
Problem was caused by CRLF line ending instead of LF line ending.
Re-generate pu_shader_material to use LF line ending and made gpu codegen
treat \r as space character.
* Fix for high quality upsampling which was wrong.
* Fix for doppler effects which were calculated wrong for scenes.
* Improved animation evaluation at the beginning and end of a scene.
- avoid needless context switching quad/tri, flat/smooth.
- dont call glNormal3vf() lighting is disabled.
gives ~2x speedup with a subdivided cube, but thats probably the best case, quad/tri smooth/flat mix will slow down a bit.
