And corresponding tests

It looks like the ChromeOS WM is resizeing the window. :(

Maybe we can figure out what the Chromebook is doing...

Fixes cur/bar weirdness when in VGA80X25 mode changing the current mode

This is how the max files/age of files in directories is enforced.

If SBBSDIR is set, add $(SBBSDIR)/exec to rpath.
Always add the original output directory to rpath.

I'm not sure what the old LD_RUN_PATH stuff was trying to do, but
I think these are the bits that make it work.

Fixes error where toprow may end up past the bottom of the screen
when the screen size decreases.

Any change to any double-height bit anywhere on the screen can
potentially change the display of every character after it on the
screen.  If the bit changes, force a full redraw from vmem
(Luckily there's no graphics in Prestel mode).

Fixes ticket 181
Ugh.

Visible in the double height graphics and engineering test pages.

With the performance enhancements, ciolib didn't realize reveal
chars needed to be redrawn when reveal was toggled.  Use a FG colour
bit to indicate if it needs to be, and hack it up in same_cell()

On an aarch64 Chromebook running gcc 12.2.0, these would wrap in
weird ways causing vector fonts to be positioned incorrectly.

It managed to say that (25 - -7) * 4 / 3 == -168
This casting dance appears to resolve the issue.

Memory is free()ed after sem is posted, so this write goes into
lalaland.

Fixes various memory corruptions with RIP

ERROR-level messages with error number, upon failure and debug-level message
upon success.

To hopefully help Deon get to the bottom of his SBBSecho echomail notification
issue.

We can't just memset() it.

Everyone gets an icon!

There's six pixels at the bottom that aren't part of text, and three
pixels on the right.

Don't touch the vmem when those pixels are updated... vmem can't
update them.

Should fix assertion in Abdul's Armor in RIP mode reported by DigitalMan
on IRC.

This should speed up loading the winnners list (e.g. when the syncdata msgbase
has a lot of messages) and scores won't be lost if the msgbase is purged.
Issue reported by Deuce (nix.synchro.net).

Bump version to 3.10.

Apparently, whatever was making this not link before has been fixed,
and so this workaround was breaking it.

Some scripts (e.g. addfiles.js) depend on the previous (undocumented) behavior
for some methods, e.g. lfexpand(undefined) would return undefined.

So for global functions that are expected to return a modified version of the
value passed (usually a string), return null or undefined when passed null or
undefined. Backward compatibility is more important than consistency in this
case. :-(

Fixes assertion/crash in 91-column mode (used in LORD intro screens)

EGA 80x43 has an extra six pixels at the bottom that are not part
of the text area.  When scrolling the botmap buffer, be sure to not
update this area.

Also, fix some screen invalidation bits when setting pixels.

This should fix ticket 180.

A high number of RIP commands use setpixel(), and it was updated
to scan the whole vmem for changes on each call.  Now we just extract
the part that is relevant and only update the one cell if needed.

This also adds parameter validation to bitmap_setpixel() which can
prevent some memory corruption and crashes.

... before the more ambiguous base filename check.

e.g. if you run 'jsexec /path/to/test.js' and get separate set of settings
from running 'jsexec test.js' depending on which sections you have added to
your jsexec.ini file.

And tests to insure the expected behavior