warning: passing argument 1 of ‘ra’ discards ‘const’ qualifier from
pointer target type

1) Unreachable return without a lock in sdl_useR_func_ret() (harmless)
2) LOR in handling SDL_USEREVENT
   - vstatlock was obtained inside of win_mutex which should never
     be done... this could (and likely does) result in a deadlock
     when using SDL mode.

Update the script to verify pixel operations are supported.

You can copy whole, or portions of a PPM file from the cache directory
to the screen, optionally applying a mask.  The mask can be specified
as a base64-encoded bitmap or as a PBM file in the cache directory.

You can also copy portions of the screen into an internal buffer, and
draw from that internal buffer.  There's plans to support loading an
image file from the cache directory into the buffer as well.

Portions of the buffer can be copied to the screen, optionally
applying a mask as with drawing PPMs.

There's also plans to support copying a PBM file into a mask buffer.

I've added the syncterm-bounce.js script which simply bounces the
SyncTERM icon around the screen like the classic screensaver we all
hate.  I initially wanted to replace the spinning cursor with it,
but there doesn't seem to be a way to replace that with a script. :(

To use the script, just add the text:
@EXEC:syncterm-bounce.js@ to the end of your logon.js script.

Log in to nix.synchro.net with the latest nightly build or build from
git to see the script in action.

Not actually part of CTerm, but documented here anyway.

This destroys graphics.  Fixes SF bug 94.

sdl_setwinsize() and eliminate another race.

Pass window size in SDL_USEREVENT_SETVIDMODE
- Fixes race condition where SDL_WINDOWEVENT_RESIZED or
  SDL_WINDOWEVENT_SIZE_CHANGED arraves while
  SDL_USEREVENT_SETVIDMODE is in the queue

Set new scaling based on old scaling multiplier
- Will resize window on mode changes, but tries to keep integer
  multiplier similar.  May break fullscreen modes, dunno.

Have aspect_fix() return higher rather than lower
- This may make it the same as aspect_correct(), but I'm too lazy
  to figure that out.

Add additional stderr debugging.
Don't try to correct the aspect ratio of the window.

This is the issue behind CVE-2022-24130 in xterm.
Reported against SyncTERM by Autumn Lamonte (Thanks!)
Discussion here: https://gitlab.com/klamonte/jexer/-/issues/105

Closes SyncTERM bug 79.

Reported by @anonymouspage, closes !211
Clean up  formatting in the touched functions while we're here.

New Medium impact and a low impact issue still needs to be triaged.

e.g. directive argument in the range [-255, 2147483645]

Not sure why this isn't defined from inttypes.h, but blah.

I'm not really happy about rounding floats in time-critical bits,
but there you go.

The previous tests to detect RGB8 appear to have been insufficient.
This may have resulted in various artifacts such as transparency with
some drivers on some hardware in X11 mode.

For programs that don't specify a scaling type (scfg, syncdraw)
use blocky scaling.

He's old, and should keep his blood pressure down.
Do not walk on his lawn.

This is basically just a scratchpad for things I want to have in
the future ciolib.

I was going to start keeping the old rect around again, but that
idea didn't pan out.  This massive memory leak is what I get for
not doing more commits while tuning.

Pixes set via setpixels() were only set in the "blink off" screen
and not in the "blink on" screen... so they all reverse blinked.

1) Keep a rectangle updated per-screen rather than regenerate each time
2) Strip palette info when putting pixels into rectangles rather than
   during scaling
3) Tighten up the screen locks a bit
4) Don't require a full resend of both screens on an update request
5) Only force a redraw for cursor movement when the cursor is visible
   (And force it whenever the cursor changes)
6) Avoid doubles in interpolation
7) Heavily optimize interpolate_height()
   interpolate_width() likely doesn't need it because it's generally not
   used and also it reads from the next pixel in memory making the
   prefetchers job easier.
8) Fix some memory-leak-on-error issues
9) For ARGB8 XImages, manipulate the data directly rather than through
   XPutPixel()

At this point, the scaling and X11 output time is heavily dominated by
cache misses.  The only really effective way to reduce this hit is to
spread the work across all the L3 caches in the system or move it into
the GPU.

With the latest updates, at the SyncTERM menu, over 90% of the time is
spent in the rendering pipeline, and over 90% of that time is spent
thrashing the caches... the only real easy win left is vectorizing, but
that's highly compiler specific.

To that end, I've switched to -O3 for release builds.  There was a comment
that -finline-functions broke Baja "badly", but that's clearly false since
-f-inline-functions has been part of -O2 for quite a while now, and Baja
doesn't seem any more broken that it ever was.