Requested by oranges and inspired by the upcoming event. A new subsyetem, non-processing (for now), aimed at providing some toggle switches that can be flipped as a last ditch effort to save some CPU cycles by sacrificing some non-critical mechanics. Below you can see each individual toggle.
Screenshot of the admin panel:
image
Surely there are more opportunities for toggles I missed, but adding new ones is not very difficult at all.
Why It's Good For The Game
Better performance during extreme pop, I hope.
Changelog
cl
code: Introduces the Lag Switch subsystem for when a smoother experience is worth trading a few bells and whistles for. Performance enhancement measures can be togged by admins with the Show Lag Switches admin verb or enabled automatically at a pop amount set via config.
config: Added a new config var: number/auto_lag_switch_pop
Add a new simple station trait, that replaces everyone's ID with a
wallet containing said ID, as well as their starting money, and if
they're lucky, a little bonus surprise.
In addition, adds a force variable to station traits, which will
always enable them, rather than having to edit weight and probability
of rolling a trait of that sign (postive, negative).
Co-authored-by: Rohesie <rohesie@gmail.com>
Part of my document on duplicating verb panel functionality so at some point in the far, far future we may be rid of it. you can read more here.
hackmd.io/tJ5H3opRQiG1GCN1sP142w
The other thing the status panel needs to do is send time dilation to the tgui chat. Having messed around with tgui chat, I am very confident i'm going to need stylemistake or mothblocks help me out with that one, it's so beyond my level it isn't even funny.
I know this is a VERY long description for a very small change but expect more of this in the future
* Converts looping sounds from a list of play locations to just the one
* Updates all uses of looping sounds to match the new arg
* Adds an area based sound manager that hooks into looping sounds to drive the actual audio. I'll be using this to redo how weather effects handle sound
* Some structrual stuff to make everything else smoother
Timers now properly return the time left for client based timers
Weather sends global signals when it starts/stops
Looping sounds now use their timerid var for all their sound related timers, not just the main loop
* This is the painful part
Adds an area sound manager component, it handles the logic of moving into new areas potentially creating new
sound loops. We do some extra work to prevent stacking sound loops.
Adds an ash storm listener element that adds a tailored area sound manager to clients on the lavaland z level.
It's removed on logout.
Adds the ash_storm_sounds assoc list, a reference to this is passed into area sound managers, and it's modified
in a manner that doesn't break the reference in ash_storm (This is what I hate)
* Hooks ash storm listener into cliented mobs and possessed objects
* Documents the odd ref stuff, adds an ignore start var to looping sounds, fixes some errors and lint issues
* Applies kyler's review
banging
Co-authored-by: Kylerace <kylerlumpkin1@gmail.com>
* Cleans up some var names, reduces the amount of looping we do in some areas
* Makes the code compile, redoes the movement listener to be more general
* fuck
* We don't need to detach on del if we're just removing signals on detach
* Should? work
* if(direct) memes
Co-authored-by: Kylerace <kylerlumpkin1@gmail.com>
Redoes how logs are generated, calling profile() does not guarentee that the output logs will contain one of each entry, if the value is 0 it just doesn't return data for that entry. So we need to do some memes to make sure things are working properly
Removes a debug proc that I forgot about
cl
server: Added configs to disable laggy hard deletes once they lag the server too much.
admin: laggy hard deletes only output once per type path.
/cl
closes#58379
1 second was too long in the current lag market.
(postpone doesn't work here because of a bug i'll fix in another webedit in another pr since its actually a mc bug)
Unobvious problem spot
#define BUCKET_POS(timer) (((round((timer.timeToRun - SStimer.head_offset) / world.tick_lag)+1) % BUCKET_LEN)||BUCKET_LEN)
With tick_lag equal to 0.1, 0.25, 0.5, rounding of division is normal. But at other values it may be shifted either more or less due to the specifics of floating-point storage and processing. Numbers 0.1, 0.25, 0.5 have blank mantissa, unlike others which lead to numbers such as 245.0000004 when divided.
PS: tick_lag is rounded to the first two decimal places.
Fixes
Rewrote the circular doubly linked list to a regular doubly linked list, because it could cause timers to loop infinitely.
Fixed re-creation of a bucket if the timer does not have a callback.
Added debug stat indicator RST to MC SStimer.
Added optional ability to log dump all timers on crash.
Fixed subsystem logic when a bucket position is misplaced due to division and rounding inaccuracy. The system now captures such rounding errors and restores the correct timer position.
References
[RU] SS220 Paradise port process from TGstation and fixes:
ss220-space/Paradise#5ss220-space/Paradise#10ss220-space/Paradise#26ss220-space/Paradise#32ss220-space/Paradise#37
Contributors
@semoro: fixes
@Bizzonium: port
Changelog
cl Semoro and azizonkg
fix: Ported fixes of SStimer subsystem from RU SS220 Paradise
config: Added a new config var: flag/log_timers_on_bucket_reset
/cl
Refactors dbcore to work off a subsystem if executed async and limits the maximum amount of concurrent async queries to 25.
This has been tested locally on a mysql docker image and there were no crashes (as long as you didn't run it with debug extools) + data was getting recorded fine.
Why It's Good For The Game
May or may not resolve terry crashes, however, each query creates a new thread which takes up 2mb, preventing the game from using that 2mb. This can lead to ooms if they stack up, e.g. due to poor connectivity. This solves that issue.
maintainer note: this did not actually resolve the crashes, but has value anyway. Crashes were sidestepped fixed by finding out Large Address Awareness works
cl
refactor: Refactors dbcore.dm to possibly resolve the crashes that happen on Terry.
/cl
This makes it so during unit tests, adding a text based overlay to something will runtime if the icon does not have an icon state matching that text. I would do this during normal compiles as well but getting the icon states from an icon is surprisingly expensive.
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
Converts most spans into span procs. Mostly used regex for this and sorted out any compile time errors afterwards so there could be some bugs.
Was initially going to do defines, but ninja said to make it into a proc, and if there's any overhead, they can easily be changed to defines.
Makes it easier to control the formatting and prevents typos when creating spans as it'll runtime if you misspell instead of silently failing.
Reduces the code you need to write when writing spans, as you don't need to close the span as that's automatically handled by the proc.
(Note from Lemon: This should be converted to defines once we update the minimum version to 514. Didn't do it now because byond pain and such)
Adds new signal - COMSIG_MOB_EQUIPPED_ITEM. Signal sent at the same time COMSIG_ITEM_EQUIPPED is and allows something to Register for every time a mob equips an item. Replaced a process() in /datum/quirk/badback with signals relating to this. Holds a weakref to any backpack it has registered signals with for use in its own remove() proc.
Removes snowflake code in /datum/quirk/light_step - This quirk no longer uses GetComponent to directly modify a component. Instead, the same code has been shifted into the component itself, which now checks for TRAIT_LIGHT_STEP and mods the volume and range accordingly.
Refactors quirk code in general - Quirks no longer do major logic in New() and no longer qdel themselves in New(). You now fully instantiate a quirk assigned to a var, then call /datum/quirk/proc/add_to_holder(). Various scenarios that shouldn't happen now get handled in this proc and calling code can cleanup properly. Quirks "support" having no quirk_holder - Since their default state is no quirk_holder until the quirk is added to a holder, and similarly quirks can be removed from a holder as well. Destroying a quirk with a quirk_holder will remove it from the quirk_holder properly. qdeling a quirk_holder will clean up any quirks attached to them.
Rethinks processing quirks - Quirks no longer all process automatically. The new logic changes mean their previous need to process() just to check if their quirk_holder had been QDELETED so they didn't become runtime factories has been removed by the refactored code. The few quirks that still process require processing_quirk = TRUE which will start them processing when added to a quirk_holder and stop processing on removal. This means there should be some many hundred fewer quirks process()ing 24/7 every shift.
Subtypes item quirks - A number of quirks are designed to give items to the player. There's code duplication and varying implementation issues, so item quirks have their own subtype with a proc and some vars to handle this. Quirks will no longer fail to give items at all (some quirks actually explictly qdel'd the items they give if the user had no free hands or slots) and will drop items on the floor in the worst-case scenario. Players will always get the opportunity to see messages related to item quirks as these are in a code path that, when the quirk is first added to a new mob, will either output immediately if the mob has a client, or wait for the mob to have a client otherwise.
"Roundstart quirks" paradigm removed - Quirks now have a path to add unique effects that aren't replicated if the quirk is transferred from one mob to another - add_unique(). Item spawning and other similar one-shot logic is done here. This means that adding a quirk to a mob will trigger the one-time effects. Transferring it between mobs (for example, slimepeople changing bodies or swapping a golem shell) will not. roundstart_quirks var renamed to just quirks since it genuinely is just a list of quirks the mob has from any source - Whether roundstart, admin or transferred.
Family Heirloom quirk - Heirloom is now a weakref.
Nyctophobia quirk - No longer process()es, now Registers COMSIG_MOVABLE_MOVED. Every footstep in the dark will slow you back to walking.
Reality Dissociation Syndrome quirk - No longer snowflakes behaviour for mindbreaker toxin on process. Now mindbreaker toxin has the anti-hallucination functionality built into its already existing HAS_TRAIT(M, TRAIT_INSANITY) check in on_mob_life.
Tongue Tied quirk - Now uses .getorganslot(ORGAN_SLOT_TONGUE) to find the tongue instead of locate() in internal_organs
Obsessed antag - Thanks to the Family Heirloom quirk now holding a weakref to the heirloom item itself, this antag type will no longer get the steal heirloom objective if the heirloom doesn't exist (ie. the weakref is null or fails to resolve, meaning the item has been destroyed)
Various quirks that were impossible to remove before are now removable - Examples being light step (which no longer directly modifies the footstep component) and bad back.
In addition, adds some extra documentation overall and improves compliance with code requirements in a number of procs (but certainly not all).
Probably various other little changes here and there to make the above all mesh together.
credit to zewaka for the idea of using underlays
turns the lighting object movables that were unnecessary and increased maptick into a datum which then applies and removes an underlay in update(). also applies a lot of general lighting clean ups (mostly using as anything in loops and fixing single letter var names).
multiz is a little different by necessity, now only the bottom turf's lighting matters in the brightness of the top turf unlike master where the bottom turf's lighting object is hidden from the vis_contents of the top turf. there are still some kinks to iron out here though, since currently objects suspended in openspace (like tram platforms) look bad and glass floors look bad too
only thing i have left to do is make multiz work (well)
UPDATE: multiz now appears the same as far as i can tell, its possible there are other situations in which its different but datum mats work and it automatically updates if the turf below changes. now i just need to make the system less finnicky if at all possible (and possibly merge managed_turf_vis_content with managed_overlays maybe?)
new update: its basically equivalent to normal multiz as far as i can tell (visually at least, in the circumstances ive tested so far)
NEW NEW UPDATE: turfs no longer have the VIS_HIDE vis_flag and multiz works without stacking the lighting from the floor below! so this shouldnt have any overt drawbacks to master anymore
1 needless movable per tile is terrible for maptick. this is probably a larger improvement than my emissive blocker change in terms of maptick. im guessing we'd get around 0.6 average maptick per player after this where currently we get 0.85 or so
Edit: according to lemon, sybil reached 0.71 maptick per person when tm'd with this
if this is a big enough improvement i might finally be able to get rid of the Gone discord avatar
Sourced from #59118 and a cursed project I'll pr later, This pr contains a lot of harddel fixes for stuff that pops up after a player interacts with something. I'm not gonna list them all here because there's something like 60 130, check the commit log if you're curious
Oh and I moved ref tracking screaming to a separate define, and made some optimizations to the thing in general. I think that's it, this pr is a bit of a frankenstine
Removed a bunch of back reference lists that were either entirely unused, or contained references that could be found within the datums contained in other lists.
lighting corner datums now get deleted with unused.
light sources no longer track lighting corners where the appiled light rounds to 0.
Fix lighting on turfs that gained dynamic lighting mid round. lazy init corner datums.
these two are related. by decoupling corner datums from the turfs dynamic lighting state, we can use them to know what level of light a non-dynamic light turf should have once it gains dynamic light.
Also should free up some memory not storing these datums in maint. Corner datums only exist on a turf that has light cast upon it by the dynamic lighting system.
Lighting corners are now lazy inited and deleted. they should always (and only exist) if there is a light source shining on it within range (even if the turf has no dynamic lighting). This is needed to support turfs that become lighting enabled mid round. On the plus side, they will no longer be generated on full dark turfs.
* Adds in support for basic send maps profiling, I'd still like to get something that tracks with the round a bit
better, but we can worry about that later
* Adds a debug verb that displays the sendmaps profiling page, as basic as it is
* Define respect
* Patches over an error where profiles would persist through rounds, somehow. Adds the maptick stuff to performance logging, so we can look at how things change over a short period of time
As described in issue #56733, there is a possibility of a race condition in SSmapping.add_new_zlevel - either during Z level incrementing, or the proc itself, as it only expands the amount of registered z-levels after being finished.
While the first hopefully shouldn't happen, it can still manifest because of the CHECK_TICK within. A practical result is that two concurrent calls to this proc will both create a new Z-level, but actually report having made the same first one.
This is problematic for map_templates that will then load blindly to the reported Z-level when using load_new_z, overwriting each other. We sometimes encounter that end result on CM13 codebase because of a map load that can be triggered by an user topic - if two people click at the same time, it's not unlikely for the first call to sleep in CHECK_TICK and let the second run, causing double-loading.
Adds a safety check to DropLandAtRandomHallwayPoint() to send mobs to the arrivals shuttle if get_safe_random_station_turf() returns null. This prevents people from being dumped in the supply pod shipping lane.
Makes datum materials support GAGS icons without too much configuration needed. You just need to specify a base config and subtype configurations can be for specific materials. I used the cleric mace as a testing item due to it having 2 distinct regions to color, so it got migrated to this system.
Moving forward we should not be using the regular color system for datum mat items as you can get conflicts with other code adding colors whereas with this the sprite has the base colors baked in.
Also I made the gold mace shimmer, for fun.
* Beep beep gamemode removal coming through
* More cleaning
* More cleaning
* Clean up station_was_nuked
* Station nuke source
* Remove name
* More unused functions
* Remove antag rep, which was unused in dynamic
* Remove more antag rep stuff, which was UNUSED IN DYNAMIC
* Station goals outside of mode
* Remove setup_done
* Remove player lists
* Move some older procs
* Fix simulations
* Attempt to fix this stupid case sensitive bullshit
* Revert "Attempt to fix this stupid case sensitive bullshit"
This reverts commit 8693c02d67602543d7d4b92130a7a32cd973e5bf.
* ???
* Delete icons/Testing directory
* Move station_goals global definition
* Moves explorer adventures to the database.
* Fixes default images not showing up in playtest.
* Review changes.
* Please
* SPACING
* Makes this a bit atomic.
Co-authored-by: Aleksej Komarov <stylemistake@gmail.com>
Redoes how geese handle eating shit, it was fucking stupid and caused harddels, and while this method is technically slower in the best case, it's a fucking goose
Fixes action related harddels, I hate how they work but at least this way they won't hold refs.
Fixes the hierophont causing its beacon to harddel
Removes the M variable from megafauna actions, it was used like a typed owner and caused harddels, so I burned it
Fixes target and targets_from harddels, replaces all setters of target with LoseTarget and GiveTarget, which should help maintain behavior. I'm not sure if this breaks anything, but if it does we should fix the assumptions that code makes instead of reverting this change
Fixes more area_senstive_contents related harddels, we need to allow the mob to move before clearing out its list.
Fixes marked object harddels (I'm coming for you admin team)
Fixes a language based human harddel
Fixes managed overlay related harddels (This was just emissive blockers, but I think this is a good safety net to have. If we clear the overlay list we should clear this one as well)
Fixes bot core harddels, I hate the fact that this exists but it has no reason to know who its owner is
Adds a walk(src, 0) to simple_animal destroy, it's the best bang for the buck in terms of stopping spurious harddels. Walk related harddels aren't that expensive in the first place, since byond does the same thing I'm doing here, but this makes finding mob harddels easier, so let's go with it
I fixed another source of part harddels, I hate fullupgrade so much
Fixes all the sound loop harddels
* FINALLY IT IS DONE FUCK
* tip of the round into ooc
* adds cyopod chat messages to the info chat filter
* last one I promise probably, mob spawner messages to info filter
* holoparasite, round will start in x and keybinding conflict to respective filters
* removed motd from constants, made it infoplain surrounded instead
* changed info to infoplain
extends the tile reskinning functionality to iron, bronze, plastitanium, carpet and pod floors
makes a bit of tile code better
moves some paths around, like elevator shafts being plating instead of floor
adds rotating as a tile reskinning function available on chapel or side floors for example
lets players customize any rooms they would want much more than it is possible now, allowing for more creativity
* Makes tanks only process when needed, rather then doing it constantly
* Makes portable atmos machines only process when needed, makes adding and removing atmos machinery cheaper, makes the tank processing logic actually work properly
* Makes pipelines only react() when their mix changes, fixes a ton of misuses of update_parents that were causing about 10ms of load for atmos just from reconcile_air()
* Adds in a new reaction framework alongside the old one for profiling related reasons
* whoops
* Cleanup, removes the react profiling code, I've chosen a direction to go here
* Cleans up some code, adds comments describing how to interact with portable atmos machines/tanks and their
mixes, adds a blurb to Atmospherics.md about the topic as a whole, leaves a line of bread for someone to follow
if I get hit by a bus in 2 days (Knock on wood)
* Removes priority from reactions, moves priority groups to the defines file
* whoops
* Converts air_contents to _air_contents, replaces all the out of file things that used it with return_air()
* Replaces the canister air contents uses, converts it back to air_contents, I decided I didn't like the _
* Fixes
* Adds timezone support for regional holidays
Also adds timezones to Waitangi Day, ANZAC Day, US Independence Day and Bastille Day
* Makes holidays span all timezones by default
Changes April Fools, Halloween and Christmas to be the correct dates instead of a range, using three timezones to cover the correct range instead
Co-authored-by: Mothblocks <35135081+Mothblocks@users.noreply.github.com>
