* large refactor of machine/power code to cut down on processing time and wasted lists (#60317)
original pr here: #59789 (Closed because he didn't think it was good enough)
came back to this because i realized that
all machines were area sensitive, meaning they had a list with at least a reference to themselves (assuming they arent in the contents of another movable which most arent) for the purposes of handling power differences when their area changes
pipes are machines
there are ~14k machines and ~6k pipes
i made this problem worse with a recent pr by making it a nested list
so i needed to track what machines needed power, and this pr had work already done that could be used for that purpose. now machines that have use_power == NO_POWER_USE do not have this extra memory overhead for no reason
currently every machine that uses power draws that amount from its area from a dynamic channel via auto_use_power() which is called every SSmachines fire(), then in apc/process() the area's dynamic power draw is reset and the power is used. with static power its not calculated then reset every loop, its just taken from the grid. so now machines handle updating their static power usage from their current area (this doesnt touch power machines that require a wire connection). in order to allow this, use_power, idle_power_usage, and active_power_usage have setters to track state correctly and update the static power usage on the machines current area and handle area sensitivity.
also goes through a lot of heavy abusers of SSmachine processing time and tries to make it faster. makes airalarm/process() into a signal handler for COMSIG_TURF_EXPOSE since air alarms only need to process for changes.
Why It's Good For The Game
SSmachines isnt the heaviest hitter in terms of total cpu and certainly not in terms of overtime, but its not a lightweight. it frequently takes > 50ms to complete a run and seems to be in the top 5 or so of subsystem costs looking at some round profilers
also gets rid of a few thousand lists since every pipe no longer has two useless lists each (and any other machines that dont use power)
Love ya kyler
Co-authored-by: Rohesie <rohesie@ gmail.com>
* large refactor of machine/power code to cut down on processing time and wasted lists
Co-authored-by: Kylerace <kylerlumpkin1@gmail.com>
Co-authored-by: Rohesie <rohesie@ gmail.com>
* small refactor to can_interact() so that borg range is fully respected (#60693)
Its a relatively small refactor that changes the previous machinery "can_interact()" proc that literally did a full override despite half of their checks already existing in not one, but TWO parent procs, so i removed the redundant checks, added callbacks to its parents and then added the cyborg range check on the can_interact_with() itself. in doing so i also moved the interaction range var from silicons only, to mobs as a whole and defaulted it to a single tile, silicons override it to 7 (so pAIs and borgs like before) but then set AI and AI.eye to "null", because i have a check in can_interact that if there is no range set, then the range is effectively unlimited. and i even added code for when AI is carded and their wireless transmission is disabled it sets their range to "0" aka, it has no range to do anything even if it could
this was really complicated for me so despite my extensive testing it probably would be a bad thing if any of you want to test my code yourself to ensure there isnt a bug with this (theres no runtimes ive come across)
note: i did a lot of searching and going through machinery to ensure i caught all the little snowflake overrides and added can_interact() checks to them, but i may have missed one or two things, especially maybe a altclick or ctrlclick somewhere, however i believe i caught most of them
one nice side effect of this refactor is that you can actually set another mobs range to something other than 1 tile and they can interact at range, rather than only silicons getting this ability, an admin could VV a human to have a 3 tile arm reach as a meme if they want
* small refactor to can_interact() so that borg range is fully respected
Co-authored-by: 小月猫 <alina.r.starkova@gmail.com>
* Del The World: Unit testing for hard deletes (#59612)
Co-authored-by: SteelSlayer <42044220+SteelSlayer@ users.noreply.github.com>
* Del The World: Unit testing for hard deletes
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
Co-authored-by: SteelSlayer <42044220+SteelSlayer@ users.noreply.github.com>
* Refactors species stomach code, makes ethereal charge inherent to their stomach (#59801)
A stomach would literally just ask the species datum how to be hungry and made a bunch of exceptions for ethereals. I mostly just moved code around so that hunger is handled by a persons stomach, thus shrinking the species.dm bloat.
This also makes it so ethereal's charge mechanic is unique to their stomache, instead of this really weird inbetween where just nothing happens and you kinda starve to death. You can stick an ethereal stomach in a human and this will replace their hunger mechanic with the charge mechanic and lets them absorb power from APCs and everything. You can also give ethereals a normal stomach and they'll use the normal hunger mechanics
* Refactors species stomach code, makes ethereal charge inherent to their stomach
Co-authored-by: Time-Green <timkoster1@hotmail.com>
* Converts the atmos senstive component over to connect_loc (#58266)
* Makes all uses of atmos_senstive pass in mapload as context
* Converts atmos senstive to connect_loc, does some general cleanup to the element, and makes it check the state of the tile the thing is on assuming creation didn't happen as a part of map loading
* Updates connect loc to match the new arg list
* Converts the atmos senstive component over to connect_loc
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
* Emissive system refactor (#58130)
The emissives system is the system that both lets computer screens and vendors glow in the dark and lets mobs and items block those glows. The current implementation relies on using filters to let mobs and items mask out the glow in the dark overlays on some structures. This is difficult to extend properly without massively increasing maptick. This PR changes the emissives system to use BYONDs native image layering to mask emissive overlays. This should prove to be a more extensible option.
tldr; There exists a system that lets computer screens glow on the dark and lets mobs and items block the glow. It isn't very extensible and this PR attempts to make it more extensible.
How emissive overlays used to work
Currently emissive overlays and the emissive blockers that mask those overlays are handled using a system of inter-masking planes. The emissive overlays and the emissive blockers are placed on separate, hidden plane masters. These are essentially rendering layers and groups. The emissive blocker plane is then used to mask the emissive overlay plane which effectively allows the emissive blockers to block the emissive overlays from being seen. After is has been masked the emissive overlay plane is used to mask the lighting plane, essentially creating holes in the shadows wherever an unblocked glowing thing exists.
Why this is a potential problem
This system works fine. In fact it works great! The computer screens glow, any person or item that winds up on a computer blocks the glow, and everything just works. However, this system runs into some issues when you try to extend it to work on things other than structures. Namely, the current system only supports emissive overlays on structures and emissive overlays that are completely unblockable by any means. As a result, several interesting uses to the system require extending the system.
As a result, if you want to apply emissive overlays to items (which exist between structures and mobs) or emissive overlays to turfs (which exist below structures) you must extend the emissives system to get the emissive overlays and emissive blockers to properly function. Doing this naively, by adding extra emissive overlay and emissive blocker planes and applying all of the relevant masking filters, is not exactly performant.
Maptick is a major contributor to lag and the higher the maptick the more free lag you, the player, get delivered fresh to your client. Trying the naive method resulted in #55782 (1f1b58bb26), an attempt to add glowing carpet to the game. Since the PR revolved around adding glowing carpet it had to extend the emissives system to allow for emissive turfs and emissive blocking structures. Extending the system was done naively as described above and you can see the results. 1.5 times the maptick across the board. Ouch.
So, we know that extending the system in it's current form is impractical. At least if done naively. Thus we are stuck.
tldr; The emissive system currently uses inter-plane masking to allow for emissive blockers to function. This is difficult to reasonably extend without murdering maptick. See #56496 (1f1b58bb26) for the results of naively extending this system.
How emissive overlays are going to work
Alright, so we know that the current system of using planes to let the emissive blockers mask the emissive overlays is difficult to extend in it's current form. The solution is to change how the system works so that it can be extended in a more efficient manner. What we want is a system that allows one set of images to be out masked by another set of images and for the first set of images to be capable of masking the light plane. Preferably, we would also like the ability to interleave the masking effect between emissives and emissive blockers with almost arbitrary layering.
Conveniently, this layering and masking is something BYOND already does to normal items and objects. If we put the emissive overlays and the emissive blockers on the same plane we can use their layers to interleave them almost arbitrarily like any normal structures and items! All we need is a way to mask away the emissive blockers from the resulting rendered plane and we can mask the lighting plane with the remaining emissive overlays.
Luckily, BYOND has provided a single filter that is capable of this task. The color matrix filter. This filter can be used to apply a color matrix to an image! Provided that the emissive overlays and the emissive blockers are different colors we can use a color matrix filter to effectively mask out the emissive blockers from the plane! The resulting emissive plane can be applied as an alpha mask to the lighting plane as it used to, to the same effect. The best part is, we get layering practically for free!
This is exactly what this PR does. It converts the emissives system from the old plane and masking based blocking to a new layer-based system which uses BYONDs native layer handling to mask the emissive overlays.
* Emissive system refactor
Co-authored-by: TemporalOroboros <TemporalOroboros@gmail.com>
* changes most uses of SSvis_overlays.add_vis_overlay() inside of update_overlays() to true overlays (#57985)
* changes most uses of SSvis_overlays.add_vis_overlay() inside of update_overlays() to true overlays
Co-authored-by: Kylerace <kylerlumpkin1@gmail.com>
* Fixes ethereals being unable to channel power into APCs, readds some defines (#57417)
* Fixes ethereals being unable to channel power into APCs, readds some defines
Co-authored-by: wesoda25 <37246588+wesoda25@users.noreply.github.com>
* renames metal sheets to iron sheets
* Merge branch 'master' into upstream-merge-56643
* yay metal
* Update DeltaStation2_skyrat.dmm
* a
* Update CentCom_skyrat_z2.dmm
* a
* a
Co-authored-by: Fikou <piotrbryla@onet.pl>
Co-authored-by: Gandalf2k15 <jzo123@hotmail.com>
* reverts Adds Neon Carpet (#55782), it appears to have been causing massive amounts of maptick (#56496)
It appears to have been causing massive amounts of maptick, this'll need a testmerge so we can test my hypothesis
* Reverts Adds Neon Carpet (#55782)
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
* Re-assesses 99% of vending prices through Arconomics to match player resources and round-length. (#54715)
* The Re-pricening
* Rewritten and adjusted for paycheck defines.
* I made the map changes finally.
* And the refills too.
* "OH YEAH REPLACING IT ALL WITH DEFINES AND SCALING IT THE EXCEL DOCUMENT WILL BE EASY, ARCANE!!!"
* And the premium ones too.
* Accidently spoiled a future pr due to dme bleedover
* Re-assesses 99% of vending prices through Arconomics to match player resources and round-length.
Co-authored-by: ArcaneMusic <41715314+ArcaneMusic@users.noreply.github.com>
* Converts A && A.B into A?.B (#54342)
Implements the ?. operator, replacing code like A && A.B with A?.B
BYOND Ref:
When reading A?.B, it's equivalent to A && A.B except that A is only evaluated once, even if it's a complex expression like a proc call.
* Converts all A && A.B into A?.B
Co-authored-by: ZeWaka <zewakagamer@gmail.com>
* Scales ethereal power capacity and interactions upwards (#53984)
Basically, all ethereal charge capacity and interactions have been scaled upwards by 20x.
Power wise, ethereals now hold up to 3000, with the max safe threshold being 2000. For reference, upgraded powercells can hold 2500, with high capacity holding 10K).
All appropriate values have been tweaked to match this change. There shouldn't be ANY sort of noticeable difference in game, aside from power sources depleting faster, and a few values which I decided to tweak for balance reasons. They are: power cell draining time is 1.5 seconds longer, and light draining time is 0.5 shorter. Also, draining cells has less of a punishment multiplier upon the cell now, (originally, the the cell lost 33x as much as you received. now its 12x). (to avoid returning to this in why its good, I did this because now that ethereals are capable of holding more, I can afford to have less of a punishment, while still maintaining a reasonable level of punishment).
Also some minor code and grammar improvements.
* Scales ethereal power capacity and interactions upwards
Co-authored-by: wesoda25 <37246588+wesoda25@users.noreply.github.com>
* Enforce preserving parent proc return values across ui_act call stacks (#53964)
All ui_act procs should call parent by default. All procs should preserve the value of the parent proc when it's TRUTHY and pass it down the call stack. No UI should be interactible when its flags or state indicate it should not be, except when explicity overriden by child procs intentionally disregarding parent return values to achieve a specific goal.
* Enforce preserving parent proc return values across ui_act call stacks
Co-authored-by: Timberpoes <silent_insomnia_pp@hotmail.co.uk>
* Replaces some numbers with defines + grammar stuff in ethereal power draining code (#53924)
* Replaces some numbers with defines + grammar stuff in ethereal power draining code
Co-authored-by: wesoda25 <37246588+wesoda25@users.noreply.github.com>
Replaces like 70-80% of 0 and such, as a side effect cleaned up a bunch of returns
Edit: Most left out ones are in mecha which should be done in mecha refactor already
Oh my look how clean it is
Co-authored-by: TiviPlus <TiviPlus>
Co-authored-by: Couls <coul422@gmail.com>
Co-authored-by: TiviPlus <57223640+TiviPlus@users.noreply.github.com>
Co-authored-by: Couls <coul422@gmail.com>
* Refactors machine_stat and is_processing() to process on demand (#53150)
* Refactors machine_stat and is_processing() to process on demand
* Fixes master (#53193)
* fixes master, damn it me/mothblocks
* fuck again
* ANOTHER ONE
* Fixes master
Co-authored-by: Rohesie <rohesie@gmail.com>
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
* Color standardization, vars moved, and signals (#52574)
Defined all the existing light_color values.
Moved their definitions to colors.dm
Made white the default color. It was so already, but that was very obscured.
Moved the atom light-related variables to the atom definition.
Wrapped changes to variables such as light_color into procs that report the event through signals.
Moved the light_on variable to the atom level, also adding a signal for its changing, to represent toggling lights.
Cleaned up a little bit of code in where new variables were defined before redefinitions.
This is all atomization to reduce changes in #52413
None of this affect gameplay at all, it's all code cleaning and refactoring.
There's more colors to standardize, a search for color = will find lots of targets, and I see little need to have both the LIGHT_COLOR and COLOR patterns, but I don't want to make this PR bigger than it already is.
* Color standardization, vars moved, and signals
Co-authored-by: Rohesie <rohesie@gmail.com>
About The Pull Request
/area/proc/usage() attempts to give list-like access to a bunch of vars. Why not make it a list instead and avoid all the proc calls? Might be room for followup here, to do something to powered(), use_power() etc.
Some legacy machinery was ignoring the default machinery use_power pulling from the machine's power channel by default
Total power usage was unused, APCs ignored it in favor of calculating it themselves :)
I also renamed the defines because they were in the danger zone of being very common words.
Changelog
cl Naksu
code: optimized area power usage calculations.
/cl
