- Clear the cached timer when the timer has fired.
- Simplify the save-timer-save cycle and just rely
on dnt_timer to do the save, to eliminate
redundant saves.
If the previously set timer is exactly the correct
value for the existing DNT, then don't update/reset
the node timer.
Not sure why this was the way it was before (it
seems like it might have been intentional) but it
caused measurable performance impact, didn't
seem right, and fixing it didn't seem to introduce
new problems in preliminary testing.
Instead of checking to make sure one of the objects
found at a location is the glowing entity (which is
very slow and has to search every object for each
light) just make sure that the entity still exists
(get_pos() does not return nil) and that it's still
within the same node position as before.
This fixes a huge lagspike that was caught on NCC
due to a bunch of dropped glowing lenses.
The problem was pointed out by Josh
(gitlab.com/krazy-j) in merge request !22.
Apparently MT is not very smart about marking
mapblocks dirty to send to clients based on
calling mt.set_node(), i.e. it will mark them
dirty presumptively even if you set the node to
the same value it had already been.
This behavior can be confirmed by registering
an ABM against a common node like grass and
setting action = minetest.set_node. This
causes every mapblock containing that node
to be invalidated every interval, causing a big
spike in the packets received each interval
that you can clearly see on the F5 graph.
Rather than just fixing it for the most easily
observed case (fire checks), add utlity functions
to check this for ALL node change situations,
and apply it more or less universally anywhere
that we are not certain that the node is being
changed and we don't need to worry about the
extra overhead cost of the check.
Note that we don't need a
nodecore.set_loud_check call, as set_loud was
only ever being used already in cases where
we were pretty sure we were actually changing
a node.
- Regen with LBM too.
- Allow different delay time for "just loaded" case.
- Use DNT auto-start for more things, replacing
regen ABM/LBM where possible.
Also tried to improve door ablation reliabiity:
- Use DNT autostart for door ablation.
- Faster DNT check so we don't need the ABM
neighbor scan anymore.
- On early ablation trigger, set DNT for right after
cooldown ends instead of letting DNT loop die,
removing one scenario that could cause
sponge squeezer stalls.
During processing of the flammable ignition ABM,
nodes could ignite, causing the new ember nodes
created by this ignition to trigger fire propagation
instantly during the same tick by nodes still yet
to be processed for ABM checks during the same
scan. This often caused large areas of flammable
nodes, like wooden plank skyblocks, to be ignited
instantly in a specific direction from any source
ignition. This led to some "cheap shots" where
players were ruined by fire far more rapidly than
they should be.
Instead of processing flammability checks in
sync, delay them until the next tick, so that we
find all flammables to check before any new
igniters are created. This should tame fire
spreading back down to normal levels.
Allow use of true to indicate no custom options. The check
was originally put in place to allow custom options, but
some mods may use it more simply like the craft check,
where it's meant to be a simple pass/fail.
- New near_unloaded API that supports custom
distance parameter, map bounds check, and
some optimizations.
- Add area automatic unload check support to ABM
API and add to applicable ABMs, replacing many
old manual checks.
Note that the explicit unload check is only strictly
necessary when something will happen due to
a node NOT being present in the area, e.g. something
igniting due to absent coolant, or soaking quantity
being reset due to missing sources. In simple cases
where the absence of a thing causes the ABM to do
nothing, then the standard check can work because
the default action is to do nothing anyway.
We were using action log level for a lot of
things because by default MT does not seem
to capture info logs to stderr. On "production"
servers though this makes too much noise and
makes it hard to find actual player actions.
Servers that want info logging will just have
to configure/compile it in.
It's possible multiple soaking ABMs are registered
against a single node, and if this happens, make sure
that the qty and time fields in the cache are kept
separate the same way they are in the node meta.
- Removes repeating patterns, creates a richer and
fuller sound experience
- Larger areas fill sound volume out better instead of
just repeating more loudly
- Spread network traffic across multiple ticks better
to reduce packet spikes
Each node position has a set delay, between 0 and 1
second, that's fixed for that pos but varies more or
less randomly between positions. This ensures that
sounds loop with the same frequency and alignment
for a given single node, but desync across nodes.
Instead of queueing inverted ABMs and then running
them out-of-band, run them inside the normal ABM
schedule time. This should:
- Prevent ABM mux lookup problems caused by nodes
changing after a delayed inverted ABM schedule and
the action actually running
- Allow the ABM scheduler to better account for the
time spent running inverted ABM actions, to ensure
that they properly count against the default 0.2s
ABM runtime budget and can't spike lag up higher.
Ensure that when meta is implicitly cleared by node replacement,
the cache doesn't keep an old copy and fast-forward through
soaking recipes if the old node is re-replaced later
This was triggering on all dirt nodes everywhere, so optimize the
most common code path, i.e. one where there is no water above
and the soaking data must be reset.
- Store next runtime instead of countdown, so metadata always has
one canonical representation for a given target time, allowing
us to more easily avoid rewriting the same data with different
representation.
- Tons of caching and duplicate write avoidance, reduced resetting
of node timers, and MT API calls in general.
- Unify logic for expanding nodenames/itemnames lists into indexes
of matching names, and comparing item names/defs to lists,
with group expansion and deferral.
- Add "autostart" option to DNTs which registers construction traps
and a patrol ABM automatically, use for some things that are a
good match.
- Note that some things are doing more complex logic in ABMs, such
as determining if a node is eligible, or custom timing logic;
leave these as-is for now.
- New fluid wandering API, mostly unified logic, only needs
some parameters, plus the "generation check".
- More sophisticated behavior seeking down-slopes beyond simple
immediate flowing node reach ("subtle slope" detection).
- More reliable timing using DNTs.
- Set firm movement rate of concrete, molten glass.
- Reduce molten glass generations, since it can now be more
reliably directed.
Soaking recipe fails will result in a lot of clearing
of old soaking values out of node meta, which is
very slow in the common case of e.g. dirt that's
not near water and thus should not be soaking in
the first place.
Skip node meta writes if the node meta already
contains the correct data, by doing a cached read
and check before writing, instead of blindly
writing.
