With the new Wii U upscaling code, 640x480 uses too much MEM1.
Honestly, maybe I should just overhaul how render targets work on the
Wii U. Surely there's a way to dynamically allocate only one colour
buffer at a time?
This commit changes which variables are static: the Mac (and
presumably the Linux) debug data tells you what variables are static,
by prefixing their names with double_underscores.
The variable names themselves also hint at this: global variables are
prefixed with 'g', and use upper-camelcase, while static variables
use whatever_you_call_this.
These were retrieved from the Mac port (v0.0.7), which kept these
functions.
Unfortunately, this port is missing local variable names, so we can't
use it to make InitBack accurate.
Storytime: Cucky's original SDL2 port work involved using SDL2's
threading API to emulate the original WinAPI threading.
I can't be assed with that stuff, so I used the same trick Cucky did
for the Wii port, and hooked Organya up to the SDL2 audio callback.
This actually opens up the possibility for perfectly-synchronised
Organya playback. By that I mean, instead of needing a super
low-latency audio callback, I can have the callback synchronise its
audio mixing with Organya itself. I haven't done it yet, I plan to
soon.
Has the same imperfections: if you regenerate the surfaces while a
text box is open (and while using Courier New I guess), the text will
regenerate with smaller spaces.
Also fixed the SDL_Window not being freed. This commit's a bit of a
blob, since I made a bunch of tweaks to Draw.cpp while adding the new
renderer. Don't worry though, I'll add the hardware accelerated code
back again soon. In fact, I've got an idea on how to make it even
faster, while still being able to survive render target losses.
Hopefully this software renderer will come in handy for the Wii U
port: its SDL2 port's hardware acceleration is broken, and the
SDL_Surface API is too slow.
Okay so WindowsWrapper.h now just includes Windows.h if it wants
non-portability. This meant I had to split the custom RECT struct
back to the original RECT and unknown nameless struct (one uses
left/right, while the other uses front/back).
Along with being more accurate to the original, this fixes switching out
of fullscreen causing textures to disappear. This was caused by DirectX
destroying textures that are marked as render-targets when a device-loss
occurs. SDL2 doesn't do anything to recover them, unlike regular
textures, so my solution is just to avoid render-target textures
entirely.
On the upside, this should make drawing text to surfaces much faster,
since we're not reading back bitmaps from the GPU.
