- Use the existing on_stack_change hook instead of creating new
on_stack_[un]fill hooks, to simplify util_stack.
- Automatically register bulk nodes, but only for nodes being
registered that request bulk optimization.
- Remove some code duplication and registration complexity.
- Hand-pick nodes for bulk storage optimization based on a survey
of actual usage from a dump of the NodeCore Community server
world map.
- Use a separate group for visinvs with hidden entities.
Players can use sneak and/or aux1 for far-zooming, so they can
far-zoom while treading water in an ocean to look at distant
islands, instead of using the aux1=climb down hack, which causes
all sorts of weird issues with other features that try to detect
movement controls
Pointable/"viscous" liquids are "solid" when present at the head
location, but not at the feet location, so that players can still
wade on the surface of these liquids.
If the liquid is "pointable" then it can block players from being
able to emergency-dig the container; in these cases, just push the
player out immedately. This should only apply to some liquids
like pumwater, molten glass, or concrete.
We can't check initial conditions because the legacy mod would
load after this one (to detect that it needs to be disabled), so
we have to provide an API entry point for it instead.
There's a "soft disable" that just hides the entities
but keeps the database up to date, and a "hard
disable" that clears the database, keeping only
minimal data, until existing transactions (taken
items, dumped inventories) are flushed to the
world.
I had wanted this to be a standard feature of the
game for a while but put off merging it because:
- I had hoped to find a better visual design for the
ghost player entities that matched the bandolier
in nc_player_wield better, and
- I had scalability concerns with large servers with
thousands of offline players.
However, in practice it seems like the scalability
issues are not as critical, and the visuals are
acceptable as they are for now.
In making this a standard game feature, it
will probably also make sense to add an option
to effectively disable parts of it (e.g. avoiding
creating the ents, or even avoiding filling the
database except where players' items are already
taken).
- Clean up group expansion logic
- Lanterns covered in lux will never net-discharge, even if they
are not getting enough power to charge, so that they can be
carried and installed as light fixtures, with a supply of
flux from above, and can be "paused" in a flux flow when not
in use without discharging.
This helps to bridge the latency gap between the dig completing
on the client side, and the server sending the result, which might
include adding dynamic lights from a now-handheld light emitter,
e.g. for torches or lanterns.
Relying on the dynamic light thing for lanterns when they are
placed on the ground causes unexpected behavior when it comes to
lens activation and the displacement of dynamic light outside of
the node space of the lantern itself. It also causes issues with
lanterns lost deep underwater, i.e. they don't emit light when
they effectively should.
When the player is moving forward, take autorun
speed into account for distance to lead, to better
counteract server latency and let players run at higher
speeds. When standing still, still extend the light in
the player's looking direction a little bit, to get more
favorable "rounding".
- Check for water in same node (carried by player) and
one node above (esp. for placed stack)
- Slightly optimize node group check
- Fix float rounding error, since meta floats are not
perfectly round-tripped, to reduce spurious rate changes
- Add debugging info to trace phase transitions on all
lantern items
- Correct initial charge setting (for /giveme'd items)
- Fix failure to detect immersion in water
- Reduce re-wield animations
Compute current charge based on past rate and
time, like a soaking AISM, but do NOT update the
totals in metadata unless the rate changed or
we crossed a charge level threshold. This means
that we should only see the re-wield anim when
the charge level changes, the item is brought
in or out of water, or moved around a reactor.
We are losing the randomness factor from the
soaking AISM, but it's probably fine to have a
more exact discharge time for this sort of item.
Every sound in the game (and all known mods for
it) should play either at a specific location, to a
specific player (e.g. chat beeps or notifications),
or bound to a specific physical object. Detect any
that might be played globally by accident, e.g.
like a bug that existed in NC ZipRunes for a while,
to aid in debugging.
TurkeyMcMac's JIT Profiler mod provides much
better performance data anyway, and these
ABM stats haven't pointed out any really
useful actionable issues for a while.
