The two weeks is configurable. But the idea is somebody creates a game
with a random name but it's never joined. Eventually the name is used
again and the first to connect joins the existing game that will never
progress. I could just prune from the DB, but this is less work ongoing.
This fixes on gtk version being able to invite to an SMS game using
relayID. I don't know why not transmitting port is ok; maybe android
doesn't use the port or gets it some other way, e.g. it's hard-coded for
each variant.
I've long wanted common code and state that outlast a single game. The
smsproto stuff is a step in that direction, but this file will be a
place for anything new. Starting, perhaps, with a mapping of deviceID ->
addressInfo (e.g. BT name or SMS number of opponent device) that will
let me stop duplicating that stuff in every game's comms state. We'll
see. It's guarded by a compile-time flag so I can play with in on the
Linux version until ready.
Crash due to typo introduced by Weblate made me realize I wasn't running
the script that ensures translated strings have the same set of format
specifiers as the English -- which might not be strictly necessary, but
it's safer to enforce it for now. Moved a few remaining %s to %1$s style
while at it.
There was only one size of png for the net status arrows. Now there are
svgs that generate the whole range. Also set color programatically to
avoid having to maintain four.
Use low-level NFC, a combination of emulated card and reader mode, to
work around Google's removal of "beaming" support from Android 10. App
emulates a card by declaring support in its AndroidManifest. When a game
is open that has data to send, it goes periodically into read mode. If
two devices are touched while one is in read mode and the other isn't,
they handshake and open a connection that should last until they're
separated. The devices loop, sending messages back and forth with or
without data (as available.)
Cleans up the code a bit, adding type-safety where it probably wasn't
needed but also letting me iterate over all tables. Not being able to do
the latter had allowed db-replacement code (useful only for debugging,
probably) to get out of date.
I suspect it's really rare to fail to get the lock when the game's not
open, but hope that trying 100ms before giving up will make failure to
save rare enough to be ok.
