## About The Pull Request
I heard you liked commas.
## Why It's Good For The Game
Trailing commas make everyone happy!
## Changelog
Nothing anyone will notice besides downstreams adding to these lists
## About The Pull Request
As the title says.
`init_order` is no more, subsystems ordering now depends on their
declared dependencies.
Subsystems can now declare which other subsystems need to init before
them using a list and the subsystem's typepath
I.e.
```dm
dependencies = list(
/datum/controller/subsystem/atoms,
/datum/controller/subsystem/mapping
)
```
The reverse can also be done, if a subsystem must initialize after your
own:
```dm
dependents = list(
/datum/controller/subsystem/atoms
)
```
Cyclical dependencies are not allowed and will throw an error on
initialization if one is found.
There's also a debug tool to visualize the dependency graph, although
it's a bit basic:
Subsystem load ordering can still be controlled using `init_stage`, some
subsystems use this in cases where they must initialize first or last
regardless of dependencies. An error will be thrown if a subsystem has
an `init_stage` before one of their dependencies.
## Why It's Good For The Game
Makes dealing with subsystem dependencies easier, and reduces the chance
of making a dependency error when needing to shift around subsystem
inits.
## Changelog
🆑
refactor: Refactored subsystem initialization
/🆑
* Fixes smoothing breaking if a map is loaded post init (#81526)
## About The Pull Request
We'd finish a set of atom creation, then try and smooth those atoms The
problem is they might try and smooth with an uninitialized neighbor,
which wouldn't have its smoothing vars parsed.
This fixes that by pooling "to be smoothed" things into a list based off
the source of the init stoppage, which we then release when we're done.
Also fixes things staying in mapload, even during a sleep. This can
cause massive headaches so it's good to avoid.
This has a cost but it's minuscule (on the order of like 0.006s (6ms
over all of init), so I'm happy with it.
## Why It's Good For The Game
Closes#77040
## Changelog
🆑
fix: Maps loaded in after roundstart will no longer have broken
smoothing
/🆑
* Fixes smoothing breaking if a map is loaded post init
---------
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
* Fix very common smoothing groups runtime by deferring icon smoothing until all atom initializations are complete (#75444)
## About The Pull Request
Fixes the "bad index" runtime that shows up on 80% of rounds.
The issue is that during dynamic maploading, the initialization of
objects is split over several ticks. However, before the initialization
is complete, an object's smoothing groups is still in its string form
(as per [the smoothing group
optimizations](https://github.com/tgstation/tgstation/pull/71989)).
Thus, code later down the line attempts to read from this string, and
errors.
It would be expensive to check all nearby neighbors every time in any
place in this code, so instead we defer icon smoothing if there are any
initializations currently in progress. Because this only happens for
dynamic loading, the stuff being loaded dynamically is usually small,
and icon smoothing already has a somewhat visible delay for shuttles,
this delay doesn't matter.
## Changelog
🆑
fix: Fixed an issue where objects on something that loaded dynamically,
like shuttles, would not be smoothed.
/🆑
* Fix very common smoothing groups runtime by deferring icon smoothing until all atom initializations are complete
---------
Co-authored-by: Mothblocks <35135081+Mothblocks@users.noreply.github.com>
* Fixes typo in `FIRE_PRIORITY` defines (#74694)
IT'S `PRIORITY` NOT `PRIOTITY` WHAT THE FUCK IS A `PRIOTITY`.
FYI nothing was broken since both the subsystem and the define had the
same name so there wasn't any bullshittery going on there - just a bit
of silliness that I noticed.
* Fixes typo in `FIRE_PRIORITY` defines
---------
Co-authored-by: san7890 <the@san7890.com>
* Micro-optimize icon smoothing subsystem, cutting down more than 50% of its init time (about .8s drop on local, more on prod) (#69741)
With atoms now being 50% of the init time instead of ~80%, I've decided to poke around other subsystems.
Unfortunately they're not straight forward at all and extremely boring. bitmask_smooth() in particular is just a big proc that is called ~60k times, so I made some adjustments to its macro to reduce ops.
I removed code that looped over two entire Z-levels to find things to smooth, added by e2dd404, at a time before the initialize system. This doesn't have huge gains since it just defers to the queue anyway, but it still shaved off 0.1-0.2s on local.
I removed update_appearance(~UPDATE_SMOOTHING) from something since the person who added it doesn't remember why and it doesn't appear to do anything. Lemon thinks it's something to do with emissives, but I tested with neon carpet and it seemed to work fine.
* Micro-optimize icon smoothing subsystem, cutting down more than 50% of its init time (about .8s drop on local, more on prod)
Co-authored-by: Mothblocks <35135081+Mothblocks@users.noreply.github.com>
* Refactors firestacks into status effects (#66573)
This PR refactors firestacks into two status effects: fire_stacks, which behave like normal firestacks you have right now, and wet_stacks, which are your negative fire stacks right now. This allows for custom fires with custom behaviors and icons to be made.
Some fire related is moved away from species(what the fuck was it even doing there) into these as well.
Oh and I fixed the bug where monkeys on fire had a human fire overlay, why wasn't this fixed already, it's like ancient.
Also changed some related proc names to be snake_case like everything should be.
This allows for custom fire types with custom behaviours, like freezing freon fire or radioactive tritium fire. Removing vars from living and moving them to status effects for modularity is also good.
Nothing to argue about since there's nothing player-facing
* Hud Image Culling By Z Level: Theft edition (#65189)
* makes hud images only apply by z level
* makes some of the atom_hud procs have better names
* fixes warning with the hud_user list and adds better documentation
* better docs for hud_images
* removes TODOs
* docs for hud_list
* adds support for linked z levels so mobs can see lower ones
* fixes merge conflict and shittily makes only shocked airlocks get added
* adds support for setting images in the hud as active and inactive
* gets rid of unatomic spatial grid change
* maybe i should actually try COMPILING my changes
* fixes merge skew and makes it compile again
* fixes huds refusing to remove from users who changed z level
* improves z level and registration logic
* fixes antag huds not appearing
* Fixes antag huds not properly setting. We now use hud_list in init, so it needs to be set before the new call, not after. Not sure why the use of appearance key was split like this, but none else knows either so none can stop me
* Ensures that hiding a basic appearance also hides the atom's active list too
* Fixes antag huds going poof
Ensures that remove_atom_from_hud will return false if the passed atom
isn't managed by it
This fixes antag huds disappearing randomly, since they assumed that if
the parent call of remove_atom_from_hud returned true, we should delete
ourselves. This is a safe assumption for them to make, since they should
only ever have one atom.
Does kinda bork if we call remove_atom_from_hud in a way that is unsure
if the passed atom is actually in that list. We were forced into doing
this by how atom huds use the qdeleting signal.
* makes basic alternate_appearance's only update themselves when setting their hud image to active and makes them not add themselves to the global huds_by_category list
* fixes mistake with hud_users list being set non associatively (bad)
* as anything in bot path loops
* Fixes merge skew problems
* Makes bot paths non global
This way they can show themselves to only the bot that "owns" them, ya
feel me?
* Fixes huds not showing up sometimes, cleans up some code
Post Kapu's limb refactor, we were calling prepare_huds twice in a human
init call chain. What was happening was this:
call prepare_huds() // Human
I gained a new hud image
I set active hud icons to mirror it
call prepare_huds() // Living
I overwrote the new hud image
I attempted to set active hud icons, which failed because it assumes
this can never happen
*cries*
* Renames add_hud_to_atom to show_to
My hope is this will make understanding hud code a bit easier, by tying
the behavior to a "verb" more closely. Also renamed a few vars
* remove_hud_from_mob -> hide_from
* Nitpicks a few comments
* Whoops/fuck/shit/damn it all/hhhhhhhhhhhh
* Moves check down, improves stack trace a bit
Co-authored-by: KylerAce <kylerlumpkin1@gmail.com>
* small touch-up
* this should do it
Co-authored-by: SmArtKar <44720187+SmArtKar@users.noreply.github.com>
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
Co-authored-by: KylerAce <kylerlumpkin1@gmail.com>
* EVEN MORE HARDDEL FIXES (#60228)
Fixes a ton of harddels, sourced from #59996
I think this brings us down to like, ~100 per round from ~200, with only like 20 of those being proper hell failures. I've seen harddel profiles below 1 second of total cost. Feeling good.
See you on the other side
Makes the cryopod control computer into a weakref, never trust bee code
Converts brig door timer internal lists to weakrefs
Fixes a harddel caused by qdeling a motion sensitive camera after it had left its source area, jesus christ why didn't we do this already holy shit
Converts the radio implant ref held by the antenna mutation to weakrefs because it isn't reliably cleaned up, makes the radio implant actually qdel its fucking radio
Removes the target var from the throwing datum, it does literally nothing and just exists to cause harddels, mostly for the singularity
Fixes a cable harddel sourced from things that try to enter blueprints after smoothing, but before roundstart. IE, shuttles. Removes shuttles from the blueprints
Fixes emmisive blockers being added post qdel
Removes some manual ghosting from cryopods, I initially did this for harddel reasons, but I figured out a better fix for that. I'm now doing it because it's got this really strange logic for like "re-entering the game" that doesn't actually link to what the ghostize proc does. We should remove this at some point
Fixes robot hud objects harddeling due to hanging refs
Fixes buildmode related hanging refs, I'm coming for you admin team
Fixes a few instances of trying to add the forensics component post qdel, hhhhhhhhhhh
Fixes some split personality harddels/weirdness
Replaces a use of disconnect_duct with an init qdel hint, I suspect there's more issues with duct harddels, I've seen some odd logs about ahhh the area_contents list, but we can worry about that later
Makes teleporter targets into weakrefs, properly types them as /atom
Makes frequency devices into weakrefs
Makes cameras remove themselves from camera nets on Destroy
Makes tgui ui datums implement destroy, this means if I ever see one hang a ref to user or whatever, I know there's an error with calling close() properly. I've seen this harddel once, but not after this change so I assume there was some error with close(). IDK maybe this is a papering over? Would have to ask @ stylemistake
I've seen logs of beartraps being in world post del, putting a return there just in case. The same is true of nerf darts, but I haven't really looked into that yet
Makes a shoe's ref to untying alerts a weakref, yes this is needed.
Moves clearing client_in_contents to the Login of the new mob. This prevents doing things like ghosting someone before a mob qdel causing harddels
Fixes a harddel set sourced from adding a status effect to a qdeleted thing. Is this an error? I'm honestly not sure.
Converts bsa code to weakrefs
Converts the partner var of heat exchangers to weakrefs
Converts camera assemblies to weakrefs
Fixes some dumb behavior with ammo casings and assuming you'll be on a turf post Destroy parent call
Fixes? merger related harddels, you were never cleared from your own members list, so origin objects would end up making a new list, creating harddels. Potential input from @ ninjanomnom about the logic
Chasms store a static list of "falling atoms", which only exists for chasms that go somewhere else. This list wasn't being cleared of qdeleted objects, which is what happens when you fall in most chasms. Fixes this, and converts the list to weakrefs.
Fixes some runtimes in both sheet code, and the weather listener element. This is here because runtime spam made testing more of a pain, didn't think it needed its own pr
Fixes colorful reagent harddels sourced from reagents that were qdel'd before roundstart. I'm only like 50% sure this actually got it, but the issue may have been solved by #60174, so eh
Turns the nuke op antag datum's ref to the war button into a weakref
Fixes some holopad code that was not nulling refs all the time
Converts camera bugs to weakrefs, this was the result of the bug being "reworked" like 6 years back without taking the existing ref clearing into account. Whole item needs a redo, but this'll do for now.
Ensures that the both pulling and pullee refs are cleared on Destroy
The crew monitor held all users in a non clearing list, makes that list a weakref because I hate everything
Oh and I removed all sources of gas_mixture qdeletion, I'm kinda unsure on this since it's not technically supported, but any harddels from it might? indicate something going wrong with like, gas passing logic. I'd like @ MrStonedOne's thoughts, since I trust him to call me an idiot if I'm wrong.
<!-- Please add a short description of why you think these changes would benefit the game. If you can't justify it in words, it might not be worth adding. -->
## Why it's not good for the game
I crashed sybil like 10 times to get this data, I'm gonna put it to good use. Don't think you're safe sybilites, I'm coming for you.
* EVEN MORE HARDDEL FIXES
Co-authored-by: LemonInTheDark <58055496+LemonInTheDark@users.noreply.github.com>
* Icon smooth refactor (#52864)
bitflag list construct added: an associative list of bitflags for cheap and quick element comparison between two lists using the same system.
canSmoothWith list turned into a bitflag list.
smoothing_groups list added to substitute the type path list.
smoothing procs turned into atom procs, refactored and optimized a bit.
smooth directions redefined in order to fit in 8 bits for a future smoothing system
some variable names changed, foreseeing a second smoothing system
SMOOTH_OBJ flag added, for things that need to scan turfs for smoothing. The old locate() optimization has the risk of returning false negatives by finding a child and returning null while there might be one of the wanted type as well, as it doesn't match the type exactly.
SMOOTH_TRUE and SMOOTH_MORE condensed into SMOOTH_CORNERS. The old behavior can be replicated using smoothing groups without loss.
Does very minor code cleanup.
Processing-wise didn't find a noticeable difference. The system loses on init a bit by setting the bitflag_lists, and by scanning whole turf contents for object smoothing (increasing accuracy), and gains by making less checks per target to smooth, through the same bitflag_lists.
Memory-wise there should be a small improvement, given that on the old system we had 63512 canSmoothWith lists (a few typelists, most unique), and on this new system canSmoothWith + smoothing_groups are both bitflag_lists from the same pool, totaling 46 in number.
Could be tested a bit to see if I missed any icons not properly smoothing.
* Icon smooth refactor
Co-authored-by: Rohesie <rohesie@gmail.com>
About The Pull Request
Fixes#45341 so that blueprints can no longer be used when not in-hand, and turfs now properly compile a list of the blueprint items it should hold by delaying when it retrieves this information, as objects such as manifolds and wires didn't initialize with their proper sprite.
Changelog
cl
fix: The CE's blueprints can no longer be used if not in-hand, and its scanning function has been fixed.
/cl
* Speeds up world init.
* Armor is now new inited for obj and the first level of subpaths.
* Actions is now lazyinited and deleted with empty.
* Actiontypes is now only inited when actually used and deleted once it pre-fills actions with the action buttons.
* Pipes now prefill their node list(s) in new() using new /list/ (count) syntax to speed up the list initaliztions and remove the init proc.
* Pipes no longer store their item version, instead creating it on the fly when deconned
* Walls no longer store their metal stacks, instead creating it on the fly when deconned.
* obj, walls, floor, plating, item, machinery, structure, pipe, pipenet, atom, and movable no longer have an (init) proc. (along with a few other smaller examples)
* Atmos can pass checking is now a var with the ability to have a proc be call in advance cases.
* (as a side effect, I had to fix a few things that were calling atmosCanPass rather then using the pre-calculated list, this should speed up chemfoam and flame effects greatly)
* Reverts upload limit
(remind me one day to defuck this, it could easily be a config thats not editable by vv to make changes easier)
* Makes apc update icon a bit faster.
APC new is some what high on the profile of world init, still not sure why, but this stood out as a waste of cpu so i fixed it.
* Fixes runtime with atmos backpack water tanks.
* Makes smoothing faster (and fixes turfs smoothing twice at init)
* Makes apcs init faster by replacing some spawns with addtimer
* fix transit turfs.
Basically, they key difference between StonedMC and CarnMC is that when multiple ticks want to run at the same byond tick, we divvy up the tick between the subsystems, rather then allow one subsystem to hog it all.
The key difference between StonedMC and GoonPS is that we allow the subsystems to tell us how to divvy up the tick using flags and priority.
The new SS_ flags allows us to select behaviors that used to be piggybacked as side effects of dynamic wait or default but sometimes unneeded behavior.
Dynamic wait is 100% gone, lower priority and SS_BACKGROUND are better more refined ways of doing this when combined with MC_TICK_CHECK
I have by design never looked at the inners of goonPS, so this is all original code but I know it uses two loops because of comments by goon devs on reddit threads, that design didn't make sense before, but when I can tell a SS how much of a byond tick it is allowed to have, knowing how many need to run this tick is helpful I also know a bit more about how it works from piecing together comments in #vgstation.
Detailed list of changes:
Subsystems now have flags, allowing fine grain control over things like rather or not it processes, inits, rather it's wait is how long between runs (post run timing) or how long between starts, and rather or not late fires should cause the next fire to be earlier.
Mc now has two loops One loop handles queuing shit, one loop handles running shit.
MC now splits up tick allotment rather than first come first serve Subsystems can even request a bigger share using higher priorities. (It will even resume subsystems it paused if other subsystems hadn't used as much as it predicted they might need)
Default fps is now 20 This is related enough to the MC and it's a change that's really long since over due
All code oddities are most likely to be necessities to lower overhead on the mc since it runs every tick
Alexis