arma3_feedback_tracker

[gaunt depot]

High ground height, _array is list of AGL positions with z=0

_array inAreaArrayIndexes [player, 15, 15, 0, false, 15] // NO matches
_array inAreaArrayIndexes [getPosATL player, 15, 15, 0, false, 15] // Matches

Lets fix the bug instead of treating it as a feature

Yeah, it treats everything as ASL, not just centre

[solid marten]

bug with inAreaArray?

[gaunt depot]

inAreaArrayIndexes, expects all positions to be ASL

We figured only center position should be ASL but no, all positions are treated as ASL

[regal nimbus]

Uh, I would assume that if you're feeding it with positions then it's just doing arithmetic.

Question would be what happens when you feed it with objects.

Given that doing an arithmetic 2d cylinder on ATL/AGL is questionable, maybe it's inAreaArray that's the bugged one :P

[solid marten]

in your example you supply to inareaarray array of agl positions and center is pos atl it has no matches, this bug?

[gaunt depot]

Looks like I messed up 🤔

Yeah, I already changed my positions to be ASL and then did the test so there was no matches

inAreaArray is fine

Still my point stands, inAreaArrayIndexes expects all positions to be ASL, not just center

Sorry for confusion

Still, I'd just fix the command to be uniform

[gaunt depot]

I wonder if mines can be made to damage buildings less?

Kinda strange to have 2 claymores destroy a stone building

[untold sky]

Let's stabilize the player character so he stays at the same angle in water, facing towards the horizon.

The player is swimming up, we need to push him down so lets push by 0.04
Okey he's now in balance its fine now.

Oh the player swam into a mine, he's now angled up a bit, we need to push him down by 0.006
Oh the player is now angled down a bit, we need to push him up by -0.006
Oh the player is now angled up a bit, we need to push him down by 0.006
Oh the ....

There is already a safety threshold, because small rotations are imprecise.
The threshold is 0.005.

[unborn acorn]

When an explosion is near(inside) a building, destroyed windows do not make hitGlass sounds when destroyed, can be improved? no sound only due explosion, all ok when shot rifles

[dreamy bane]

wish we had destroyable house doors. (like wooden doors etc) the game already has destroyable glass doors so it should be doable.

[sinful kettle]

Does anyone else find the curator camera annoying? It seems to use ATL height which means it keeps bobbing up and down as you move
It's not that bad over land but over water it's horrible

Imo at least it should be AGL not ATL ( I personally prefer ASL, but AGL is the most "backward compatible" solution)

[sinful kettle]

It makes no sense for a wooden door to just disappear tho...

[dreamy bane]

better than not being destroyable at all 🙂

[sinful kettle]

Even if it was acceptable that way it's not gonna happen because it needs modification to model animations

[dreamy bane]

hmm i didnt realize glass doors had breaking animation, it look like there was just particles flying around and door become invisible. maybe im wrong

the ones in Tanoa have animation it seems

well have to just rely on my script that opens the doors "when destroyed"

[unborn acorn]

Weapon tail sounds glitch ticket was updated with adds
https://feedback.bistudio.com/T164449

[unborn acorn]

FT ticket https://feedback.bistudio.com/T182471

[dreamy bane]

though having ability to hide the door would also work and be much better than just opening it

[sinful kettle]

That "becoming invisible" is an animation (hide animation)

[dreamy bane]

no change to just change the texture to invisible or something?

[sinful kettle]

Do they even support hidden selection textures?
(and even if they did, you'd still see the shadow)

Plus the geometry is still there

[dreamy bane]

yeah disabling geometry would be needed. idk about selections if they have those

[solid marten]

probably some config value for the building

[solid marten]

would you even hear shuttered glass over explosion?

[gaunt depot]

From my experience all destructible buildings are affected. Not sure how this can be fixed though.

[solid marten]

as I said earlier the command was introduced awhile back it wont be a good idea to radically change it now

[solid marten]

mod?

[gaunt depot]

No, fully vanilla game.

[solid marten]

with mod I mean

you can make a mod and tweak all the buildings

[hallow sun]

You can even shoot a building with bullets and destroy it, just takes a while.

[gaunt depot]

Yeah, but this might make other explosive sources weak, people got used to have buildings fold from a single HE shell

[solid marten]

this could be done by ignoring small damage completely, and I think there are some mechanisms in place already, maybe not working as expected. If you make a repro with a single bullet downing a building I can have a look

[solid marten]

previously damaged, modded? There is not enough damage in vanilla HE shell to bring down vanilla building with one shot

at least it wasnt like that when I was playing with building destruction years back

[unborn acorn]

nope, falling glass is make no sound when explosion?

[hallow sun]

Just tested real quick, shooting with a lynx or cyrus at buildings.
Looks like it was only modded buildings that can be damaged this way.

[gaunt depot]

All stock of course. By folding I meant building part being take out, not completely destroyed. One HE shell usually does that.

Nothing wrong with that

Having claymores also do that is the issue, not sure if it can be solved by tweaking the buildings

[solid marten]

tweak claymore, is there like a separate anti-personnel damage setting or something

[cyan basin]

oh yeah, the building damage is very arcadey

[gaunt depot]

@solid marten Can I ask for a bit of insight? Does setMissileTargetPos [0,0,0] stops the missile from going to target position or will it actually try to go to [0,0,0]?

Basically I want to stop missiles from stop going to target position

And another question, what is object complexity requirement for setMissileTarget? Does it have to be full-fledged entity? I was trying to have missile follow flare shot but it doesn't work

I wonder if its just an oversight or its possible to have missiles target even simplies entities

Having missiles follow actual flare shots would be cool script (flare shots are global)

[regal nimbus]

The trouble is that the blast damage doesn't distinguish between concussion and fragmentation, so if you want an explosion to kill people then it'll kill buildings, and vice versa.

[gaunt depot]

Yeah I figured there is no easy solution

Wish there was something property in CfgAmmo to fraction the damage depending on armor or something

or whatever buildings have

[regal nimbus]

Honestly the blast damage vs units calculation is so bad that I was seriously wondering if they could improve it without breaking backwards compatibility.

[gaunt depot]

Guess its 11 years too late

[solid marten]

[0,0,0] definitely has special meaning, not sure if it stops it, have you tried?

[gaunt depot]

I did brief tests but couldn't say for sure.

Also I wonder if setMissileTarget also does some internal weaponLockSystem checks and stuff like that and its not about entity complexity

If it does, I wonder if having a flag to bypass them could be a thing

[solid marten]

yeah it takes object but then for some reason tests if it is a transport. Gimmie repro with flare will see what I can do

[gaunt depot]

Maybe missiles does transport function calls during flight or something?

[solid marten]

if you do setMissileTarget objNull does it not cancel the targeting?

[gaunt depot]

Do you want repro now? Can cook it in a few mins.

[gaunt depot]

It does, cancels properly

Hmm, maybe setMissileTarget objNull also cancels setMissionTargetPos too?

[solid marten]

now or later, I will look at it over weekend

setMissionTargetPos [0,0,0] this is the init default so maybe you have to use both to cancel targeting, try it

[solid marten]

next dev

[gaunt depot]

1. Select BIM 9X (AA SR)
2. Shoot without targeting, script will try to setMissileTarget to flare
3. Observe that target wasn't set

[daring wagon]

15:21:36 Serialization of Display variables is not implemented
15:21:36 ➥ Context: Serialize Game Variable: _display
15:21:36 Serialization of Display variables is not implemented
15:21:36 ➥ Context: Serialize Game Variable: _this
15:21:36 Serialization of Display variables is not implemented

Would it be possible to print the parent script or something where this variable is being used/defined?

[solid marten]

Maybe you would want to update biki. The sure way to cancel locked missile would be setMissileTarget objNull; This sets target to null, sets direction to [0,0,0] and sets missile state to LOST

So maybe you need to also set position to [0,0,0]

[gaunt depot]

How many states are there? Maybe a worthy thing to return with getShotInfo?

[devout wave]

50 and five territories, but that's not important right now

[solid marten]

missile has own command set

[solid marten]

https://community.bistudio.com/wiki/missileState

https://community.bistudio.com/wiki/setMissileTarget

[gaunt depot]

Nice!

[regal nimbus]

Sounds like we're gonna get more feedback on missile AI than any other Arma AI :P

[surreal bough]

comparing the two gameplay of the game Dayz and ARMA in my experience, I really lack more sound for steps, opening doors and manipulating equipment

AI literally does it silently

[unborn acorn]

more sound for steps
there are in game indoors footsteps but undefined, so used same step sound as for outdoors, can be improved https://feedback.bistudio.com/T181949

[valid ice]

outdoors and indoors sounsets are handled by sound shaders, I dont think they are intended to use specifically different sounds for each state

Ah nvm, just saw the actual ticket content

[fast lodge]

It would be really nice if addMagazineCargo could get used on local containers just like addItemCargo

[gray wharf]

it can…?

[fast lodge]

let me test it again because some month ago I tried to add a magazine with 50% ammo count into a local weaponholder on a server and it did not work

oh wait Im talking about a different command

addMagazineAmmoCargo

this one

it would be nice if it can used on a local container

[rose cloak]

wut intensifies

[devout wave]

I dunno if the global effect part of it has any unforeseen consequences, but in theory you should be able to use it on a local-only object.
Note that:

• you must use it on the machine where the object is local; other machines don't know it exists
• you must get the reference to the object on the machine where the object is local; other machines don't know it exists, and if they have an equivalent copy, their own local reference will be different and won't work on different machines

[rose cloak]

it werks

called from client in 3DEN MP

[fast lodge]

yea I tested it again and it does work for some reason when I tested it a year ago it wasnt working at all

[surreal bough]

🤪

[cyan basin]

how does that happen?

i have seen the other month on the milsim group someone dead with their limbs stretched on a window lol

[daring wagon]

https://feedback.bistudio.com/T182548 Ability to attach entity with given class name to eden editor cursor, same as it happens when one selects an entity from the asset list

[gaunt depot]

I had this happen when you use wrong TakeBag action (there are 2, one is broken)

[gritty moth]

I think the 2x sight should be 4x, and 4x sight has a 6x option

[solid marten]

So it happened that for some strange reason the helicopter was declared dead prematurely when it was crashing upside down. Now the next simulation cycle will determine this, and because of this it actually nicely explodes now with a bang

[surreal bough]

I had two tanks collided with players and AI, ai tank immediately disappeared, and units were sitting on top of it on the armor - these units did not leave and left such a legacy

[untold sky]

Zeus camera has a flag to disable the ATL mode.
Eden sets it by default, and Eden has a "camery copy terrain" mode that can disable it.
So probably we can just default it on in Zeus. But would people then complain who preferred the ATL mode 😄
Wouldn't AGL make it much worse, when instead of bobbing with terrain, you also bob over every building?

[outer frigate]

atl works better when <100m atl imo

[untold sky]

Serialization of Display variables is not implemented
15:57:50 ➥ Context: Serialize Game Variable: _display <- [RootScopeName] L5 ()
[SubScopeName] L8 ()
<- Serializing Spawned scripts
next week

It shows the whole script callstack. Even if the variable is in RootScopeName scope. You'll have to check all the files in the stack to see which one has the variable

[untold sky]

Somewhat simple to do for create unit operation.
But there's also Steam composition, custom composition, group, marker, trigger, module, waypoint.
And not all of these have a classname to go with

[daring wagon]

How does the Engine handle that. What information does it need?

[untold sky]

It stores it all in the units tree.
What tree it is decides type and side.

Entries in units tree contain classname, icon, creation flags.
Composition item has flag whether steam or custom, and the foldername
Marker item has icon, classname, markertype and color
Trigger item has texture and classname
Module item has icon, classname and creation flags
Waypoint item has classname and icon

I would need to implement new wrappers for all of these..

Unit I already got done

[sinful kettle]

No that's AGLS

AGL is terrain + water

[daring wagon]

No idea what the best approach here would be.

[sinful kettle]

Actually now that I think about it AGL moves with waves so it shouldn't be AGL either

[untold sky]

There is no "backwards compat" to keep in mind with how the player moves a camera around.
So it doesn't really matter.
If Eden is ASL then that makes the most sense to use there too

[sinful kettle]

Yeah let's go with ASL

[untold sky]

I think the most universal would be array of hashmaps.
Hashmap has type entry that says what kinda thing.
And then per thing there are different sub-info's.
And then you can pass multiple with offset entry in each, to make a composition.

Though composition is also very hard..
For composition I would have to basically spawn them all, generate a copy-paste out of it. Delete them all and create a dummy composition with that copy-paste memory..
I'd rather not do that 😄

All options are a pain. Unless I just leave it at "unit only"

Maybe for future if someone feels like being pained.
We can make it forward compatible and make it array of hashmaps. But for now basically only allow
[{classname: "B_Survivor_F"}]
one entry, only classname.

Later on we can add "type" for others, or offset if someone feels like composition

[daring wagon]

Yeah I guess unit/prop is the most important one.

[untold sky]

Do you actually currently need anything other than single-unit?

[daring wagon]

Let me check

Nope, I disabled markers and stuff from adding it to favorites so units only

[untold sky]

Then I'll do jus that and we can maybe extend it later

[untold sky]

Actually only hashmap as argument.
Then you can maybe do

set3DENAttachedCursorEntity {type: "composition", objects: [a, b, c, d, ...]}

We need some syntax to create a actual hashmap 🤣

Something like

#{
  "a": 1,
  5: _variable,
  missionNamespace: player
}

would be neat 😄

[outer frigate]

would maybe close with }# to make it easy to identify when used in code blocks?

[untold sky]

Mh.
<> thingies are unused for opening/closing blocks

two character closer feels weird

Implementation wise easy to do.
And code editors would already support matching {} pairs and could just ignore the # around it

[outer frigate]

given <> is less likely to be supported by code editors id say #{}# works better imo

even if two character closer feels weird... its sqf

[gray wharf]

so HashMap select x would be #{}##x? 😛

[untold sky]

# select command doesn't work on hashmaps.
But that would look like that if that worked

[gray wharf]

make it

[outer frigate]

hashmaps are unordered so select is useless

selectRandom works though iirc?

not according to biki... hmmm 🤔

[gray wharf]

_myHashMap # 'indexX'

[outer frigate]

get 🙂

[gray wharf]

select

[outer frigate]

😠

[gray wharf]

I dislike # really 😄

[outer frigate]

i love it

it means i can make my code even more unreadable

[gray wharf]

https://tenor.com/view/go-leave-me-alone-golden-girls-you-have-to-go-now-rose-gif-22287138

[outer frigate]

😄

🐸

anyway im pretty sure there was a ticket for selectRandom in hashmaps, trying to find it

hm, cant find it - any interest in expanding it to hashmaps?

[untold sky]

A random key-value pair 🤔
But why 😄
selectRandom keys _hm is doable but expensive

[outer frigate]

ive wanted to randomly select an item and its data before but cant remember precisely why

(i.e like id'd items)

[untold sky]

@daring wagon Have you tried getting composition folder paths?
I don't think there is a way for scripts to read what compositions are available.
Besides reading the elements and treeItemData in the right-side tree inside eden.

Adding composition from "folderpath" would be easy to do, but not worth if its not possible to find what compo's are available 🤔

I have now implemented unit and marker cuz it was easy.

[daring wagon]

Compositions are messy as there are several implementations. I don't feel like messing with it if I dont have too 😄

There are

• CfgGroups
• Steam Compositions
• Local Compositions
• Config Compositions (Of these I can get the path to the file IIRC)

[untold sky]

CfgGroups would be "group" type with classname.
local/custom compositions yeah... Eden gets it from treeItemData so I assume scripts can read the tree UI too. But that's quite a chore to enumerate compositions

https://community.bistudio.com/wiki/create3DENComposition also only supports CfgGroups and config compositions from Cfg3DEN.
The 3den config compositions isn't even documented 😄

[daring wagon]

I remember I made a ticket related to that command.

Some type of composition didn't work with it

(findDisplay 313 displayCtrl 56)tvData tvCurSel (findDisplay 313 displayCtrl 56) //"B_APC_Tracked_01_AA_F"

[untold sky]

When you click on a composition what does it say?

[daring wagon]

Checking

"OutpostE"

[untold sky]

That should be foldername

[daring wagon]

"WW2\SPE_Assets_c\Compositions\3den_c\sqe\SPE_Comp_GER_Camp_Small"
It is

[untold sky]

and tvValue says if its steam or custom/local

[daring wagon]

It's folder name if it's a eden composition defined in config.

The other compositions are CfgGroups again I guess

"C:\Program Files (x86)\Steam\steamapps\workshop\content\107410\3248475523"
For steam composition

[untold sky]

Ugh.

CfgGroups builds config paths by checking parents

parent parent parent itemData >> parent parent itemdata >> parent itemData >> itemData

"CfgGroups" it finds the class that has "side" entry matching the currently selected tree.

faction >> type >> group

config composition is in a "side" tree, instead of the compositions tree. And there it has the classname.

in the compositions tree its either steam or local based on tvValue

Well that's all alot of fun 😄

group spawn would probably be by config path then, not classname

[daring wagon]

Yeah, makes sense I guess

[sacred river]

Is there any chance for having this in 2.18? Currently, its only possible via extension or external tools to get the BUID. Would be really helpful for scripted bans

[hallow sun]

You can just use the UID to ban

private _UIDToBan = "76561197960287930";
"RCONPW" serverCommand format ["#ban ""%1""",_UIDToBan];

[sacred river]

also true, same for #kick - but getPlayerBUID or/also (getPlayerBUIDFromUID) is still useful in some cases

[gritty moth]

Starting the game will cause PS to crash

[gray wharf]

PS?

[solid marten]

play station

[solid marten]

The command is useless, I have edited the description, we have other commands that do a better job returning the same info and more. But since it is nular and meant to obtain only player settings, you can forget about it.

[keen sable]

what about a currentNamespaceName command so we dont have to use a function ```SQF

private _fn_getCurrentNamespaceName = {
if (currentNamespace isEqualTo missionNamespace) exitWith { "missionNamespace" };
if (currentNamespace isEqualTo parsingNamespace) exitWith { "parsingNamespace" };
if (currentNamespace isEqualTo profileNamespace) exitWith { "profileNamespace" };
if (currentNamespace isEqualTo uiNamespace) exitWith { "uiNamespace" };
"error";
};

[alpine tulip]

https://community.bistudio.com/wiki/currentNamespace
?

Not sure what is the usecase there

[gray wharf]

https://tenor.com/view/but-why-wtf-wth-gif-18627417

[keen sable]

easier to get the namespace name for debug or something else, copy paste the _fn_getCurrentNamespaceName func its to hard requires to much work

[gray wharf]

what about

hint str currentNamespace
```?

[keen sable]

[gray wharf]

(well, does it work?)

[keen sable]

never tried

1s

[gray wharf]

but yeah no, 0% chance for the currentNamespaceName command to exist
it's a bit of a pain to create commands so they better do something nice 😄

[keen sable]

na it doesnt just return the namespace type

but in a string

whatever not a big deal

https://feedback.bistudio.com/T176268

[gray wharf]

?

[gray wharf]

it can be worth a ticket to request for stringified namespace to return its name, IDK

[keen sable]

the ticket explains it well doesnt it? i know dedmen subscribed to the ticket so he might be interested on doing it

[gray wharf]

yeah but why share it here? unrelated to the namespace thing and pointless to post it here no?

[keen sable]

i mean its a old ticket so maybe it fell into oblivion 😄

[gray wharf]

https://tenor.com/view/perhaps-cow-gif-14441933648207730172

[solid marten]

Namespaces do not have debug name

[gray wharf]

could they?

[solid marten]

they are currently return "namespace" so it cannot suddeny return soemthing else

[gray wharf]

yeah, I mean could it be made so that str missionNamespace would return "missionNamespace" or is that .ToString() a tough task?

[solid marten]

can extend name command

[gaunt depot]

I very much doubt anyone uses str missionNamespace == "Namespace" check in their mission

[solid marten]

name currentNamespace - "missionNamespace"

[gaunt depot]

typeName missionNamespace maybe

[solid marten]

thats what I meant :0

but it returns "NAMESPACE"

so no

name it is

[gaunt depot]

I'd just change str NAMESPACE to return namespace name

[solid marten]

or type

can use type currentnamespace

[solid marten]

no too late in the game

so it tis name or type pick one

[gaunt depot]

type as in new command?

[solid marten]

both return string so

type is existing command

[gaunt depot]

There is type LOCATION and name LOCATION which seem to do the same thing

[solid marten]

I go with type though

yeah

type is almost never used so i go with type

idea for the return values?

[gray wharf]

I'd vote for name tho :(

[solid marten]

why do you have to always be difficult

[gray wharf]

because namespaces are a same type, they are just different usage

[keen sable]

i vote for name also, i have never seen someone using the type command xD

[solid marten]

NAME NAMEspace meh

type namespace 👍

[gaunt depot]

space NAMESPACE

[gray wharf]

mespa naMESPAce!

[solid marten]

namespacename namespace

[keen sable]

ye better

[gaunt depot]

Speaking of str of stuff, would be so great if entities also had classnames in it, its so confusing that you get same string for both unit and vehicle of unit if they're commander:
[player, vehicle player] => [B Alpha 1-2:1 (Sa-Matra),B Alpha 1-2:1 (Sa-Matra)]

but I guess its 11 years too late again

[solid marten]

changing return values is VERRRRRRY risky

[gaunt depot]

[player, vehicle player] => [B Alpha 1-2:1 (Sa-Matra) (B_Soldier_F),B Alpha 1-2:1 (Sa-Matra) (I_Heli_Transport_02_F)]

[solid marten]

so

type or name

[gray wharf]

typeName

[solid marten]

taken

[keen sable]

hahaha

[gray wharf]

waitaminute

[gaunt depot]

[u, vehicle u] => [B Alpha 1-2:2 (B_Soldier_F),B Alpha 1-2:2 (I_Heli_Transport_02_F)]

[solid marten]

no u

[gaunt depot]

no u => object id

[gaunt depot]

Speaking of which, is there a command to return that object id?

211f9600b80# 1832934: dummyweapon.p3d

1832934

[solid marten]

getObjectId?

[gaunt depot]

Oh yeah, that it is

Fairly recent too, that's why I missed it

[gray wharf]

wait, that's a thing?

[solid marten]

visitor id i think

[gray wharf]

https://community.bistudio.com/wiki/getObjectId
👀

ah yeah
https://community.bistudio.com/wiki/getObjectID

casing ^^

[gaunt depot]

I.D.

[solid marten]

getObjectID

but yeah thats correct casing

[gray wharf]

getObjectIdentificationNumber

[solid marten]

add redirect

[gray wharf]

done!

[solid marten]

so name or type?

[keen sable]

name

[gaunt depot]

Btw you can use hashmap for fastest return

it can use namespaces as key

[solid marten]

ok I will extend name after I finish "ArtilleryShellFired" EH

[devout wave]

On the subject of name, setName on units handles fullname/firstname/lastname separately, but name only returns a unified string. There's no way to actually get the separate names (firstname and lastname can contain spaces so can't assume based on splitting the fullname at first space).
I don't really have a use case for this, it just occurred to me while you were talking about this. i think nameName should be added to do this

[gaunt depot]

UAV locality gets unplayable in live games, turrets sometimes take several minutes to switch locality, sometimes they don't seem to become local at all, I couldn't figure out any connections, feels random

I'll have a ticket soon-ish

[dreamy bane]

btw speaking of "name", name location does not return anything (empty string).

I used text instead

[gaunt depot]

Thank you Arma, very cool

Looks like empty world you see in the main menu, no idea how this happened

Still says VR for worldName

Anything else to run to try to figure what's going on?

Also looks like pre-Apex lighting

[alpine tulip]

More like default lighting (CAWorld to me)

[gaunt depot]

A3 Limbo

[gray wharf]

see its wiki page

[dreamy bane]

i was looking at it. noticed just now some ones comment that its only for 3D editor?

[alpine tulip]

Actually what the heck is happening/you've done out there

[gray wharf]

ye

[gaunt depot]

Joined my game with another client instance and got this

I think I know why it happened though, I started hosting by accident, then cancelled and joined

Which seemed to break the game

[alpine tulip]

Thanks, very useful repro step

[gaunt depot]

I wonder if you'll end up in limbo on a live server too

[gaunt depot]

1. Host game with 1st client
2. Host another game with 2nd client
3. Cancel hosting without starting any missions on 2nd client
4. Join 1st client's game with 2nd client

[solid marten]

yeah this will do it

[solid marten]

does it still do it (think not)

[gray wharf]

lemme try

[gray wharf]

doesn't seem to do that anymore
do you happen to know when that changed?

[solid marten]

should be on the logs somewhere

[gaunt depot]

Almost 1:30 to switch turret locality on dedic server

Couldn't get locality never switch though, only happens randomly on live server

[desert trench]

name _deadUnit should return what now?

[gaunt depot]

https://feedback.bistudio.com/T182587

Made a ticket with repro, both showing unreasonbly long time for UAV locality switch on dedicated server as well as turret taking the longest (and sometimes never switching locality in live games)

[solid marten]

ok

[solid marten]

name

[gaunt depot]

Wasn't it something like: ERROR: Dead unit or something?

Error: No unit

[desert trench]

unless this was fixed/changed in last few updates, its still the same AFAIK

you need cache the unit name for dead units - otherwise you get that error spammed

[solid marten]

last few LOL

dont even remember when it was changed

[desert trench]

"Error: No vehicle"

for dead bodies

with 2.17

edit: with cursorObject - which may not be the body

name cursorTarget

does work indeed

[solid marten]

imposibru

[desert trench]

it seems the identity gets reset or sth - like previous setName is no longer applied

its weird. seems where the AI is remove, the name gets reset once the unit (AI) has respawned

but the vehicleVarName is also different

cursorTarget,cursorObject,name cursorTarget,name cursorObject]

host

[bis_o2_3275,bis_o2_3275,"François-Marie Farage","François-Marie Farage"]

2nd client

[1f770fed900# 715175: spe_us_hbt44_uniform.p3d REMOTE,1f770fed900# 715175: spe_us_hbt44_uniform.p3d REMOTE,"McGregor","McGregor"]

[gritty moth]

Photoshop

[azure crane]

does not crash mine

[gray wharf]

G.I.M.P doesn't count

[cyan basin]

you didn't get any dialog boxes about missing stuff? this behavior normally happens if the world is missing (but in that case it loads altis), seems like the load call went through incorrectly and loaded the empty world fallback somehow. did you attempt to host with a game that was launched to -world=empty?

the empty world is a special one used when you're disconnected and have no game/scenario/mission running at all

been a thing since arma 2 oa

[gaunt depot]

Nope, no errors. I later rejoined and it was fine too.

[cyan basin]

odd

fun edge case

[solid marten]

https://community.bistudio.com/wiki/name

[keen sable]

i seen that isThrowable command getting added on 2.18 would be possible to add also a isRunning command ? https://feedback.bistudio.com/T177374

This is the workaround i currently have. idk if there its a better way of check if is running apart of comparing the player speed that it is not accurate at all

    isRunning = false;
    0 spawn {
        private _standingAnims = ["amovpercmevasraswrfldf","amovpercmevasraswrfldfr","amovpercmevasraswrfldfl","amovpercmrunsraswrfldf","amovpercmrunsraswrfldr","amovpercmrunsraswrfldl","amovpercmrunsraswrfldb","amovpercmrunsraswrfldfr","amovpercmrunsraswrfldfl","amovpercmrunsraswrfldbl","amovpercmrunsraswrfldbr"];
        private _crouchAnims = ["amovpknlmevasraswrfldf","amovpknlmevasraswrfldfr","amovpknlmevasraswrfldfl","amovpknlmrunsraswrfldl","amovpknlmrunsraswrfldr","amovpknlmrunsraswrfldf","amovpknlmrunsraswrfldfl","amovpknlmrunsraswrfldfr","amovpknlmrunsraswrfldb","amovpknlmrunsraswrfldbl","amovpknlmrunsraswrfldbr"];
        
        while {true} do {
            private _anim = animationState player;

            private _runningStanding = _anim in _standingAnims;
            private _runningCrouch = _anim in _crouchAnims;

            if (isRunning && { !_runningStanding && { !_runningCrouch } }) then {
                uiSleep 0.28;
                isRunning = false;
            };

            isRunning = _runningStanding || { _runningCrouch };

            uiSleep 0.1;
        };
    };

[fast lodge]

you could also check the speed instead of the animations

[lament escarp]

While this was likely intentional at one point, I have found that you cannot unload the magazine from a muzzle with showToPlayer = 0; to your player's inventory, but you can unload it to an external inventory object.

https://feedback.bistudio.com/T182614

[sinful kettle]

You can just get the current action group and see if the player is using the sprint actions (called FastXX, where XX is the direction):

_state = animationState _unit;
_actions = getText(configFile >> "cfgMovesMaleSdr" >> "states" >> _state >> "actions");
_actionCfg = configFile >> "cfgMovesMaleSdr" >> "actions" >> _actions;
_isRunning = _state in (["FastF", "FastRF", "FastLF"] apply {toLowerANSI getText (_actionCfg >> _x)})

I guess it'll return true if you're "prone sprinting" too

[keen sable]

yeah that works as expected in less lines, thanks

should i cache the configs reads or its not noticeable the impact ?

[alpine tulip]

Latter

[keen sable]

ok

[unborn acorn]

Since you do some in artillery there one issue I want send is about soundFakeFall[]={}; sounds on falling artillery projectiles not work if shot in a range more 1000 meters https://feedback.bistudio.com/T159917

[untold sky]

https://xkcd.com/2954/ I shall take this as inspiration for hashmap syntax 😄

xkcd: Bracket Symbols

’"‘”’" means "I edited this text on both my phone and my laptop before sending it"

[solid marten]

interesting, i have seen something strange in code possibly related, have to investigate

[hallow sun]

Can someone repro that ?
I might have found a cause for that Object 2:1029 not found rpt spam.
https://cdn.discordapp.com/attachments/476133397224488960/1258420913620389999/image.png?ex=6687fb4e&is=6686a9ce&hm=455bb059b1644b9215e0bcb4c3006af5d7db682d1c09a3238a4ebacd80a45a32&
See netID of the UniformContainer in the debug console in the screenshot (Unit was spawned and geared serverside (Not sure if it matters for the repro)).

After sever exec of this code to remove the Uniform the error will continuously spam the rpt.

_targetplayer = objectFromNetId "2:1030";
removeUniform _targetplayer; 

And you should see the error spamming with the netID of your UniformContainer.
This also triggers for removeVest but not for removeBackpackGlobal.
If you execute on the client there is no error spam in the server rpt.

15:54:38 Server: Object 2:1029 not found (message Type_463)
15:54:43 Server: Object 2:1029 not found (message Type_463)
15:54:48 Server: Object 2:1029 not found (message Type_463)
15:54:53 Server: Object 2:1029 not found (message Type_97)
15:54:53 Server: Object 2:1029 not found (message Type_463)
15:54:56 Server: Object 2:1029 not found (message Type_125)
15:54:58 Server: Object 2:1029 not found (message Type_463)
15:55:03 Server: Object 2:1029 not found (message Type_463)
15:55:08 Server: Object 2:1029 not found (message Type_463)
15:55:13 Server: Object 2:1029 not found (message Type_97)
15:55:13 Server: Object 2:1029 not found (message Type_463)
15:55:19 Server: Object 2:1029 not found (message Type_463)
15:55:24 Server: Object 2:1029 not found (message Type_463)
15:55:29 Server: Object 2:1029 not found (message Type_463)
15:55:33 Server: Object 2:1029 not found (message Type_125)
15:55:33 Server: Object 2:1029 not found (message Type_97)
15:55:34 Server: Object 2:1029 not found (message Type_463)
15:55:39 Server: Object 2:1029 not found (message Type_463)

[gaunt depot]

This also triggers for removeVest but not for removeBackpackGlobal.
OA good, A3 bad

[fast lodge]

At the end it is probably connected to the bug that vest and uniform vanish sometimes for others and the server but the backpack always stays for some reason 🤣

maybe that comes down to that a backpack is a object (cfgvehicles) and uniform and vest are defined via cfgWeapons

but yea arma inventory code is super spaghetti so I doubt anyone is gonna touch it

[solid marten]

It was not connected t distance but to the altitude, so it was not playing falling sound is the shell was above 500m or something like that. The whole trigger routine was suspect, like returning sea or ground position (either either) but then if sea altering it more which made no sense.

[regal nimbus]

Anyone run into an issue where hideObjectGlobal on terrain objects doesn't work on some init timings?

[sinful kettle]

Shouldn't you always use hideObject for terrain objects anyway?

[dreamy bane]

have to say yes. i had problems using hideObjectGlobal first in test mission, but then somehow got it working in my mission. still dont know what happened there

trees would become invisible but you could still collide at them

[regal nimbus]

Why?

The hide terrain objects module uses hideObjectGlobal by default.

[regal nimbus]

Also this is a localhost test case, so I don't see why there would be locality issues.

[daring wagon]

Can you elaborate on that?

[regal nimbus]

oh hey, sometimes it unhides only some of the objects :/

Generated the same issue with hideObject.

[solid marten]

object streaming issue?

[regal nimbus]

I'm guessing so but it's so hard to replicate reliably that it's hard to find the conditions.

The test case has the camera within draw distance of the area where the objects are hidden, so I'm guessing that's a risk factor.

Often switching map to altis and back again makes it worse.

[daring wagon]

We had the issue where players would see different objects. For some tree 1 was hidden for others not.

[regal nimbus]

Yeah, back when we had issues with destroyed buildings I figured it was an issue specific to MP, but it turns out you can generate it in localhost...

My destroyed buildings workaround probably worked by luck. The method was to wait for the client to ask for the list of destroyed buildings.

[sinful kettle]

I mean they're local anyway

[sinful kettle]

Maybe the message never arrives?

[solid marten]

So, bare with me, when one enters vehicle turret, the gunner changes and this forces turret update so turret changes locality. When you take control of the unit in turret, it forces unit update but not turret because nothing changed, the gunner is the same. turret lock is in the same turret update message, could you test it by locking turret to force update and report back?

I think the same true for taking control of the driver (havent looked yet but all points to it)

Thats a major oversight if true

[gaunt depot]

Lock turret right as gunner locality switches?

Did lockTurret on server and it instantly transferred the locality!

You have to change it though, can't just true-false it in the same frame.

oh wait, you can't unlock it as turret is no longer server-owned anymore

gunner uav addEventHandler ["Local", {
    params ["_unit", "_local"];
    if(!_local) then {
        uav lockTurret [[0], true];
    };
}];

Doing this in SERVER EXEC as client in repro mission makes turret instantly switch locality alongside the gunner

can't figure out proper moment to unlock the turret though

Is turret lock stored and synced in the vehicle or the turret? 🤔

[devout wave]

Given that it's a UAV, I wouldn't think you actually need to unlock it other than to force an update again next time it changes locality

[gaunt depot]

locks are also used to prevent connection and taking control of uav crew

You won't be able to reconnect to gunner if you don't unlock the turret

[devout wave]

Oh, I didn't know locking did that

[gaunt depot]

Yeah should probably add a note somewhere on wiki

I previously had a complicated system of enableUAVConnectability and disableUAVConnectability to control UAV access but turns out all you have to do is to lock/unlock it for each player

No need to mess with terminals, wait for terminal updates, etc.

UAVConnectability commands are stupid anyway, you can connect terminal and ditch it somewhere and these commands don't work for non-unit stored terminals

You need to mess with per frame locking though to make locks local

But you only have to do that when scroll wheel menu or terminal is open

Nothing is easy in Arma

[gaunt depot]

gunner uav addEventHandler ["Local", {
    params ["_unit", "_local"];
    if(!_local) then {
        uav lockTurret [[0], true];
        {
            addMissionEventHandler ["EachFrame", {
                uav lockTurret [[0], false];
                removeMissionEventHandler [_thisEvent, _thisEventHandler];
            }];
        } remoteExecCall ["call", _unit];
    };
}];
```So there is nothing complicated about it really, I was just using `onEachFrame` to unlock on client which stopped my debug hint and I assumed turret locality didn't change, but it did.

So basically, if gunner locality goes away, also lock the turret and have new gunner owner unlock it back

Glad there is a hack to fix this issue but would love to have this fixed properly in the engine

[unborn acorn]

Spam
https://feedback.bistudio.com/T182345#2633888

[desert trench]

[BUG] hideObject[Global] at times keeps collision and pathfinding for AI - even locally (SP/hosted)
https://feedback.bistudio.com/T169443

our conclusion was its due to PX state not reliably updated

[solid marten]

so it worked?

[solid marten]

there is an assert which is triggered in internal build warning of entity leak. we will need a proper ticket @hallow sun

[gaunt depot]

Yes, locking turret insantly moves the locality

I haven't tried in a real game yet but will be trying this hackfix soon-ish

[untold sky]

drawLine3D thicc edition

[solid marten]

what it looks like on bends?

[untold sky]

drawLine3D only goes between two points, so there are no bends. The next segment would be a new line and the corners will overlap eachother

[solid marten]

I mean if you join 2 lines at different angles? I abandoned this after it looked mega ugly the thicker you go

[alpine tulip]

I mean it's less likely to have the feature, but what about "line style" option like solid, dotted etc? What I know is line pattern is defined in CfgCoreData IIRC

[solid marten]

it needs some radius on ends and that's a job

and performance loss

[untold sky]

It will look mega ugly because the ends will be flat and overlap

[untold sky]

Sure. You can do line style in SQF by splitting your line into multiples.
Only limit is your creativity

[solid marten]

Not to mention that if you make lines semi opaque you will see the overlaps

one possible solution would be to provide option for the end radius

roundEnds = true

[untold sky]

But engine line drawing doesn't support rounded ends. That'd be alot more work.
Adding width was just using a parameter that was already there anyways

[alpine tulip]

Uhh yeah that's true

[gaunt depot]

If we could draw a quad in the world similar to lines we could make it draw a circle to act as a radius at the end

Not sure how to even call it, drawQuad3D? Name would assume you can specify everything for a quad 🤔

drawTexture3D also sort of sucks

drawIconWorld

drawIconWorld ["circle_texture.paa", PositionAGL, ColorRGBmodifier];

on line edges and you have smooth line angles

Or even:

drawLines3D [[Pos1, Pos2, ..., PosN], Color, Width]
```which adds these edges by itself

Less SQF too

[gaunt depot]

That fuzzy border of the blue line looks like stretched texture. I wonder if you could feed it another texture to see what happens, maybe have drawLine3D take texture path instead of color?

[sinful kettle]

Can you also add flag to show them thru walls?

[gaunt depot]

Hmm, I wonder if drawIcon3d can be made to show inside walls with a flag too?

[sinful kettle]

It already does show thru walls

[gaunt depot]

Sorry, meant to not show through walls, be culled similarly to drawLine3D

Wow, drawLine also got width parameter

I wonder if other map-drawing commands can get it too 🤔

Having to calculate pixel width and redraw same shape 4 more times to make thick line right now

[untold sky]

Refer to #arma3_feedback_tracker message

[untold sky]

Refer to #arma3_feedback_tracker message

[hallow sun]

https://feedback.bistudio.com/T182649

[untold sky]

Fix the 8192 limitation on format
👍 / 👎 ?

[alpine tulip]

But is there side effect?

[untold sky]

No I'm just being funny.
I just found there is a decade old "//todo" on there which says to use a resizable buffer

[tacit kettle]

also what is the new limitation if there is one ?

[untold sky]

Strings longer than 4096 characters would be (very slightly) slower. Because the buffer will resize and copy the data into a new one
99% of uses are probably less than that so no difference there

[gray wharf]

strings don't have a max length in A3, right?
max 10M
https://community.bistudio.com/wiki/String

and why 2056 in OFP

[untold sky]

The output of this command is limited to ~8Kb.
It was 8191. The buffer had 8192 elements but the code wasn't very nice

The new limit will be whichever power of two is smaller than 10mil

[gray wharf]

. . . 65,536 ?

no, my bad ^^ wrong column

8388608

2^23

[untold sky]

yeah.
I can make it go up to 10 mil. But thats extra code and at this length I don't care enough

[gray wharf]

tbh, yeah ^^

if someone wants to format something this long they are already wrong in their mind :D

or just redirect to str formatText ["%1", _string] you know 🤣

[untold sky]

It now also throws script error if the max limit is reached and the output is truncated.
But its not an error that stops execution so it'll just continue after notifying you

[untold sky]

Took the chance to rewrite it and make it faster

_array = []; 
_array resize 8191; 
_array = _array apply {32}; 

[
diag_codePerformance [{ format ["%1aaa", toString _this]; }, _array, 10000],
diag_codePerformance [{ formatOld ["%1aaa", toString _this]; }, _array, 10000]
]

[[0.024775,10000],[0.0315558,10000]]

Here the new one has to resize and copy the buffer twice. But still performs better.
The magic of not calculating string size twice and using memcpy (just blasts memory from one end to another) instead of strcpy (needs to check every character for whether its the end)

Also a little free extra feature.

formatOld ["%%%1", 5] -> "5"
format ["%%%1"] -> "%5"

You can now escape the % in format string to print it out.
Chances of this breaking someones backcompat. Well ya shouldn't have used it wrong in the first place!

[alpine tulip]

format ["%1","%"] 😄

[limpid rune]

lol I needed this previously but couldnt see a way around it 😬

ended up just writing it out

[daring wagon]

I always added % as argument

[untold sky]

Today I learned we have two string format's that are copy-pasted..
https://community.bistudio.com/wiki/textLogFormat

[gray wharf]

effin' finally! 😄

[devout wave]

(the effin' stands for formattin')

[limpid rune]

dont have a choice for field manual

it just eats any % you throw at it

I guess actually I could have passed it as an arguement

[regal nimbus]

Anyone recognise the map in this ticket? Looks like Tanoa but apparently has a marker called "poisto". Would like to replicate: https://feedback.bistudio.com/T120756

[devout wave]

That's a marker, not a location, so it's not part of the map. I'm pretty sure that is Tanoa.

[regal nimbus]

great, no replication then :/

[devout wave]

Yes, that's Tanouka on Tanoa, viewed from the hill to the south

[regal nimbus]

forgot that vanilla maps don't have markers for the towns.

hmm. hideObject doesn't just have a limit, does it?

Doesn't look like it.

[uncut briar]

It does not. You can hide all trees on tanoa without any issues for example.

[regal nimbus]

Yeah, in all artificial testing it seems fine. Hiding is fully persistent. Doesn't matter how much you hide or where. Doesn't matter whether you use hide object modules or init scripts.

So I think I'm down to init timing issues that only happen with real missions and so are impossible to create minimal replication cases for.

[daring wagon]

Is that format as well just for rpt?

[untold sky]

A new kind of spam 😄

[untold sky]

basically diag_log + format. And its enabled in internal builds only

[gray wharf]

soon I'll get FT files access and burn block 80-90% of spam…

[tacit kettle]

But how will you know about all those sweet facebook deals 🤣

[untold sky]

What cases?

[gray wharf]

I'll make the server e-mail you

[untold sky]

It did.
Easy fix.
Done.

[sleek scaffold]

I never added a subscriber! Was my account hacked? I don't even know how to add a subscriber other than myself

[devout wave]

That's a text comment written by a spambot that says "LSValmont added a subscriber". It's seen the system notifications and copied them to fit in.

[sleek scaffold]

How do other platforms prevent bots from doing this? One would think by making registration more strict but then you can have human created accounts handled later to bots...

[gray wharf]

the FT uses an obsolete platform that should be replaced… Soon™, in the meantime I should get access to its files and write home-made filters to prevent 80% of these spams

[alpine tulip]

80% more spams per spam!

[unborn acorn]

ye remove that spam pls https://feedback.bistudio.com/T182345#2633888

[gray wharf]

no u

[outer frigate]

and the remaining 20% is the funny spam right

[gray wharf]

Should be normal users 😄, spam users and spam comments

[cyan basin]

ft could have one last upgrade to a supported fork of phabricator called phorge. at least its on a supported platform instead of a discontinued piece of software. it wont really fix the issue but it will mean better support for issues and maybe some tip for combating this spam better

should be ideally painless to do

[gray wharf]

we are aware of alt solutions - just that I will try to provide filter an acceptable level of spam until IT replaces it all

[devout wave]

Why does that make it sound like you're going to be posting the spam yourself 🤔

[gray wharf]

I might

[cyan basin]

lol wut

[cyan basin]

after all, since its still getting use and the replacement seems to not have a high priority, the move would be the best of both worlds since it enables getting support for the issue and benefiting from possible enhancements and security patches for what is running

[gray wharf]

I am not taking responsibility for screwing up a prod server 😉

[cyan basin]

copy it, inspect any issues that arise from upgrading and then attempt a prod upgrade with a prior backup

[gray wharf]

no

"oh, the new version updates the db"
*rolls back files*
db is not compatible with old version anymore

and that's only one possibility

[cyan basin]

thats why you make a backup of everything

[gray wharf]

yeah, no

that's what I mean: no

[cyan basin]

okay fine, even if you got developer support?

well i tried ¯_(ツ)_/¯

[gray wharf]

I more or less know how to take precautions ;-p I just want to limit my own support to some filters easy to add/remove
the real operation itself will be done professionally by IT peeps, with a preprod server and everything that's good

[sacred river]

Actually, all the cases I had in mind are solvable with the Steam ID. For a long time I thought the ID in the Wiki for #kick and #ban is some weird internal ID as it is shown 47114712 in the example, rather using a example Steam ID starting with 7565 or something like that.. With just testing it a couple of days ago with Lou, I can convert all my systems to Steam ID.

[cyan basin]

heh

it could be clarified that the id is a Steam64 ID integrer

[solid marten]

Should be fixed in Revision: 151965 and it is good fix for profiling

[solid marten]

Here goes my week end

[hallow sun]

Found the issue ?

[solid marten]

thats what weekend is for

[fast lodge]

😂

[gray wharf]

https://tenor.com/view/stahp-stop-breakfest-club-breakfast-club-gif-5219633

[unborn acorn]

duplication of Launcher_MRAWS_closeShot1_SoundShader and Launcher_MRAWS_closeShot2_SoundShader, presence of sound shader for a completely different weapon (Vorona), and missing the Launcher_MRAWS_closeShot3_SoundShader in list
looks strange so I want to ask is this all as designed?

[solid marten]

Revision: 151966

[solid marten]

@sterile nova

[gaunt depot]

Thanks! Any chance to push this for 2.18?

[sterile nova]

looks like a mistake - is there a feedback tracker ticket for it?

[unborn acorn]

no yet, making

[daring wagon]

Thanks for digging up all these things!

[unborn acorn]

created: https://feedback.bistudio.com/T182700

Also I want to inform about some more configuration mistakes, listed here: #arma3_feedback_tracker message
It will be great to see fixes and improvements from there

[gaunt depot]

Tiny config tweak from me: https://feedback.bistudio.com/T172185 (M2 HMG minElev values are very small making it very hard to aim it)

[solid marten]

[stone tree]

https://feedback.bistudio.com/p/lewishamilton0998/

Spam account

[dreamy bane]

two new commands i'd like to see are case-insensitive **in ** (for string array search) and **add **command, similar to set but would add instead (numbers, arrays, maybe strings)

[rose cloak]

inb4 namespace/object change [string, code] for muh_knee change ["arrow", {_x + 1}]; to replace muh_knee set ["arrow", muh_knee get "arrow" + 1]; 🙃

[dreamy bane]

imo add command would be nice for that

or with the {_x + 1} would be cool too

was thinking of arrays where the add command could be used

[gaunt depot]

KEY in HASHMAP is key sensitive and much faster than STRING in ARRAY search for large number of values

[dreamy bane]

yep faster but not nice to use hashmap for everything , if you have simple array like ["B_Competitor_F","B_engineer_F"] and would like to have insensitive search for that

[gaunt depot]

Like 2/3 of my new data sturctures are hashmaps now

They're amazing!

Speaking of classnames, depending on source and some other nuances, I always try to stick to config casing and this eliminates having to care about casing, lowering everything, etc.

In cases where commands return lower cases or it comes from unreliable source (user input, etc.), you can do configName(configFile >> "CfgWeapons" >> _weapon)

Still, having case insensitive in wont hurt at all

inCI or something

As for adds, wanted something like that for a long time now

Basic form of ARRAY add INDEX, ARRAY sub INDEX, then more complex like @rose cloak suggested

Searching 960 files for ") + 1]"
16 matches across 11 files

Lots of cases where I just add one to array index

[dreamy bane]

yeah i like hashmaps too they are easy to work with. but i just dont want to turn even the shortest of arrays to hashmap because that it self requires work 😉

[sinful kettle]

What about insert?

[dreamy bane]

oh yeah thats good for arrays, forgot that one. But I was mainly thinking of adding numbers

so you dont have to do sqf _array set [ (_array # 0) + 1 ];

it would be nice for arrays too so you dont have to do ```sqf
(_array # 0) insert [-1, [123] ];

[clear glen]

or maybe modify set to take code on an array similar to apply

_array set [3, {_x +1}]

erp..edited because I spoke to fast

wouldn't work on a hashmap because itself can set code. maybe a seperate command like modify which could take an array index or hm key

ah as artemoz said already I see... yes. I would +1 that command

[sacred river]

Does anybody have also some issues with client crashes (multiple players at the same time) when some high polygon vehicles are used? This issue occured almost a year ago (doesn't matter if players run prof or stable), but the server is running prof all the time.

[alpine tulip]

What vehicle and what it does mean "used"?

[sacred river]

We had this issue with Redd'n'Tanks vehicles (all of them) and BWA3 NH90. Probably some new CUP additions too but not confirmed. And with used, being in them. Driving, flying. Or just be close to them when spawning them.

[alpine tulip]

Unrelated Qs:

1. Is it possible to have optimized map control? Don't process drawing polygons if forsts or sea and such has no color to draw? (mainly because a map is sometimes the best way to draw something in GUI)
2. A way/game option to tilt camera aligned with plane/heli roll?
3. A procedural texture to save/reuse current screen render?

[regal nimbus]

If it last happened a year ago then it might be obsolete. A3 client used to be capped to 4GB by default, and now it's... higher. Probably.

Although if an object is causing Arma to run out of memory that regularly then you should probably avoid using it.

[sacred river]

It started to happen a year ago (at least the first cases happened there), but it happens very often lately. It seems that it is related to graphics setting set to Very High or Ultra. The hardware doesn't seem to matter as it happens on high-end and low-end systems.

[regal nimbus]

If you can get some crash dumps then dedmen would probably be interested.

[sacred river]

And there is the problem. The game crashes with random memory error (access violation) but there are no crash dumps at all

[regal nimbus]

Universally?

IME the crash dumps are hit-and-miss at worst.

[sacred river]

I asked a couple of players who had the issue and none of them had any crash dumps

[regal nimbus]

That doesn't seem very likely. Crashes on shutdown at least are super common...

So even if they didn't have crash dumps for that issue, they should have crash dumps.

[sacred river]

Well that's what I thought, but sadly it isn't the case. The only thing everyone had in common that they were close to a source with high polygons (one of the players always with the BWA3 NH90). In the last happened case (before I asked here) the source is unknown, but probably something like that. Earlier this year we removed Redd'n'Tanks because players crashed around the Fuchs and Gepard when these vehicles spawned, but I didn't. And afaik all of these vehicles have high pylons (or at least very good models, but I know that the NH90 got a couple of million pylons).

[gaunt depot]

1. Could be some config property flag instead of checking color alphas. New CT_ even? CT_MAP_EMPTY

[rose cloak]

i had set in my message as well, but then i understood that the current syntax can set code as a variable value. Something like ```sqf
MUH_fn_array = [];
MUH_fn_array set [0, {_x + 1}];
[1,2,3] apply (MUH_fn_array#0);

[gaunt depot]

Needs another command name because this assumes you're setting code

_array applySet [3, {_this + 1}]

_array applySet [[1,2,3], {_this + 1}] even

Using this opportunity to raise awareness of setN command idea: https://feedback.bistudio.com/T178841

Imagine the world were you could do private _pos_xy = getPosWorld player setN [2, 0];

Or instead of building large arrays for commands that take shitton of arguments you could modify and feed them in a single expression:

    // Imagine _array is pre-made array to draw an icon
    _map drawIcon (_array setN [2, getPosWorld target1, 6, "Target 1 text"]); // Modify position and text
    _map drawIcon (_array setN [2, getPosWorld target2, 6, "Target 2 text"]); // Modify position and text

[gaunt depot]

3. This will let us do TF2 style kill cam freezes

Though knowing Arma it will probably end up with half of assets on low LODs after sudden zoom in

[alpine tulip]

Maybe better to have drawLineUI drawIconUI etc

[rose cloak]

getSet

[gaunt depot]

get assumes you return one value by index/key though, isn't it?

My original name idea was setRet

[solid marten]

letGetSetRet

[clear glen]

applyDontReturnCopyWantOriginalModifiedSet

[gaunt depot]

+Please

[desert trench]

are freeze dumps still worth to upload to A3 FT?

[untold sky]

Most crashes are easier if I just get them via DM

[desert trench]

alright. just mdmp+bidmp or rpt/some 1-3 lines of basic description too? (repro steps if known are also useful i guess)

[untold sky]

For unknown crashes to check just mdmp to me.

For crashes with reproduction steps, so you basically know how it happens and we know that its actually a relevant bug, FT ticket.

There are alot of unactionable crashes on FT that I really can't do anything with, I don't want these to waste space

[dusky tendon]

I discovered a bug in obtaining selectionPosition directly through a string.

(player selectionNames 1) select {_x == "proxy:\a3\characters_f\proxies\weapon.001" } apply {player selectionPosition [_x,1]}; // //String Proxy by selectionnames
return [[0.223065,0.549794,1.4695]];

"proxy:\a3\characters_f\proxies\weapon.001" in (player selectionNames 1);
return true;

((selectionPosition [player, (player getVariable ['unit_weaponProxy','']), 1])); //how String Proxy by unit variable
return [0,0,0];

((selectionPosition [player, "proxy:\a3\characters_f\proxies\weapon.001", 1])); //how String Proxy 
return [0,0,0];

The bug occurs on a dedicated server on the client. It repairs itself when the LOD changes. (moved the camera further away from the player).
The bug may happen again, but I can’t describe specific situations.

[gaunt depot]

Some LODs are probably accessible only when they're loaded

[dusky tendon]

I checked all the LODs using the syntax from 2.06.
These results were returned to me when I started the mission
Lod 1 = [0,0,0]
Lod 2 = [0,0,0]
Lod 3 = [0,0,0]
Lod 4 = [n,n,n] has coordinates
Lod 5 = [n,n,n] has coordinates but not always
For now, I will only be looking for a position at 5, and all my hope is in him. I don't want to make a choice from all possible Lods.

[untold sky]

In short

player selectionPosition ["proxy:\a3\characters_f\proxies\weapon.001", 1]; Works
selectionPosition [player, "proxy:\a3\characters_f\proxies\weapon.001", 1]; Doesn't work?

To get the same LOD, you would need to use
selectionPosition [player, "proxy:\a3\characters_f\proxies\weapon.001", 0];

You cannot just go "I'll take LOD 5"
What LOD 5 is, depends on the model. A model with 3 res lods will have something completely different than a model with 6 res lods

If its a bug, then surely I should be able to reproduce it.

If I check the same LOD (The first res LOD), then it works the same.
If I check a different LOD (The second res LOD), then it also works the same (in that it doesn't find the proxy).

Indeed the command only searches loaded LOD's, it doesn't force-load the LOD itself

it would be neat if allLods would return if a LOD is currently loaded

fun fact, selectionNames command force-loads the LOD

_name = ((player selectionNames 1) select {_x == "proxy:\a3\characters_f\proxies\weapon.001" }) select 0; // Force loads the first LOD
selectionPosition [player, "proxy:\a3\characters_f\proxies\weapon.001", 0]; // Fetches from first LOD

_name = ((player selectionNames 1) select {_x == "proxy:\a3\characters_f\proxies\weapon.001" }) select 0; // Force loads the first LOD
selectionPosition [player, "proxy:\a3\characters_f\proxies\weapon.001", 1]; // Fetches from second LOD (which might not be loaded)

[untold sky]

allLods will return whether a LOD is currently loaded.
I don't really like scripts force loading LODs, because everyone will just use lod 0 or res 1. And force load all kinds of models in their highest resolution even if they are very far away and irrelevant.
But people are probably doing that already using selectionNames

[dusky tendon]

This may well depend on the Texture settings. Or the quality of the models.

But I tried playing with super low textures and models and didn't notice a difference.

[solid marten]

maybe we should extend preloadObject to pass LOD index as well

[untold sky]

Didn't know that command existed.
Yeah that would be nice and explicit

[dusky tendon]

Now I spoke with a player on the server. And his proxy lod position is not shown as it should. (And now I’m using 5). And the other player is shown correctly.

[gaunt depot]

Wish there was a way to know if object is fully loaded, preloadCamera doesn't seem to be reliable at all

Fully loaded is pretty vague though, I guess LOD 0 is loaded? 🤔

Oh, preloadObject already returns boolean so it could also return load state for LODs, so that would let me know if objects are fully ready

I assume preloading LOD 0 doesn't mean lower LODs are loaded too, isn't it?

So you don't have to preload every single LOD one by one

[untold sky]

allLods will return if a LOD is loaded

[untold sky]

I would expect so. But Andrei's stuff shows that lod1 might not be loaded even though lod0 is

[gaunt depot]

Speaking of preloading, I wish there was a way to know if map control finished loading the textures, I wonder if there is a method that can be translated into a scripting command.

[dusky tendon]

allPlayers apply {100 preloadObject _x};// Return [true,...]

I call global and no have effect. Proxy by Visual [0,0,0].
The player who received the position [0,0,0] also did not receive selectionposition Lod 5

[gaunt depot]

ctrlMapLoaded _map to return true if map is currently loading anything or false if not not

[gaunt depot]

Try lower distance? Say 0 meters?

[gaunt depot]

I've been trying to render map controls onto a texture, but you often end up with textures like this:

[dusky tendon]

i try 10m and [false,false,false...]

[gaunt depot]

Yeah, from my experience preloadObject is not very reliable too

[dusky tendon]

we have Lod 0? I never see positions on Lod 0 (all time this return how [0,0,0])

[solid marten]

good opportunity to fix/extend then

[untold sky]

lod index 0 != lod resolution 0

lod index 0 is lod resolution 1

[dusky tendon]

it would be nice to be able to get not just a proxy, but also the object itself that is held by this proxy. I use it to get the position from where I draw the laser. But I have to store an array with all guns that have proxy a3\data_f\proxies\weapon_slots\side.001 and then calculate the position geometrically. And all this is a dance only for unit modeltowolrldvisualworld [n,n,n,]

[untold sky]

The thing that is in the proxy, is not an object that you can get

[gaunt depot]

You can get some proxies like the weapon in soldier hands with cursorObject

Wonder is that's the only exception

[cyan basin]

how i can get more details about this so i can unsubscribe the removed mod?
2024-07-07 00:19:36,064 [ERROR] 22 SteamLayer: 0.0: SteamQuery: class SteamUGCFileDownload - failed - Code: 16 - STEAM_API_DLL TODO @untold sky

[cyan basin]

have you guys considered closing registration to prevent the bot spam problem?

or restricting users

[devout wave]

The FT isn't just for the few known diehards on A3, it's also for Reforger and DayZ which have much more development and support activity. People need to be able to create accounts and use it.

[cyan basin]

ah okay, there's no eta for the migration to a better system or upgrading from phabricator to phorge which is a community supported version of the software?

[devout wave]

Didn't you have a whole conversation with Lou about that, like, last week?

[cyan basin]

i mean yeah, but nothing further than that unfortunately :(
im hoping i could hear from someone who is actually involved in the process

sorry

[outer frigate]

yeah, lou is involved in the process

and has been for months at this point

its not an overnight swap

[cyan basin]

oh i see, well it seems to be either confidential or there's no progress to share about

[cyan basin]

um, im making a ticket but could i get some confirmation on whatever the following assumption about said information was correctly guessed please? (see source of reply, tldr: whatever the error in launcher log corresponds to a unavailable/deleted item)

im not sure how to reproduce otherwise

and im looking for someone experienced in workshop mod management

[cyan basin]

um, what is the maniphest form id 2 used in/for?

[cyan basin]

i also observe two other games being tracked in FT

[cyan basin]

https://feedback.bistudio.com/T182809

[unborn acorn]

13. https://feedback.bistudio.com/T182814 Some explosion ammunition misses SoundSetExplosion={}; in configuration: no explosion adds, no echo(tail), no debris

[gray wharf]

clever girl…

[cyan basin]

lol

wait, where is the reforger ft channel?

shouldnt this channel be named arma_feedback_tracker and be put in a more general category?

[outer frigate]

no

[cyan basin]

so?

[azure crane]

#reforger_feedback

sometimes I wonder if you ask those questions while having everything righht on the plate just in spite of whatever or is Discord really fucked up so much it hides half of the channels

[cyan basin]

oh, no wonder, the channel inconsistency didnt make me think of it

[azure crane]

no offense :D

[cyan basin]

adhd shenanigans really. i dont pay enough attention or i hit a self-fluke

[azure crane]

when I need to find something I usually press Ctrl+T and start typing whatever I need

[cyan basin]

i also have all channels shown

oh interesting, didnt know about t but knew about k

[cyan basin]

its fine

[cyan basin]

i did spot the channel but i had conflicting thoughts that ultimately ended in nah its not the channel, without ever looking.

[unborn acorn]

I want suggest to add opening and landing sounds for parachute (all sounds in data but not used)
Config params soundGetOut[]={}; success plays landing sound
soundGetIn[]={}; nothing plays when player uses "open para" action
But if use getIn EH in config the sound will work

So I want to ask, is it will be easier to add sound when parachute simulation triggers animationOpen=""; and animationDrop=""; i.e. in engine not data?
and why soundGetOut[]={}; won't work when player enters to parachute, if it will work the opening sound will be so ez to add too

[unborn acorn]

Cluster bomb missed its long-range explosion sounds and using the same sound files for 'medium' and 'distance' ranges, although there are separated ones for it in files
https://feedback.bistudio.com/T182834
compare it https://www.youtube.com/watch?v=xEQGEgX-UCU
https://www.youtube.com/watch?v=a_e0d1RvgJ8

[untold sky]

Has anyone noticed that loadAbs command is broken? 😄

Put 1 bandage in container: 0.6
Put 1 uniform in container: 40
Put 1 uniform with one bandage inside it in container: 41.2 😮

The game is also bugged when checking if you can put items in container

Container can hold 7000.
Fill it up to 6959.4

A uniform with one bandage inside it should be 40.6.
Can it fit in the box? yes.
Do any more bandages fit? no.
Take one bandage out, now there is again space for one more bandage to go in. Can you put it back in? no.The box thinks its full.

Everything inside a uniform is counted twice

Put my uniform, vest, backpack on.
loadAbs player -> 648
Put all of it in a box
loadAbs box -> 1117
Manually script calculated mass of all items in the box -> 301.6

Have empty uniform, vest, backpack
loadAbs player -> 180.001
loadAbs box -> 180.002
Script calculated: 180

Put one bandage with 0.6 mass into each uniform, vest, backpack.
Expected load 180 + 0.6*3 -> 181.8
loadAbs player -> 181.801
loadAbs box -> 183.602
script -> 181.8

ahhh

[surreal bough]

the reason is that no one needed to count the mass of items in the box 🙂

[regal nimbus]

Curiously we only have one marginal use of the command in our codebase.

God knows what the arsenal uses instead.

[untold sky]

Deleting the currently selected row in a ListNBox.
Every deletion also triggers onLBSelChanged.
Except, the second last row delete. You end up with no row selected, and onLBSelChanged doesn't fire

It should be selecting the last row which is now row 0. But instead it just unselects.
Must be some off-by-one error.

Only happens when you have two rows and delete the top one.
Deleting from bottom upwards is fine

[desert trench]

would it be possible to extend https://community.bistudio.com/wiki/getEventHandlerInfo with an optional bool to log the linked code to the EH (if necessary limited to -debug) - unless i am missing something there seems to no way to figure out what is the source of an event id. the engine must know tho

[fresh hawk]

https://feedback.bistudio.com/T167469

[dreamy bane]

will this be ever look at https://feedback.bistudio.com/T167202 ?

[solid marten]

what does freeExtension return when you execute it?

[dreamy bane]

I forgot, not using Dev anymore

[solid marten]

You want fix for a command that only works on dev but you don't use dev anymore?

[dreamy bane]

yeah was no point continuing with Dev when stuff didnt work. but I also wonder if this command could be put into stable?

[untold sky]

What do you need it for in non-development use?

[devout wave]

At a guess, it's still for development but they want to develop their extension to work with the current stable branch, rather than with a version of the game that won't be released until an unknown time in the future. Given the current A3 release schedule, tying your extension to dev branch might delay your own updates by a lot.

Also it's a bit annoying to switch branches in general. Normal Steam management doesn't let you have both at once, and I dunno if the standalone game updater ever got fixed to work with Steam Guard.

[dreamy bane]

for stable dev use . and like Nikko pointed out switching between branches is not fun

[untold sky]

We could put it on profiling binary

profiling generally also gets new extension features if possible. So you can also test new things early there

ever got fixed to work with Steam Guard
It got fixed once, but then valve broke it again

[dreamy bane]

for profiling you can only download the exe? 🤔

[untold sky]

yeah

[dreamy bane]

that sounds good

[untold sky]

We could probably also push the dev branch exe to google drive, but that's a bit of effort

[solid marten]

doooeeeet

[alpine tulip]

Ignore all QA, an exe directly from Dedmen

[near cedar]

The code dealer

[untold sky]

Spam clicking on a empty combo box == game crash.

[gaunt depot]

Having dev build exe downloadable same as prof is a great idea

[cyan basin]

you can have multiple versions of the game but you'd need to use steamcmd force_install_dir every time you want to manage the installation, and things are going to be more manual, you also might need to force/pass -mod= for making use of the install relative folders.
the arma 3 project stuff while mainly for buldozer and engine tools it does fill the purpose of having a dev install i believe, there's a utility in the tools app to manage installations

[cyan basin]

but what about the data changes that get made in dev?

[untold sky]

You won't have them 🤷

[untold sky]

there's a utility in the tools app to manage installations
Yes. As he said
I dunno if the standalone game updater ever got fixed to work with Steam Guard.

[cyan basin]

what if theres code that necessitates data changes?

[untold sky]

Well then you're doomed, because you don't have the data changes

[cyan basin]

oh i see, supposedly it should work

[cyan basin]

right, then what's the point of posting executables without data if data changes are likely to happen?

[untold sky]

They are not likely to happen

[cyan basin]

would it be possible to request an appid allocation for dev builds exclusively?

[untold sky]

Yes. You can request anything.
No we won't do that.

[cyan basin]

fair, i mean it would help the situation

of having multiple installs

[cyan basin]

it could be issued on a whitelist basis for people who want/need this scenario.

anyway i do understand that at this point in the game lifecycle bi is not willing to put more resources than the minimum necessary to maintain the current setup

[untold sky]

Even more effort. How about no

[cyan basin]

okay fine, sorry for asking

i was just exploring possible ideas to tackle this

about the data changes, it could be posted in gdrive with the exe... or would it be too much storage to do?

[untold sky]

Not feasible. Won't do

[cyan basin]

because it's intensive to separately pack the data that did have changes?

tickets won't help getting trivial infrastructure changes like this implemented, correct?

[untold sky]

I already said its not feasible. We will not do that. Its not trivial.
I said its not feasible, too much effort.

[cyan basin]

okay sorry for insisting

[dreamy bane]

can we get copyFromClipboard working in editor-MP? I understand it's disabled in MP but why in editor also?

[cyan basin]

wdym editor-mp? multiplayer sessions hosted from within eden?

[dreamy bane]

yep

[cyan basin]

well, because its basically the regular client hosting wrapped into the current editing eden mission...?

[dreamy bane]

dunno whats the difference between "real" server hosting but that's how I do most of my testing

[cyan basin]

"real" server hosting is done via dedicated server, im not sure about whether missions hosted with eden have the same full network access as regular hosting.
i dont remember, anyways, i think it might be for security reasons.......

[dreamy bane]

yes i know what dedi is. will get back to this topic later

[daring wagon]

I would like to request an event handler for saving. Something like "OnBeforeSave". We only have a scripted event handler for that right now. But that can be circumvented easily.

Let me know if I should. Create a ticket for that later.

[untold sky]

Good idea yeah. I'll do that

[untold sky]

yes

[untold sky]

End of 2011, a bugfix was added to fix wrong mass calculation for backpacks.
The fix was that in function A, the mass of contents was subtracted. And in function B it was added back.

Early 2013, a bugfix was added to fix wrong free space calculation in containers.
For that the fix in function A was removed. And function B was forgotten.

This was broken since February 2013...

This does not only apply to loadAbs. load command, the UI inventory box-fillage indicator and putting items into boxes are also affected.
If you have a backpack that including contents is 365 mass (B_Soldier_A_F backpack).
Then a box with 2000 space (PlasticCase_01_large) can only fit 3 of them and then its full. But it should have space for 5 backpacks.
It actually when broken should only fit 2. Putting the 3rd in puts it to 100.5% full... Another bug

I think I noticed this bug months ago doing logistics and putting backpacks into a box, because you can fill a backpack very quickly in arsenal, and dump it into a box.
But I noticed it takes up much more space and much less items fit in the box. I thought it was the backpack by itself just being so heavy. So I went back to manually right-click tons of items into the box.

Turns out, it wasn't the backpack.

And I found even more bugs.
The preview bar in bottom left 😄

[devout wave]

Beautiful. A true classic Arma moment

[azure crane]

and here I thought it's on purpose

[untold sky]

The bar display is on purpose. Its even documented

// if draggingObj is a container
// let's get mass of a single empty item
// like uniform, vest or backpack
But it just doesn't make any sense. If dragging a filled container, why would we show a weight preview of an empty one

it already does it correctly on the player load bar. Just not on the container bar

[limpid rune]

I think zeus ammo box bar is bugged too last I looked at it i'd show a different value each time I opened the menu

[outer frigate]

those bars are all sorts of jacked up

[fast lodge]

could that explain why items dissapear while putting them into a almost full container ? A lot of players were reporting to me that they lost their weapon which they swapped into the container with a gun from the container

[untold sky]

I don't think so. They wouldn't disappear, they would over-fill the container

[untold sky]

I don't know which one you mean

[limpid rune]

this one

[untold sky]

🤔 mh. But that's scripted, I don't want to look at scripts

[limpid rune]

and naturally after I took that screenshot I couldnt get it to bug out

[daring wagon]

https://feedback.bistudio.com/T182887 OnBeforeSave

[untold sky]

in general the EH was never firing when you delete a row. Except when deleting the last one.
It updated the current selection, but it copy-pasted the function that would do it instead of calling it, and without the EH trigger.

[untold sky]

Would you accept a diag binary only?

[untold sky]

Actually we can't. Its dev-only for security reasons. So it will not appear on retail

[desert trench]

sure. the cmd mostly for debugging anyway i'd say

[untold sky]

Ugh. That's harder than expected.
This was not made to look at the handlers contents.
I'll only give you the string of the handlers contents

And not the music eventhandlers

And not projectiles

next dev I only tried mission and entity. UI's should also work. MP entity is not implemented

[daring wagon]

@solid marten Since you claimed ticket T180930, perhaps you find the time as well to look at https://feedback.bistudio.com/T180633. Both tickets are about missing defines.

[daring wagon]

Cheers, that was quick. https://community.bistudio.com/wiki/Arma_3:_Mission_Event_Handlers#SaveGame

[gray wharf]

https://tenor.com/view/michael-scott-thats-what-she-said-he-said-gif-25165146

[young oasis]

https://github.com/acemod/ACE3/blob/master/addons/dragging/functions/fnc_getWeight.sqf#L33-L38

¯_(ツ)_/¯

[untold sky]

Comment says backpack, code says all containers. Its all containers

[dusky tendon]

If we talk about the weight of the equipment. Then I remember a very interesting bug in calculating the stamina for a player’s movement and his speed modes.

1. player setUnitTrait ["loadCoef",0.nn ];
   Reducing weight to increase the stamina bar gives more stamina but does not change the overload weight. Therefore, the player has a stamina bar but cannot run.

2. Any change in weight in the player's inventory recalculates fatigue as if ["loadCoef",0.nn] had not been applied to the player. Therefore, literally all endurance is removed when overloaded. (Even when reloading weapons or changing attachments)

[gaunt depot]

I've also seen this thing. Needs a ticket.

[untold sky]

Yeah that looks like more work, for a ticket

[dusky tendon]

Can anyone create a ticket? I don't want to create an account on feedbackBis, only to create one ticket and forget about the existence of this account.

[gray wharf]

it's not like this forgotten account will haunt you at night or anything

[untold sky]

Most accounts on FT just post spam once and then forget about themselves

[rose cloak]

https://feedback.bistudio.com/T182907 infinite tinfoil hats

[daring wagon]

https://tenor.com/view/thumbs-up-conspiracy-tin-foil-hat-gif-20017163

[gray wharf]

why did you report it?? I don't want it fixed 😄 😄 😄

[rose cloak]

inb4 closed with "Feature, not bug"

[devout wave]

I mean you should definitely be able to make more than one hat from the average roll of tinfoil

[gray wharf]

yeah that
infinite, maybe not, but 20 at least 😆

[outer frigate]

maybe the scale of the tinfoil hat model needs to be increased instead 🤔

[gray wharf]

https://tenor.com/view/perfetto-yes-gif-26163077

[desert trench]

if sounds via sqf commands are to be played while you are tabbed out with -noPause (in SP only?), it seems the sounds get added to a queue and played once tab back in - can someone confirm that?

similar thing seems to happen with draw3d[Icon?] - the creation gets stacked while you are tabbed out, all drawn for all the time you were out drawn in on go, and disappear the next frame

[gray wharf]

2.10 added -noPauseAudio (if needed)

[desert trench]

well thats a workaround, but above would still be bugged

[daring wagon]

Yes..when one tabs back in, oftentimes many sounds are played at once. Super annoying as sometimes they are also very loud

[unborn acorn]

KK I can see you fix some functions, I want promote ticket about Jets DLC scripts work in MP in case it was forgotten
https://feedback.bistudio.com/T174022

[untold sky]

They are both bugs but.. So minor that.. eh

[solid marten]

too many issues looks like a lot of rewrite needed, I will ask

[unborn acorn]

imo priority in this ticket should be given to BIS_fnc_planeEjectionFX as the most used: 1) global RE the say3D, 2) write correct params to say3D and 3) do something with particles creation

[unborn acorn]

well to other issues only 1) and 2) steps needed (expect BIS_fnc_ejectionSeatRelease) idk if it can be called lot of rewrite

//from
_ejectionSeat say3D _ejectionSoundExt;
//to
[_ejectionSeat, [_ejectionSoundExt, 1400, 1, 0, 0, true]] remoteExec ["say3D"]; //BIS_fnc_planeEjectionFX

[untold sky]

beep boop, requesting feedback.
"Attached" entity event, for attachTo/detach.

If the event is fired before the attach, you can get the position of where the object was, before it was teleported to the attach point.
If fired after, you get where the object now is after attaching.

I assume the after is usually wanted.
As far as I can see, detach doesn't change position so there before/after shouldn't matter

[outer frigate]

can you not pass previous position as param if after? if fired frame before movement imo just before would be more valued

[untold sky]

Extra overhead Attaching is rare though
Extra code to write

[outer frigate]

fired after with previous position would be best imo but if that is too much work then just before would be better as we have not yet invented time travel to get position before

[gray wharf]

If fired after, you get where the object now is after attaching.
can't you just getPosASL?

[untold sky]

wha

I am not passing any position currently. But you can do it in script ofc

[untold sky]

Question is. Is position then enough, then you have position. But want orientation too

[outer frigate]

can you not just get it all yourself with getposasl/vectordir etc or am i missing something

youd just need to pass obj param surely

[untold sky]

I only pass objects.
Sure you get it yourself

But if it fires after it changed position, then you cannot get the old position in the eventhandler script

[outer frigate]

yeah

[outer frigate]

hence this

if its before position change, can get all your stuff, exec next frame and get all your new stuff

[untold sky]

Ugh before is hard. Before doing the attach, I don't know if it'll succeed

[outer frigate]

those are the best type of events 🙂

attach on each frame for random chance for event fire each frame 😄

[untold sky]

Do you also find it annoying that Attached has two T's, but Detached only has one?

[gray wharf]

nope, just you

(not that I find you annoying, more that it only annoys you, I mean, I swear yeah!)

[outer frigate]

sometimes

it depends

[untold sky]

Arma must make everything so hard

How does ViV transport unload work?

• Detach the cargo, teleport it to unload position
• Re-attach the already attached cargo at a different offset so it is floating at the unload position. Then detach the cargo.

Two choices. What's it gonna be

[gray wharf]

reattaching most likely prevents a collision, sooo yeah

[alpine tulip]

3. Just blame someone who wrote the ViV idea back when 2014

[hallow sun]

"Re-attach the already attached cargo at a different offset so it is floating at the unload position. Then detach the cargo."
If it was done like this, it's probably to avoid crates flying into your face because of #ArmaPhysics.
I can speak from experience as player with such a system (ExileMod Viv) and trying to make it better as a scripter 😆

[untold sky]

The re-attach and detach are done in the same frame. There is no physx simulation inbetween.
Doing setpos would also be done in the same frame

[gray wharf]

I expected it to be all scheduled tbh 😄

[sinful kettle]

Well if you're going to do 2 you might as well just detach and set position at unload pos + offset?

[regal nimbus]

What's the locality going to be for attach/detach?

[gaunt depot]

Make local flag for attaches

[desert trench]

@solid marten as you had looked into the respawn code for vehicles for 2.18 - any chance to look into unit respawn too?
in particular for playable AI. it seems https://community.bistudio.com/wiki/setPlayerRespawnTime is only for players/clients

only workaround i can see would be: addMissionEventHandler ["EntityRespawned", + if (not (isPlayer _x)) then {_x setDamage 1;}; but thats super bad for multiple reasons

[untold sky]

Global probably.
Remote will call attach/detach via network message. Its a very low-level eventhandler

[untold sky]

Did you know that Arma will not create a crash dump if it crashes due to a stack overflow (infinite recursion usually)?
Fun. It will crash inside the crash reporting code because there is no space 😄
Finally solved the mistery behind crashes that don't create dumps

[gray wharf]

but is it avoidable?

[untold sky]

I found out how to reserve a bit of stack space.
But the crash reporter does lots of memory allocations (which needs more stack space than we have, and also would crash if memory allocator is in some way corrupted or broken)
Lots of rewriting to be done 😄

Did you know that on game crash, Arma copies game version information into clipboard?
It includes a list of all loaded Addons (pbo's)
To do the formatting for that it only needs a couple kilobytes of stack space and several dozen memory allocations.
While we are handling a game crash where everything could be broken/unavailable. Ugh

[devout wave]

Wanting to change that because it's a bad use of resources 🤝 wanting to change that because I hate unprompted clipboard overwrites

[untold sky]

I wanted to fix it. But getting enough memory space for all mods, without allocating memory.. nah.
I didn't even know that feature existed, so I doubt anyone was actually using that for anything productive.
All the info is printed to RPT anyway

[outer frigate]

can we get rid of arma copying errors to clipboard at game start too

[untold sky]

It seems to be some windows problem telling the game that its in focus when its not.
We could disable it for all errors, but I don't want to do that

[outer frigate]

launcher option to disable it for all errors? 🥺

[untold sky]

This huge block of all loaded pbo's.
that is always so super long.
Its so super long, because the line width was limited to 80 characters.
We could stuff it all into one line and get it out of the way. Not that anyone cares about that blob of stuff anyways

[daring wagon]

Yeah that's annoying. I wish we had error levels printed at the front so we could filter stuff more easily.

[untold sky]

I think I'll probably compress it into a data URL.
Then if you want to have the data in it which basically no-one wants, you can paste it into the browser

[regal nimbus]

If you load CDLC using ModLauncherList in the arma3.cfg, it doesn't show as "official" in getLoadedModsInfo. This isn't intentional, is it?

[untold sky]

probably not

Isn't ModLauncherList some old legacy Arma 2 stuff that shouldn't even exist anymore? 🤔

[regal nimbus]

It's used by the Arma 3 Unix Launcher.

Found out eventually because we have some code that depends on the "official" flag being correct.

If you load mods using the "expansions" dialog inside Arma 3 it also uses it.

[untold sky]

Ticket

[regal nimbus]

https://feedback.bistudio.com/T182946

[cyan basin]

sounds more like something for third party launchers to hook onto

yeah, you can still manage mods the old way without specifying -mod= at all and using the expansions menu, list gets saved in arma3.cfg

[untold sky]

Turns out we have no deflate compression.
Only je old LZ which is blocking that a bit.
Maybe some other day, for now I just put the big addons list into a single line.
So a crash report now looks like.

And, we can now dump dumps on stack overflow 🥳

[desert trench]

-connect= -port=0 -port=0

could this cause issues?

https://cdn.discordapp.com/attachments/1263788240364638228/1263788320794742887/VID_20240719_000434513.mp4?ex=669b8217&is=669a3097&hm=1f4cdb1396b2e11bdba7477da895ffa6fcee1ce0fec0edeb94a11d84633b5315&

like to break campaign hosting

nothing special is found in the rpt - mission doesnt get even started

[outer frigate]

Is there any chance of getting renderer changed to display normal details in shadows?

[untold sky]

don't understand the question. Where does that come from, it might cause loading screen at game start

[desert trench]

this guy: #spearhead_1944 message cant launch the campaign in SPE - but normal missions just fine

I'm using the normal launcher from steam
And I've had trouble with the ports and multi player since me and my friend didn't have our arma names the same as steam which seems to have caused issues but idk if that could be causing issues

as said the rpt showed nothing beside the unusual startup parameters

as the hosted campaign was hacked into the engine, its hard to say whats going on there/whats the specific issue

[desert trench]

removing -port=0 -port=0 fixed it

[dreamy bane]

is it a bad idea to have some new net commands to be able to communicate between server and client currently in the lobby? not to run any commands really but just to exchange some data

so you could do like ```sqf
[true, false, 123, "string"] sendToLobby;

[gaunt depot]

publicVariableLobby STRING

[outer frigate]

do scripted events not work in the lobby

[dreamy bane]

not a whole lot works in lobby

[hushed birch]

~~not sure if this is a bug, but cant use commands like ctrlMapAnimAdd in a RscMapControl created with titleRsc or cutRsc (RscTitles) #arma3_scripting message

It only happen ingame, on debriefing dialog, its seems to work~~

i was executing cutrsc on each frame 🙂
i dont know how to delete a ticket 👉 👈

[limpid rune]

https://feedback.bistudio.com/T182166 hope this can be looked into at some point, it's pretty bad that launcher effects just straight up don't work in MP.

it does seem related to the magazine being deleted when fired and presumably some 'check' in how gunparticles are spawned blocking it

my guess is similar weapons with the magazine being deleted upon firing (like underbarrel grenades) would have a similar issue

[hallow sun]

Any reason why MissionProfileNamespace doesn't save automatically on leaving a game/server like profileNamespace does ?

[regal nimbus]

Is that confirmed behaviour?

Our usual issue is that missionProfileNamespace doesn't work correctly on Linux servers, but the exact mechanics are unconfirmed.

but then we use saveMissionProfileNamespace and it still doesn't work.

[hallow sun]

Almost sure yea, gonna repro in vanilla to be 100% sure

[deep plaza]

Why does hideobject command hide the players geometry lod making the player go through objects, i mean what is the purpose behind this? and could we get a "fix" for this?🫣

[devout wave]

It does the same thing to any object you use it on, player or not, and that's just...what the command is meant to do. Player/notplayer inconsistency would be bad (and likely unreliable as units can stop being players or suddenly start being players).

Changing it would be non-backwards-compatible.
What's really needed is a separate new command that only hides visual/shadow LODs and leaves collision alone.

[deep plaza]

That would be amazing if we had a command like that

[outer frigate]

arent we getting the reverse of that for physx objects

[devout wave]

I think I heard something like that

[deep plaza]

Should I create a request ticket for such a command or would it not receive any attention?

[deep plaza]

What does this mean?

[devout wave]

https://community.bistudio.com/wiki/setPhysicsCollisionFlag

[deep plaza]

Does this mean that you can enable collision for a hidden player?

[solid marten]

yes it is related to magazine removal, but this is not why effect is not seen, it is started fine on remote client, but then updates comes and remote client removes and reads all its weapons thus making effect to abort.

[limpid rune]

ahhh

[untold sky]

I would like that.
But I would need a ticket, and also a use-csae/repro script. Where you would receive such data, in a script running inside the lobby screen

[untold sky]

Looks like a bug.
The save call at game shutdown was missed. Ticket.

[untold sky]

Afaik @sinful kettle already had exactly that ready to go? So won't need a ticket then

[dreamy bane]

cool. i dont know either whats the best way to receive the server sent data. not sure if event handlers can be used in lobby. Maybe just call a function like onServerData everytime server sends data

[untold sky]

The main problem for me is getting any script to run in lobby

If you can get a loop running, that can read public variables. That'll be enough.
We'll then figure out what we do, eventhandler or just allow publicVariable to send there.
Maybe remoteExec

[dreamy bane]

maybe the onServerData function would be in uinamespce function because missionNamespace probably doesnt work there

[untold sky]

It should work there

Its either cleared before or after lobby. That's something to find out.

[dreamy bane]

calling onServerData in client would probably be simplest and you can access the UI from that function

then no loop is required

[untold sky]

Same would be if remoteExec could call a function there.
But you first need to register the function/eventhandler

[dreamy bane]

I thought remexec wont work in lobby..

[untold sky]

it doesn't now

[dreamy bane]

ok

this is how I would "normally" call UI function ```sqf
call (uinamespace getvariable "guiTestFn")

guiTestFn defined in with uinamespace do {};

[untold sky]

Just give me a ticket with example how to execute code in lobby screen.
I need code running in lobby screen that can set a variable. The rest I can figure out

[dreamy bane]

ok will put something up

[untold sky]

That worked nicely.
90kb of #arma3_feedback_tracker message That.
Turns into a single line url that when pasted into browser gives this

With 2.20 most likely this text->dataURL compression will also be available to script. Though not the decompression way.

[hallow sun]

https://feedback.bistudio.com/T183037

[dreamy bane]

@untold sky I did my best describing the best method I know how to run code in lobby but I'm unsure if this will help you: https://feedback.bistudio.com/T183039 I hope that is of some use. let me know if you need anything else

[outer frigate]

initLobby.sqf event script? 🤔

[untold sky]

The client hasn't received the mission file at that point. Its only downloaded after slot is selected

[outer frigate]

ah ofc

[desert trench]

Invalid turret path

possible to get this in public build and with more info

we could never figure out what this is about (with tanks/vehicles)

[sinful kettle]

It just means that the path it's trying to use is empty/too big/out of bounds index
Getting a turret by path is used by many operations involving turrets, e.g. weapons/magazines/pylons, R2T, MFD, SQF code, etc.
It's not possible to tell what it is about without passing a context to it too to print with the error

[desert trench]

too big in what sense - turret layer depth?

or some index value of some of the turret specific paramters

gunnerCompartments = "Compartment5";
is another weird one. the engine doesnt like Compartment5 - instead of taken from model.cfg(?) it seems some hardcoded enum system instead

[deep plaza]

It would be so cool if we could get a command like this. I wanted to create a vanilla prop-hunt game mode and this command is essential🙏🙏🙏

[sinful kettle]

Yeah depth (as in array size; turret paths are arrays, so e.g. if your turret depth is only 2, using [0,0,0] causes that error)

[sinful kettle]

From what I see in the code it expects number there? Try just 5 (or maybe 4 since it's index and afaik component numbers start at 1, which is index 0)

[sinful kettle]

No it only worked on geometry/physx lods

But you already added that anyway 😅
Oh you mean the ticket. nvm

[sinful kettle]

Come to think of it, using 4 should be correct since in the game config it lists "compartment1" which just parses to number 0, which incidentally is the same as the index 😅

[alpine tulip]

Wild and workthought: a diag command/mode to check... drawcalls or something? Graphic performance?

[untold sky]

only 2.20 btw. So it'll be a bit before I start looking at it

[untold sky]

There are config enums, maybe Compartment4 is a enum, and it is able to turn it into a int cleanly. And Compartment5 isn't.

[untold sky]

diag_enable "PerfInfo" might be useful, I don't know if thats available in diag.
It shows gpu "usage"

"ObjPassCount" shows render passes, which is essentially draw calls

[alpine tulip]

At least not on BIKI, let me try on Diag

Both are invalid

[untold sky]

Oh there is a flag whether something is allowed in dev-branch diag binary

I don't know why these are off

Might be that they don't work in that build, don't have time to test that

The pass count seems to be draw calls indeed

I'll turn them on in next dev-branch.
As far as I can see the PassCount should work fine. The PerfInfo I don't know

[alpine tulip]

Just checked using diag_list and
["Force","Animation","CursorTargetObject","Shots","Mines","HitPoints","AnimSrcUnit","AnimSrcTarget","Suspension","EPEForce","EPEVehicle","ParticleNames","Particles","SurfaceUnder","AISkill","AIDriving","AIAiming","AIBrain","AIDynamicError","AISuppression","DynSimGrid","DynSimEntities","DynSimGroups"]
Yeah it's not there

Sounds a good news to have this. Thanks!

[desert trench]

ok. so if you use Compartment5, it wont get converted, the engine doesnt like the string - so defaults it to 0?

[untold sky]

probably yes

[untold sky]

Ugh. There are about a dozen of these. And considering they are now silent on retail builds. We might cause some RPT spam with that.
And in most cases (except triggered directly by a script command) you wouldn't get useful context out of it.

[daring wagon]

Debug only?

[untold sky]

internal binary and diag only

[daring wagon]

Good enough

[untold sky]

OH! .. Well I guess yeah.. duh. Making it debug only is easy

[limpid rune]

some data changes will be needed then, as Compartment5 is used in vanilla too:
\a3\air_f_heli\Heli_Transport_04\config.cpp
line:1456

[desert trench]

alright. basically looking for a way to verify our vehicles. some tanks create that entry when createVehicled

[untold sky]

You'll have it in internal build with context tomorrow

Found it. Yeah "compartment1" and "compartment2", ... are local variables in the config parser.
Not sure where they come from. But it will be an enum, or maybe a __EXEC() setting a variable

[limpid rune]

weird, wonder why that was done

[untold sky]

The game is bamboozling me.
I cannot see anything ever setting that variable. But it is there in the end 😠

[untold sky]

raises fist mumble mumble grrr
"bin\config.bin"

Up to compartment32.. But not even Compartment5 is set

[untold sky]

I can't find that
Line 1456 is item_xx(Medikit,1);
That config file only uses Compartment1
And I cannot find any compartments in any other included files there

Ah there it is ffv bench
CargoTurret_04
gunnerCompartments = "Compartment5";

[limpid rune]

sorry must have got the number wrong

[rose cloak]

// for dedmen to fill eventually

[limpid rune]

actually probably debinned vs source or something lol

[desert trench]

is anything being worked on that could affect custom mission end types to no longer work?

or does this cause it?

15:51:15 Error in expression <criptionParams];};
        _descriptionText = format ([_descriptionText] + _descriptio>
15:51:15   Error position: <format ([_descriptionText] + _descriptio>
15:51:15   Error 0 elements provided, 1 expected
15:51:15 File \A3\Ui_f\scripts\GUI\RscDisplayDebriefing.sqf..., line 45```

the function has:

        if (_isDebriefing) then {
            /////////////////////////////////////////////////////////////////////////////////

            //--- Static debriefing
            if !(isnil {missionnamespace getvariable "RscDisplayDebriefing_noise"}) then {
                ("BIS_fnc_endMission_end" call bis_fnc_rscLayer) cuttext ["","plain"];
            };```

so might - will verify via console

hm seems not related

"end_ger" call BIS_fnc_endMission;

[untold sky]

yes

[desert trench]

rpt:

15:56:57 "BIS_fnc_log: [BIS_fnc_endMission] Mission completed with 'end_ger' ending."

no error raised, but fade to black goes fade in again

[untold sky]

That format error you posted there, is being fixed. That's what we're working on.
The error itself doesn't stop the script from working

[untold sky]

I have seen that a week ago, fixing a crash bug on mission end.
I couldn't tell if thats a bug or intended. Because previously the game just crashed so what was supposed to happen

https://feedback.bistudio.com/T182375 yeah here it was

I'll need a repro that doesn't crash before the fix, to see what went wrong there

[desert trench]

https://community.bistudio.com/wiki/endMission these listed here still work (albeit killed is weird as it leads to couple seconds blackscreen - edit: infinite black till you hit ESC)

[untold sky]

Maybe custom ones broke in the enum rework recently. I'll need a way to repro

[desert trench]

BIS_fnc_endMission* only seem to call endMission or failMission. probably the custom end names are the prob

repro:

• SP
• showcase
• IDAP
• start
• hit space twice
• execute endmission "Behavior" or ["Behavior", false] call BIS_fnc_endMission; (for visual feedback)

i guess at least

class CfgDebriefing
{
    class End1 
    {
        title = $STR_A3_CFGVEHICLES_MODULEENDMISSION_F_ARGUMENTS_WIN_VALUES_WIN; //TODO: nicer phrase?
    };
    class FriendlyFire
    {
        title = $STR_A3_Orange_Faction_IDAP_CfgDebriefing_FriendlyFire_title;
        subtitle = "";
        description = $STR_A3_Orange_Faction_IDAP_CfgDebriefing_FriendlyFire_description;
        picture = "A3\Ui_f\data\GUI\Cfg\Debriefing\endDeath_ca";
    };
    class Behavior
    {
        title = $STR_A3_Orange_Faction_IDAP_CfgDebriefing_Behavior_title;
        subtitle = "";
        description = $STR_A3_Orange_Faction_IDAP_CfgDebriefing_Behavior_description;
        picture = "A3\Ui_f\data\GUI\Cfg\Debriefing\endDefault_ca"; //TODO: Custom picture?
    };
};```

the mission itself doesnt seem to use the custom endtypes itself - shouldnt matter tho

74                 ["Behavior", false] call BIS_fnc_endMission;
112                             ["Behavior", false] call BIS_fnc_endMission;
133             ["Behavior", false] call BIS_fnc_endMission;```

alternative would be the Helicopters showcase with [[missionName, 2],false] call Bis_fnc_endMission;