i've disabled font face (font source) instance sharing - it will load one per
size again due to performance reasons. i need to tackle this with the ft2
guys and see if theres an acceptible solution.
i COULD shadow all the glyph and font metric data i use myself per size -
thats fine... EXCEPT for kerning - thats the thing i can't sanely (figure
out how to) shadow myself... if someone figures that out for me! be my guest!
:) let me know!
SVN revision: 8634
allows me to be able to virtualize he canvas co-ordinate system. right now
it's doubles. i can now move to floats, int's etc. with a recompile (and well
recompile all depending apps too). it's still ACTUALLY doubles, just all
typedef'ed now. i've also changed booleans to actual boolean types (not an
int), all code will keep working - but i'd highly suggest moving your code to
use these types if interacting with evas.
SVN revision: 7644
with texture upload. it does NOt check error returns anywhere from gl... this
may mean issues with LOTs of images, LARGE images etc. need to fix that later
SVN revision: 7432
speed of texture uploads. anyone want to help? i've tried many things... and
nothing semms to work. this is a major bottlneck for evas gl engine
performance (apart from text - which is simply a matter of finishing off
properly)
SVN revision: 7428
directories... as long as the fonts are ttf's it will work. you can use
x-style font definition. ie.
-*-helmet-medium-r-*-*-*-*-*-*-*-*-*-*
:)
(fonts.alias is also supported!). the old style is still supported too.
SVN revision: 6870
1) Added 3 prototypes to evas_test_main.h for setdown, scroll_setdown, and scroll_setup.
GCC prints a warning for first declaration without a proto, but other compilers (FORTE) error on it.
2) Added type to did_fps, currently a boolian var, which was declaired only static... typed it as int.
benr.
SVN revision: 6595