There are both some good changes and some terribly hacky changes in this
commit.
- Server uses a different struct than client now for windows, so
arbitrary compositor things like effect parameters can be added
without having to rebuild everything. That's good.
- Client window structs no longer attempt to track or use window
locations. That's good.
- Windows do not destroy in response to the destory command, but rather,
a special animation is triggered; the end state for this animation
calls the actually window destruction code. That's a hack.
- Animations are based on rendered frames, not on time. That's a hack.
* Dumb bounding box top_at logic replaced with select buffer
* Select buffer rendered through cairo with AA disabled
Using rectangles for window shapes - this should easily
be expandable to 1bpp bitmaps. Currently, the select buffer
is very inefficient, using twice the space it needs (plus,
it's double buffered, so in reality, 4x the space needed),
however, it's also very accurate and fast, and I like that.
* Window rotation is controlled through Ctrl+Shift+{z,x,c} where
z = rotate 1 degree left
x = rotate 1 degree right
c = reset rotation
* Input is remapped based on window rotation, so you *can* use the
draw app, and it is totally epic.
Kernel driver sends raw scancodes, compositor or terminal handle the
rest. Support for F* keys will be added soon, in the mean time userspace
applications can know about the status of modifiers (control, shift,
alt, and super).
Compositor actions have been changed to Alt+Left Click for move and
Alt+Middle Click for resize. I'll work on adding more mouse events once
I get them to be faster (they're annoyingly slow at the moment... I want
more accuracy and less latency).
Also add functionality to the terminal to toggle it on and off.
There's a small, but nice-to-have performance increase from disabling
alpha transparency in the terminal.
Also, the display server sets DISPLAY=... to the shmem endpoint for the
compositor, so if getenv("DISPLAY") returns a pointer, you can assume
you are in a graphical environment, instead of having to rely on things
like arguments. Good for those applications that want to be able to run
both full-screen and windowed.
The compositor itself still needs work, but the compositing engine
within now does full blitting and is faster than the old method.
Transparency is now supported properly, though telling the compositor to
use it on a window will degrade performance. One terminal is usually
okay, and everything runs faster than it did before; two terminals is
pushing it; three will make you very sad. The stacking logic has also
been updated. Presumably, alpha blitting for transparent windows could
be done with SIMD instructions and be extremely fast.
All graphics libraries have also been updated to (hopefully) work
properly with alpha bits.
* Finally bring syscall.h up to speed and include all syscalls in the
syscall module of the C library.
* Remove the third-party obfuscated C demos (we have nyancat, good
enough)
* Fix userspace apps to build without complaining about undeclared
strtok_r by disable __STRICT_ANSI__
* Fix .eh_frame by including the proper stuff with libgcc.
All graphics library commands now take a gfx_context_t pointer, which
points to a simple datastructure describing a rendering context (width,
height, depth, total size, front buffer, backbuffer; where backbuffer =
front buffer when not in double-buffering mode, thus we always render to
backbuffer except on a flip). This may have caused a minor speed
reduction, but I don't really care as it's far more important that we
support multiple graphics contexts.
TODO:
- Shared Memory Fonts library (there are a couple of apps that use these
so-called "shmem fonts" on their own; we need a dedicated library for
them)
- Break off "TTK" GUI toolkit into its own library. Since it's just a
callback-based button framework, this shouldn't be too hard right now.
Also, with the previous tick, I'll be able to put labels on controls
and start using text in more places.
