arma3_feedback_tracker

[sinful kettle]

no I just want to see the new spheres

[sinful kettle]

hmm. that should fix it
I sent a "fix" to Dedmen for this a while ago but he said let's wait until after 2.20
@untold sky

[snow cairn]

so whats the fix?

[sinful kettle]

the fix would just use the model BB instead of lod BB

for this issue. I thought it only affects animated objects but looks like not

[snow cairn]

just for the carrier model or would it apply universally

[sinful kettle]

universally (engine code change)

the LOD bbs are borked for some reason. they don't account for anims so they're too small. but this makes 0 sense because it has no anims

[snow cairn]

as someone that doesn't do modelling, this is eye opening.

[sinful kettle]

ofc this "fix" has downsides too. some model BBs are much larger than LOD BBs which can slow down calculations. that was actually one of the reasons why Dedmen didn't like it

but the only alternatives are:

1. leave the bug be
2. recalculate model bounds at runtime

both are bad

[snow cairn]

also putting concrete panels over where you'd fall through lol

[sinful kettle]

the slowdown will be pretty insignificant imo

[sinful kettle]

well that's just a "fix" for this specific case

many other models are still affected

[hallow sun]

I had the same problems with the USS Liberty for players walking inside the ship some years ago, had to place a floor under two seams.

[true blaze]

Dunno i used script i posted in wiki for testing. I wasnt able to find seams but tbf i didnt look very hard. One could pull the carrier parts towards the base for 1 cm to fix that.

But would have to scale overlap for each part lenghtwise otherwise youd just end up with seam in another place.
Or a dirty solution would be to use setobjectscale and make each part slightly larger

Can you tell me what rotation was used where seam was most noticable?
Basically pos/dir vector?

[dreamy bane]

which carrier are you trying to spawn?

[true blaze]

Freedom, vanilla

[dreamy bane]

hmm isnt that just "Land_Carrier_01_base_F" createVehicle _pos;

[true blaze]

No. Carrier is assembled from parts. Script for that executes on init of base part.
Issue is with seams between said spawned parts of ppl falling thru

[dreamy bane]

ive never had problem with that. here's my full code, I think the last line here is important: ```sqf
_pos set[2,0];
private _carrier = "Land_Carrier_01_base_F" createVehicle _pos;

_carrier setposASL _pos;

_carrier setDir _angle;

[_carrier] call BIS_fnc_Carrier01PosUpdate;

[true blaze]

Ofc its important, otherwise you just roate base and actual parts stay the same.

Here is an exaggerated example of bad stitching, spawn carrier in eden, and set rotation vector to [43.124, 21.563, 24.097]. Observe the results ingame

[dreamy bane]

ok i guess using other than setDir is problematic then

[true blaze]

Even when using setDir, but above really brings the issue to light. Apparently seams are still big enough, that you can fall thru, if you walk on seam midline

While function i posted seems to fix the large offset, small seams, that make you fall thru still remain.

[dreamy bane]

weird, ive never noticed such problems

[gray wharf]

go with 90° angles

it has been discussed somewhere very recently, it is due to a bad scripting

[true blaze]

Its really hit or miss, but when it happens its annoying. Worst case you fall thru, best case you skydive and break legs

[gray wharf]

setDir 0/90/180/270 for best results

[true blaze]

It also appears to be bounding issue. Ill take another look and see if i cant remedy it

Hence id like his pos/rotation as this one seems very noticable ^

[gray wharf]

if too far away from 0 0 0 then yes, most likely the float precision thing

[true blaze]

This aint looking "too far". And as Leopard explained, error would be relative, and in order of mm, not cm

[verbal harness]

Might have fixed the math

[true blaze]

Set vector dir and up instead of bank roll?

[verbal harness]

[true blaze]

Welcome to the club
#arma3_feedback_tracker message

But seam issue persists

[alpine tulip]

Nice

[verbal harness]

What is that rotation? I can't seem to find any seams big enough to fall over/into after fixing the rotation

[alpine tulip]

x, y, z in Eden = 0, 341.454, 0

[true blaze]

Id also like to know. Hypoxic found it here
#arma3_feedback_tracker message

There is also visible z offset. So i wonder if mem points themselves are also crooked. Ill check in the evening

[sinful kettle]

this has nothing to do with the seams (as I already explained before it's related to small bounding spheres)

[sinful kettle]

yes but rendering glitches can still happen

even if you do fix this seams issue by overlapping the models, you still do get glitches (this time z fighting)

[unborn acorn]

Not only airplanes can give a sound too big delay on approach to player but also any sound from SoundSet. A large value of the audibility range and fast movement of the object will give non normal delays it's sound for players, which depends on the simulation of the speedOfSound if turn it off we can notice some changes
https://feedback.bistudio.com/T124855#2818171
https://www.youtube.com/watch?v=DC0sZzSvWRI

[true blaze]

Ill just check if mem points are aligned. Might just aswell make it look good regardless if ppl fall thru 😄

[daring wagon]

Since we have addonFiles why don't we have missionFiles Giving a list of all files inside the mission folder, security?

[true blaze]

@snow cairn After fiddadling with it a bit more, and finding places where i could reliably fall thru, i found a working but dirty solution.
If you dont mind rendering glitches along the seams, it works, i ran up and down along the seams for some 30 minutes and didnt fall thru.
initCarrier.inc

params[["_carrierBase", objNull]]; 
private ["_dummy", "_pos"]; 

//create and assemble the carrier on server 
if (!isServer) exitWith {}; 

private _carrierDir = vectorDir _carrierBase; 
private _carrierUp = vectorUp _carrierBase; 
private _carrierPos = getPosWorld _carrierBase; 

// For overlap scale adjustment, 1.007 seems to be lowest to prevent fallthru
private _scale = 1.007;
private _zOffset = (_scale - 1)*20;

private _cfgVehicles = configFile >> "CfgVehicles"; 
private _carrierParts = (_cfgVehicles >> typeOf _carrierBase >> "multiStructureParts") call bis_fnc_getCfgData; 
private _carrierPartsArray = []; 

{ 
    _x params ["_dummyClassName", "_dummyPosName"]; 
    
    _pos = _carrierBase selectionPosition _dummyPosName;
    if (_dummyPosName find "hull" == -1) then
    {
        // Compensate for scale
        _pos = _pos vectorAdd [0,0, _zOffset];
    };
    _pos =  _carrierBase modelToWorldWorld (_pos); 

    _dummy = createVehicle [_dummyClassName, _pos, [], 0, "CAN_COLLIDE"]; 
    _dummy setPosWorld (_pos); 
    _dummy setVectorDirAndUp [_carrierDir, _carrierUp]; 
    _dummy setObjectScale _scale ;
    _dummy allowDamage false;
    
    _carrierPartsArray pushBack [_dummy, _dummyPosName]; 
} foreach _carrierParts; 

_carrierBase setVariable ["bis_carrierParts", _carrierPartsArray, true];

Also note this will only work close to sea level. If you wanna make flying carrier you will still drop thru.

[unborn acorn]

~~ https://feedback.bistudio.com/T191859#2818230 spam. Ticket can be closed no bug see #arma3_feedback_tracker message ~~

[gaunt depot]

Maybe config (or even script?) set radius to hotfix the issue where it matters without touching the models?

[sinful kettle]

there are too many cases and if you ask me they all matter.
script won't work because it's a model parameter and models are shared.
also config fixes won't fix the models (e.g if you make super simple objs)

[alpine tulip]

This is real random and wild idea: is it still worth of effort so a proxy can have multiple P3Ds/objects?
My usecase: more slots for character cosmetics

Yeah, the answer will be a big "no" but...

Also one thing... Will you guys do model data fixes? Do you guys (or reyhard himself) prefer to have such reports? Like Taru Medical (IIRC)'s wrong texture, CTRG uniform's wrong hand weight, US Parade Suit have unretexturable parts, etc. I think I have never heard a yes or no answer

https://feedback.bistudio.com/T168411
Speaking of a model, this still is a big killjoy

[sinful kettle]

you mean like a chain of proxies? I thought this was already possible?

[alpine tulip]

I don't mean proxy in proxy, but having more than one proxy model in one proxy. Legitly I have no clue how a proxy works tho

[sinful kettle]

I thought this was possible too

[alpine tulip]

So basically, having more cosmetic (mask with glasses, hat with headset etc) is easily possible

[gray wharf]

2.16?
diag_testScriptSimpleVM <code>, returns nothing?

[daring wagon]

nothing makes no sense I guess

[gray wharf]

I believe it prints in logs directly, but I may be wrong

waiting on Dedmen (in #community_wiki?) for info

in doubt I said it returned an array
https://community.bistudio.com/wiki/diag_testScriptSimpleVM
we'll see

[untold sky]

It returns array afaik.
I think it's disg only

[hot lark]

Two very easy to fix issues:
http://feedback.arma3.com/view.php?id=25091
http://feedback.arma3.com/view.php?id=27442

[daring wagon]

https://feedback.bistudio.com/T191578 any news on this? (Can't group entities in Eden via script)

[alpine tulip]

Strange idea: alternative format where take args like:
["%ArgFirst, %ArgSecond", "%ArgFirst", "Hello", "%ArgSecond", "World"] so it prints "Hello, World"

[regal nimbus]

I don't get it.

Oh, like replace string #1 with #2 etc?

[alpine tulip]

Jup

[regal nimbus]

Isn't that what regex is for

[alpine tulip]

🤔 maybe

[fallen elk]

Unintended behavior from getOrDefaultCall or is it basically the same limitation as event handlers?

createHashMap getOrDefaultCall ["something", {
    if (true) exitWith { -1 };
    1
}];
// nil

[regal nimbus]

I asked before, apparently it's intentional :/

I don't understand the logic though.

I felt like getOrDefaultCall could really do with a magic variable with the key name too.

[fallen elk]

import it is I guess

[hollow garden]

Hello admins of Arma Official Servers. 2 Years ago i played on official Servers of Arma 01 and 03.
There were available in the past:

1. UAV,
2. Drawing on the map,
3. captured Verhicles of green fraction did not disappear
4. also killed green soldiers disappear very fast...after 5 min. if i want to take their weapon, i dont have a chance

My questions are:
It was very interesing gameplay with those 3 aspects, which i wrote. Is it possible, that those 3 aspekts you, admins,
turn on again on official servers 01 and 03?

Reference point 1
-without UAV, on BLUFOR side, you can not use rhino shooting by laser any more
in comparison to opfor, angara is much stronger as rhino, if you play with rhino like with angara.
-also you can not drop bombs on laser of UAV.

Reference point 2
-if i order launcher or other weapon, i can not find it anymore, because i can not put a point on the place, where i got it.
-i can not make drawing informations for me and my teammates

In total i buyed this game with all these features. Now you, admins, disable these features. Why i should have disadvantages because of your changes.
If you do such changes, i want to get also a share of price, which i payed for this game, back.

Give please Information when exactly you make the changes back.

Thank you very much for your attantion!

[gray wharf]

why are you crossposting?
also this channel is not for reporting, use https://feedback.bistudio.com/ instead

[hollow garden]

thank you

i posted because admin HorribleGoat wrote: Feedback tracker would be the place to post suggestions.

[gray wharf]

ah yes! the feedback tracker itself ^^

In total i buyed this game with all these features. Now you, admins, disable these features. Why i should have disadvantages because of your changes.
it is a mission design decision, you can still play on other servers
in the end Dwarden will or will not adapt this as he said

[hollow garden]

i know, that i can still play on other servers. but there is not a server like 01 or 03 and has these 4 points, which i wrote above. So it is very important for the gameplay, that these 4 features will be adapt. 5 players (who play on servers 01 and 03) more told the same problems and does not like, that these changes implemented. Because of it, i decided to write.

[desert trench]

@hollow garden there is also #arma3_official_servers

[hollow garden]

thanks

[untold sky]

It has been in profiling branch since v31

[untold sky]

I'd say because no-one thought about it.
But also, I'm pretty sure a mission is an addon. You just don't know the prefix to pass to it, we can just make it a special argument to addonFiles to get it.
I'll take a look later

[sinful kettle]

really?
@snow cairn so give profiling branch a try (if you didn't do it then)

[daring wagon]

Mission in form off the mission folder while in editor. Just to be clear.

[snow cairn]

I've been on the profiling branch. Remember we said try it out back on stable and I had to go back?

But I'll go back and sphere it if you want on profiling. I don't remember if we sphered on profiling, but I found the issue first on profiling.

[untold sky]

But why?

Is that not the same as
format ["%1, %2", "%1", "Hello", "%2", "World"]
Which you could just...
format ["%1, %2", "Hello", "World"]
I don't see the point

It would let you provide arguments in different order than just first, second,...
But, you can use whatever script to sort the arguments first too

[untold sky]

smart!
I guess it would be in _this?
but that's dangerous could break existing code. Same for _x
But I would hope people just don't do that 🤔
I don't want to invent a new magic variable name

[alpine tulip]

Yeah, pretty much the point I tried to make were you told and also pretty much not a good improve

[untold sky]

The bounding box fix thing, will not be in 2.20 v0 though, it'll be back in v1

[sinful kettle]

maybe the fix doesn't work with roadway LOD then (it probably checks the BB in a different function)

[vocal abyss]

Obviously _key 👀

[gray wharf]

is 2.20 multithreaded? 🏃‍♂️‍➡️

[untold sky]

I dislike that only because someone will make an entry on the "magic variables" wiki page about it

[dreamy bane]

*only the elite will know the magic variables *

[verbal harness]

Can we have a steamID check for Lou and only run it on 1 core in that case?

[uncut briar]

_this WILL break existing code.
example:

fnc_example = {
    private _fnc_getArmor = {
        // wrap in call to prevent issues with exitWith in getOrDefaultCall
        call {
            params ["_item", "_hitPoint"];
            // ...
            // ...
            getNumber (_entry >> "armor") // return
        };
    };

    vgm_c_medical_armorCache getOrDefaultCall [_this, _fnc_getArmor, true] // return
};

[untold sky]

Well not in that specific case.
Because _this being the key, would again be _this

[gaunt depot]

Make new commands that include magic keys and leave old ones untouched

[uncut briar]

yeah, but you get the gist

[gaunt depot]

getOrDefaultCallKey

countIndex with _countIndex, applyIndex with _applyIndex

Or just _index or even _y

selectIndex too

[untold sky]

I wanted to turn _forEachIndex into a command, that fetches the index.
Saves us from having to set a variable, so makes it faster if you're not using the index, and starts working in every loop. Even while loops could get their iteration counter
But, breaks people who do _forEachIndex = 5; Which shall be forbidden anyway

[alpine tulip]

Just a thought actually
configfile >> "CfgVehicles" >> "MBT_02_base_F" >> "Turrets" >> "MainTurret" >> "animationSourceBody" is "mainTurret"
But is it possible to get "otocvez" from "mainTurret" using a script command? So basically fetching a selection name from an animation name, useable for attachTo

[untold sky]

2.20 changelog says

Tweaked: setUnitLoadout will now abort if a unit is in the process of switching weapons - https://feedback.bistudio.com/T167015
@fast lodge

[fast lodge]

yea that must be it I guess

[untold sky]

A second change to setUnitLoadout is that it will no longer delete/recreate the backpack/vest containers when you replace them by one with the same type

[tropic jewel]

Something like getSkeletonInfo

[untold sky]

@fast lodge according to ticket.
Add a diag_log isSwitchingWeapon
https://community.bistudio.com/wiki/isSwitchingWeapon
before your setUnitLoadout, if it returns true, when the applying loadout fails ,then thats that

[fast lodge]

Thank you , will do !

the weird thing is that the units I placed are playable civ units without weapons, I guess when the player starts taking control over it does an switch Weapon animation

[true blaze]

What about using vector dir and up stead of pitch bank roll?

[untold sky]

I don't know anything about that

[true blaze]

My comment of this ticket:
https://feedback.bistudio.com/T124858

Rotating carrier along x,y,z will break it (like screenshot below), well fragment carrier. Solution in comment prevents that.

https://cdn.discordapp.com/attachments/105465701543723008/1379528656061206668/2025063.JPG?ex=68525dee&is=68510c6e&hm=e495f4ecc6b3924f80bdd04df263ab650820962b6ab4d8c98e9170a131a6c946&

[untold sky]

I don't know I'm not involved I don't have time for that

[true blaze]

Ah...

[verbal harness]

I'll deal with it when im dealing with other script errors for 2.22

[scarlet sage]

Is it intentional that these little lights on the locked doors are visible from space (but not streetlights and stuff)? I'm guessing no, but I don't have time for a real bug report on the tracker.

Also, now with the illumination flares, I feel like it's weird that the mortar has more flare rounds than smoke rounds.

[fast lodge]

flareMaxDistance is probably set to high

[meager radish]

looks like when u start control the player without any weapons isSwitchingWeapon fires true. Idk why

And after 5-10 seconds this fires false and all works fine

So any chance for "force apply" arg in setUnitLoadout?

[dusky tendon]

Why, after all these years, is there still no command that brought back the one who activated the mine and can detonate it? (in this case, a demolition charge)
Of the corpses present, only one FIA engineer was clearly robbed of one charge. And I definitely haven't seen him personally install it. And it is impossible to find out the owner of the explosives at all.

[scarlet sage]

Use getShotParents (allMines select 0). Does that not work? Maybe you could use a script with getAllOwnedMines if not.

[dusky tendon]

yes its work thx.
I forgot this command. activated mines = shot by weapon "Put"

[final isle]

Why does the blackfoot have DARs now instead of DAGRs? (after 2.20)

[last reef]

Kinda maybe dumb question but with regards to v2.20 is that on profilining branch or on main one? What gives better preformances now?

[alpine tulip]

Both are 2.20

[last reef]

So in theory both should have same preformances?

[alpine tulip]

As for now there shouldn't be much difference

[near cedar]

Use #warlords_discussion or similar channels e.g. #arma3_official_servers

[verbal harness]

and answer is no for now
#warlords_discussion message

[untold sky]

I cannot reproduce that in singleplayer or multiplayer.
With direct spawn at start or respawn screen at start. You must be doing something special, but you are the second person to report that.

[unborn acorn]

Ticket about incorrect value of fuse distance for SPG-9 ammunition https://feedback.bistudio.com/T192706

[true blaze]

Big love

[desert trench]

steam overlay, steam fps and steam key actions still dont work in 2.20 for me. anyone else still with this prob?

[gaunt depot]

Call the command differently so old behaviour remains? iteratorIndex or something

[limpid rune]

I think that'll defeat the purpose of it because the variable would have to still be made which is the peformance issue

[gaunt depot]

How on earth ctrlTextWidth can be -nan(ind)? Maybe add a failsafe inside the command to return 0 in such case or something?

Or force recalcuate the width

    _ctrl_line ctrlSetStructuredText _text;
    ctrlTextWidth _ctrl_line;
    private _line_width = ctrlTextWidth _ctrl_line + KILLFEED_WIDTH_EPSILON;
    if(!finite _line_width) then {
        diag_log text format ["line_width somehow is not finite!!!"];
        diag_log text format ["ctrl_line = %1", _ctrl_line];
        diag_log text format ["line_width = %1", _line_width];
        diag_log text format ["ctrlTextWidth ctrl_line = %1", ctrlTextWidth _ctrl_line];
        diag_log text format ["ctrlTextWidth ctrl_line + EPSILON = %1", ctrlTextWidth _ctrl_line + KILLFEED_WIDTH_EPSILON];
        diag_log text format ["ctrlPosition ctrl_line = %1", ctrlPosition _ctrl_line];
    };
```=>

15:49:41 line_width somehow is not finite!!!
15:49:41 ctrl_line = Control #-1
15:49:41 line_width = -nan(ind)
15:49:41 ctrlTextWidth ctrl_line = -nan(ind)
15:49:41 ctrlTextWidth ctrl_line + EPSILON = -nan(ind)
15:49:41 ctrlPosition ctrl_line = [0.840027,0.0487704,0.159973,0.0364]

Not creating a ticket because no way this can be reproduced

Since I use a lot of ui2textures in that structured _text my guess the bug stems from some bugged ui2texture texture

[untold sky]

simpleVM already gets rid of _forEachIndex if you're not using it. But SimpleVM is quite limited

[untold sky]

Why not?
Same font and same text should always produce same result

Oh I see, that command is not just checeking text width

A structured text, that contains ui2tex images?
I didn't know that was possible 😄

jo I see.. image width..
But it pulls it from the underlying DirectX texture, which's size is set in the procedural texture string.
Maybe its possible its somehow not initialized yet.

I need atleast some repro that triggers this code, with a ui2tex so I can follow along and see what might be going wrong

[gaunt depot]

I'll try but it happens so rarely and only on live servers

I guess I can just make it wait until ctrlTextWidth is not nil or something

[regal nimbus]

Is there a hidden branch with 2.18 on it? Need to check if something changed.

[fickle root]

legacy

[regal nimbus]

Does that have a code?

Ah, Arma3LegacyPorts

[verbal harness]

technically Arma3Legacy218
LegacyPorts is for MacOS compatability

[regal nimbus]

thx

[cursive tusk]

https://feedback.bistudio.com/T192750

[regal nimbus]

What's the process there? Do you have to beg Nvidia not to allow cheats in your game?

[cursive tusk]

im not sure, however plenty of other games have disabled the ability to use filters in their game, so kinda hoping it would be something simple like that

[fallen elk]

Helicopters can't seem to land on top of objects (terrain or otherwise) with land or landAt as of 2.20. Namely a problem for ships as they appear to prefer swimming

2.18 landAt
https://youtu.be/2DbSuyPcPFY

2.20 landAt
https://youtu.be/U2jtFTgOlXk

2.20 landAt Sea
https://youtu.be/mH26eyJCNNE

[modest sonnet]

I'm not sure about land but the landAt syntax changed in 2.20, it was extended to be able to land at arbitrary positions instead of just helipads https://community.bistudio.com/wiki/landAt

[solid marten]

whats the code?

[solid marten]

NVM got it reproed

[untold sky]

Do not spam tickets, thank you

[solid marten]

Revision: 152971

[cursive tusk]

Was just switching it to public and couldn’t figure out how, so closed the initial as a duplicate

[unborn acorn]

MK200 machinegun magazines use the wrong ammo classes: https://feedback.bistudio.com/T81009

[devout wave]

The ammo classes are correct, the weapon is intended to be caseless.

[verbal harness]

correct, see caseless cartridge here

[unborn acorn]

ugh brain, ye links can also carry caseless ammo and eject. sorry, close please

[true blaze]

These are caseless links

[cursive tusk]

What does Restricted Maniphest Task mean?

[devout wave]

It means the task's visibility has been restricted to certain users or groups of users. Could be an internal task, a security-related task, a task related to an unrevealed project, or a spam task that's been hidden.
Maniphest is one of the feedback tracker's systems. Maniphest tasks are FT tickets.

[cursive tusk]

I just wanted to view the task that I created but thats all I get when trying to view it, it says whoever authored it can always see it, but I cant now. guess im just confused

I guess im just trying to find out if its been deleted or something. Even on my account I cant even see that I created it

[gray wharf]

T192750?

[cursive tusk]

Yea

[gray wharf]

done

[cursive tusk]

Wait wym done?

[gray wharf]

you should have access (no, I did not fix the issue 😛)

[cursive tusk]

I don’t think you understand how excited I got for a second there😂 thank you though!

[gaunt stratus]

Guess thats the most correct channel to post this.
Our closed community start to feel crazy desync since the 2.20 update, in additional with broken player slots when relogging after briefing stage.

Can not provide any additional rpt files atm.
Games include TFAR, ACE, RHS and other mods, which was stable on 2.18. We did not really updated anything after release.
Desync caused invisible players and 'enabled' (?) godmod for several players, as shown on video. Player count was ~100 players

Hope this will help to understand something.

[regal nimbus]

@gaunt stratus Is the server running on perf or stable branch?

[gaunt stratus]

I would tell you tmrw, more likely that this one is perf. We did roll back to 'stable branch' after that, but im not 💯 sure.

Im sorry if this is obvious case, and did not bring any new info for this issue.

[regal nimbus]

Dedmen did find a desync (well, guaranteed queue stall) bug earlier but I think that was perf branch.

[gaunt stratus]

Well, we did not experienced this level of craziness after rolling back on stable branch. But TFAR is being crazy with its sync. Sometimes you can’t hear your teammates and its one-way issue usually (they hear you, you hear only few). Fixes by time or relogging on slot.

Not really related to arma problems, ig its TFAR Beta plugin stuff

Anyway, thanks for info.

[gaunt stratus]

In this case we've been on profiling branch.

[regal nimbus]

In that case you should report any networking issues on #perf_prof_branch. And if you run into major issues then try switching back to stable and see if it helps.

Perf v4 was released yesterday with a desync fix so try that first.

[gaunt stratus]

We did not tried v4 yet. But rolling back on stable version really helps us. We'll try v4 i guess.

[desert trench]

sukhoi191 has Arma 3 20 hours ago

Seems like there's an undocumented change in https://community.bohemia.net/wiki/ctrlSetEventHandler function. After 2.20 update, it returns a number instead of "Nothing". The function is obsolete but still available in ArmA 3. This change might break some old scripts (it did in Ravage mod).
See this post for more in-depth info: https://steamcommunity.com/workshop/filedetails/discussion/1376636636/591774577804196423/#c591774664103065074

from https://steamcommunity.com/app/107410/eventcomments/664957874066182489?snr=2_9_100003_#c591774664103071407

[gray wharf]

(↑ changed rev 152597)

[spring ice]

https://feedback.bistudio.com/T81303 please fix 🙂, this is majorly game breaking for any large scale Ai scenarios. 10 years it's been!

[regal nimbus]

Oh, this was actually better in A2?

[spring ice]

@regal nimbus yes

[regal nimbus]

Currently intending to drop waypoints entirely in favour of scripting because the behaviour is all busted. But I figured it always was.

[spring ice]

It's beyond busted and is probably not too hard to fix...

[regal nimbus]

GET OUT has a funny bug where they'll get out on previous waypoints instead for some reason. And then you have vanilla vehicles occasionally stopping before they reach the completion radius.

[spring ice]

Sick

[desert trench]

to see any chance of this getting fixed, i would strongly assume a tightly tailored repro/demo mission is needed.
the outcome may very well depend on various factors (vehicle type, group leader setup, combat and behaviour state, waypoints, etc)
unloadInCombat should play a key role. the config setting seems unchanged since OFP (active for car and motorcycle).

potentially a change in the system with PX, or AI could have influenced/broken it
the said different commands given by the effectiveCommander could indicate something

still would need debugging with https://community.bistudio.com/wiki/canUnloadInCombat and other AI state commands,
as well as said repro steps/demo mission

i am not 100% on this: i think Ondra did also some fixes in late OA with AI get in/out behavior (which may not have been merged to A3)

meaning it might matter when it was working as suggested (OFP, A1, A2, OA)

[daring wagon]

13:47:49 "Rastergröße erhöhen"
13:47:49 Bad conversion: string
13:47:49  ➥ Context:     [] L137 (x\enh\addons\main\functions\GUI\x\x.sqf)

What's that about?

I get the text like this private _text = getText (_x >> "text"); and directly print it to rpt

And this is the path configFile >> "Display3DEN" >> "Controls" >> "MenuStrip" >> "Items" >> "ToggleFlashlight" >> "text"

[spring ice]

Done, i'll add a link in the tracker

One Mech and one motorized with MOVE waypoint where the squad simply does nothing after the engagement and one of each again with a Seek and Destroy waypoint where the truck is leaving behind units while driving away.

[dusky tendon]

The players on our server started complaining that their ranks and experience and all the data were missing. Since version 2.20, callExtension ExtDb3 now returns a record error where "" is present. Before version 2.20, everything was fine and "" did not affect the return of the string in the callExtension

Screenshot 1. The name of the player with "".
Screenshot 2. The player changed the name where there is no ""

[regal nimbus]

I don't get it. Where's the ""?

Oh, they actually wrapped their profile name with quotes...

[dusky tendon]

Every players with "" lost all after join server...
Only the ЯПОНЧИК got the correct data because he deleted "" in his profile name

[regal nimbus]

I think you'd need to boil down the issue a bit more to get any response. The ScusDatabase functions are your code, right?

[hallow sun]

This was always a problem with extDB3 in my experience from years ago with Exile.
Had similar problems with players with "" in their name back then.

[dusky tendon]

yes my code

params [["_CustomStateName","",[""]],["_args",[],[[]]]];
(call SCusDatabase_fnc_ConnectSettings) params ["_db_name","_protocol"];
_plugin = ["0",_protocol,_CustomStateName]+_args;
_formatArray=_plugin joinString ":";
if IFDEBUG(2) then {diag_log text "CallCustomProtoloc";diag_log _formatArray;};

_return="Extdb3" callExtension _formatArray;
if IFDEBUG(2) then {diag_log text "CallCustomProtoloc";diag_log _return;};

parsesimpleArray _return;

[regal nimbus]

well, you have the diag_logs there. Is the issue introduced before or after the callExtension?

Well, or after the parseSimpleArray

[dusky tendon]

//[parsesimpleArray _return,_return]

[
    [1,[[""]]],

"[1,[[""""Andrei Gud"""","""","""",[[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0]],[[0,0],[0,0],[0,0]],[[1,0],[1,0],[1,0]],[[0,0],[0,0],[0,0]],[0,0,0,0,0],[],[[0,0,0,0,0,0,0],[0,0,0],[0,0,0],[0,0,0]],5,0,11145,0,0.0,0]]]"
]

Yes parsesimpleArray "Impostor"

[dusky tendon]

now i must return code in

call compile _return

But i try and return Nill

//[compile _return,call compile _return]

[
{[1,[[""Andrei Gud"","","",[[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0]],[[0,0],[0,0],[0,0]],[[1,0],[1,0],[1,0]],[[0,0],[0,0],[0,0]],[0,0,0,0,0],[],[[0,0,0,0,0,0,0],[0,0,0],[0,0,0],[0,0,0]],5,0,11145,0,0.0,0]]]},

<null>
]

Yes thx...

Find Error in log

22:20:37   Error position: <parsesimpleArray _return,_return]>
22:20:37   Error parseSimpleArray format error

//Error
parseSimpleArray "[1,[[""""Andrei Gud"""","""",""""]"
//Work
parseSimpleArray "[1,[[""Andrei Gud"","""",""""]]]"

Need try call this in 2.18

Also Call compile have problem with this string 🤔 #arma3_feedback_tracker message

[solid marten]

because the correct format is:

"[1,[[""""""Andrei Gud"""""","""","""",[[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0]],[[0,0],[0,0],[0,0]],[[1,0],[1,0],[1,0]],[[0,0],[0,0],[0,0]],[0,0,0,0,0],[],[[0,0,0,0,0,0,0],[0,0,0],[0,0,0],[0,0,0]],5,0,11145,0,0.0,0]]]"

if you want string you do str if you want quotes in string you do str str

str [str "name", "", ""]

The result is

"[""""""name"""""","""",""""]"

and there were no changes to parsesimplearray

[dusky tendon]

callextension result - "I don't care about your rules, I'll write two instead of four. It doesn't affect me anymore"

[desert trench]

File '@WW2SPEX\addons\ww2_spex_assets_c_characters_americans_c.pbo' not found.

any chance to see this type of message improved please

from what i understand this is actually a client missing a mod/pbo (and the message is logged on the DS)

at least it should mention the player in some form to have the context

Warning: Cannot evaluate '' would be nice to have context from -debug (from parsing configs stage)

[untold sky]

It's trying to say the value that was read, is not actually a string, so it cannot be read as stribng

But its not the config entry thats not a string.
Its a script variable.

[untold sky]

yes, but I need to repro it to find where that's from

[untold sky]

It already prints the context it has.
Need repro to add more

[desert trench]

the weird one with this is that AFAIKI its from either bugged mod.cpp or defines/macros in configs.
however it only appears in one session for that server. with multiple sessions have the same modset (same hash)

Slow network for player name='XXX', ID=DDD, steamID=DDD. DisconnectTimeout too high (11/30000 ms)

first time i noticed this one. does the value display at the end make sense?

like here it does:

Slow network for player name='XXX' ID=DDD, steamID=DDD. Ping too high (1208/ 999 ms)

[untold sky]

The default time is 90 seconds.
That shown there is 30, which seems fine.
server.cfg disconnectTimeout

[devout wave]

I think they're asking about the (11/30000 ms) which seems to imply that 11ms is being considered "too high"

[desert trench]

yeah the message doesnt quite make sense to me

[desert trench]

i'd love that they fix the broken script in the campaign mission " ", where, if your officer dies before you take the city, the mission breaks and you cannot advance and finish it. I mean, it's an early important mission in the main campaign; and i like the main campaign. I still recommend the experience, but it's broken.

(if anyone knows what the guy is referring to)

[devout wave]

I think it may be Tipping Point

In that mission, you assault a town, you have an NPC squad leader, and the Fandom wiki claims there was a soft lock if that squad leader died too early. It also says the issue was fixed in 2.14, though.

[untold sky]

The message is printed when the lower layer says that the connection has been dropped.
The logging code assumes it was dropped because the timeout had expired.

[unborn acorn]

Interesting to know after 2.20, will fixing this↓ bug still make big hit to player performance or maybe something can be done now?
Ticket Wulf comment: https://feedback.bistudio.com/T126021#1759434

[desert trench]

alright. so basically it says it has waited the disconnectTimeout + 11 ms and then dropped the guy? like the x/disconnectTimeout value seems a bit weird.

[untold sky]

No.
It said it dropped the connection (without saying what the reason for the drop was)
And informed you what the time and timeout were

In this case the time is waaaay less than the timeout

[untold sky]

No

[desert trench]

Ref to nonnetwork object 1213717: shell_tracer_red.p3d SPE_S_37L57_M51

these have no context, so engine created from ammo?

https://community.bistudio.com/wiki/setUnitLoadout has global arguments and effect - is that wrong?

_unit setUnitLoadout _unitClass; (without prior local check)

leads to

Ref to nonnetwork object 200fd21c700# 412648: spe_civ_pak2_swetr.p3d SPE_FFI_Doctor
➥ Context: [] L617 (WW2\SPE_Missions_p\GameModesBI_p\SPE_Combat_Patrol\CPInit.sqf)
[] L596 (WW2\SPE_Missions_p\GameModesBI_p\SPE_Combat_Patrol\CPInit.sqf)
[] L49 (WW2\SPE_Missions_p\UtilityFunctions_f\equipment\fn_SPE_setUnitLoadout.sqf)
from server.rpt - script only executed by isServer

this might explain also why i am seeing naked units now in 2.20. aka the setUnitLoadout no longer apparently gets applied to remote units.

is it possible setUnitLoadout network behavior was changed with 2.20?

[vapid cradle]

It would be great if two AIs could coordinate their effort to make a vehicule aim at a specific target. I'm working on this kind of cannon and this king of cannon takes 2 persons to handle it ( a driver and a gunner). The problem is this kind of cannon cannot rotate left and right, it can only rotate up and down and two AIs cannot coordinate their actions to aim at the target but players can.

[desert trench]

check the static weapon turning script in SPE

[deft marsh]

@hexed sonnet can you get @trim acorn to check his discord and check his pm's have stuff for tracker issue I posted above in a PM with him

thanks 😃

[vapid cradle]

Where do i find this script ? What is SPE ? The cannon is not a static weapon by the way, it's a car

[desert trench]

SPE_fnc_System_Vehicles_Static_MainAI_Thread
P:\ww2\SPE_Core_f\System_Vehicles_f\staticWeaponAI\fn_mainAI_Thread.sqf

SPE_fnc_System_Vehicles_Static_turnHandler
P:\ww2\SPE_Core_f\System_Vehicles_f\staticWeaponAI\fn_turnHandler.sqf

or
SPE_fnc_System_AI_SPG_mainAI_Thread
p:\WW2\SPE_Core_f\System_Vehicles_F\aiSPG\fn_mainAI_Thread.sqf";

[gaunt depot]

https://feedback.bistudio.com/T170835

Want to bump this. Tiles are great for procedural textures and controlling them through script would be even better

[untold sky]

I'll need repro then. I heard about a lot of ammo things indeed

[untold sky]

SetUnitLoadout in general is broken on 2.20 stable, we already fixed it in Prof.
But that no network only makes sense if that unit is local only?

[desert trench]

Alright. Not sure if armahost allows switching to prof - will check.

The code is

        _units = [];
        {
            _units pushBackUnique _x;
        }
        forEach (allPlayers + ([switchableUnits,playableUnits] select isMultiplayer));

        {
            _unit = _x;

            [_unit,BIS_infantryPlayer] call SPE_MissionUtilityFunctions_fnc_SetUnitLoadout;

            ...
        }
        forEach _units;```

[untold sky]

Are you actually setting a local-only units loadout?

Although man type units should not be local-onlyable. But I probably allowed it assuming people knew what they were doing when using that checkbox

[desert trench]

BIS_infantryPlayer is the class type (aka string)

so cant see how these units would be local only

i can test tonight locally with prof (DS+client)

Ref to nonnetwork object DDD : xxx

shell_tracer_green.p3d SPE_PzGr4042_KwK42_APCR
shell_tracer_red.p3d SPE_76mm_M1_M62_APHE
shell_tracer_red.p3d SPE_76mm_M1_M93_APCR
shell_tracer_red.p3d SPE_76mm_M7_M62_APHE
shell_tracer_red.p3d SPE_M101_M67_HEAT
shell_tracer_red.p3d SPE_M61_M1_AP
shell_tracer_red.p3d SPE_PzGr3942_KwK42_AP
shell_tracer_red.p3d SPE_S_37L57_M51
shell_tracer_red.p3d SPE_S_37L57_M74
shell_tracer_white.p3d SPE_50mm_PzGr39_AP
shell_tracer_white.p3d SPE_50mm_PzGr40_APCR
shell_tracer_white.p3d SPE_PzGr39_KWK40_AP
shell_tracer_white.p3d SPE_PzGr40_APCR
spe_105_shell.p3d SPE_M101_M1_HE
spe_shell.p3d SPE_50mm_SprGr38_HE
spe_shell.p3d SPE_76mm_M1_M42_HE
spe_shell.p3d SPE_76mm_M7_M42_HE
spe_shell.p3d SPE_76mm_M89_SMK
spe_shell.p3d SPE_M48_HE
spe_shell.p3d SPE_S_37L57_M63
spe_shell.p3d SPE_SprGr_KwK36_HE
spe_shell.p3d SPE_SprGr34_KWK40_HE
spe_shell.p3d SPE_SprGr39_HE
spe_shell.p3d SPE_SprGr42_KwK42_HE

@limpid rune would you happen to have an idea or a hunch what these would be/created how?
like its not regular weapon ammo, nor submunitions - yet how could these be generated?

[limpid rune]

might be something to do with some sound thing

like the interior hit sounds, ricochet sounds etc

thats the only thing I can think of

[limpid rune]

that isnt to say the scripts are creating the projectiles its more likely the scripts are trying to 'reference a nonnetwork object'

[untold sky]

projectiles are indeed a bit quirky. Some of them are local-only but every client has their own copy kinda thing

Most likely the real fix is to add local checks into engine, but for that I need to trace out exactly what thing that is

[vapid cradle]

Unfortunately i don't have spearhead.? But tell me if i'm wrong, this function is for StaticWeapons right ? Mu cannon is not a static weapon but a car, characters are supposed to be able to move it

Actually this cannon will require a driver to position and act as the gun turret and a gunner to handle elevation (and a third guy to bring the ammunitions). The cannon is not supposed to rotate on z axis to get the proper orientation but you're suppoed to drive it to the right orientation

[ripe crystal]

Any updates on this- we are still having issues with this 🥺

We check players weapon type a lot in onPlayerKilled and when this doesn't sync it breaks stuff in our mission

[unborn acorn]

Sounds of a shell exploding when hitting a vehicle are duplicated and produce glitches, cannot be heard when shell hit ground seems only vehicle
https://feedback.bistudio.com/T192872
https://www.youtube.com/watch?v=eUyZdX5aJLE

[desert trench]

SPE_fnc_System_AI_SPG_mainAI_Thread it used by tanks with a non moving turret.
you can get the SPE compatibility workshop mod and check the code via function viewer

[desert trench]

tried to narrow it down a bit without much success. cant see a createVehicleLocal of ours related to these ammo types, nor does the sound FX classes related

one possible angle might be:

_projectile addEventHandler ["Deflected", {
    params ["_projectile", "_pos", "_velocity"];

    [_projectile,_pos,_velocity,"default"] call SPE_fnc_positionalRicochet;
}];```
it doesnt have a local check for the _projectile.
it then tries to `attachTo` the soundFX to it, `setVariable` and `addEventHandler ["Deleted"`
however wouldnt this then create a script line trace?

[solid marten]

Just fixed it

[vague moss]

Thanks a lot! I'll check later

[solid marten]

I actually even had it screenshotted and attempted to fix it many times, today I succeeded

Revision: 152992

[vague moss]

Awesome, out of professional curiosity and if you can tell, what caused it/fixed it?

[solid marten]

The AI unit was still considered real player after control was passed back to player, while everything was AI, it even had all the AI scripts restarted

[vague moss]

Ah, very strange. Thanks for fixing it! 🙏

[solid marten]

Could you do me a favour, after it gets merged with DEV or Profiling could you check this ticket https://feedback.bistudio.com/T83435 if it works too?

[vague moss]

Can do! How long do you reckon until it is in profiling?

[solid marten]

🤷‍♂️

[vague moss]

I'll keep my eyes open then :)

[vapid cradle]

I'm afraid ww2_spe_core_f_system_vehicles_f is an ebo not a pbo, i can't open it

[alpine tulip]

You can check it in Functions Viewer

[vapid cradle]

Oh thanks

Hmm i see, it calculates a new position according to heading of current target

i'll try it

[dusky tendon]

What is the best way to fix the signatures of the game files?
I changed the version to beta 2.18 and after I went back to 2.20 I have a ui_f.pbo file signature error.

[devout wave]

Verify your game files through Steam

[sleek scaffold]

I am having the issue of some modded weapons not casting shadows. Is that something known? (2.20 Main Branch). The mod in question is "The Free World Armoury"

But I also experienced it with other mods

[untold sky]

Actually the ref to nonetwork don't have a script trace.
They are printed when the message is serialized, not when its created.

I assume its the attachTo, is soundFX thing a global object?

[untold sky]

Probably tuesday

[untold sky]

ui_f was updated in the hotfix from yesterday. Either your file or the server's file is outdated

[dusky tendon]

Yes need was update server.

[desert trench]

local - however as said this may be totally unrelated. its just the only case i could find potentially related

    private _soundSetSource = createVehicleLocal [_soundSetSourceClass, _pos, [], 0, "CAN_COLLIDE"];

    _soundSetSource hideObject true;
    _soundSetSource attachTo [_projectile,[0,0,0]];
    _projectile setVariable ["SPE_soundSetSource",_soundSetSource];

[vapid cradle]

I think the weapon should define the position and the direction of the gunparticle but it's the ammo that should determine what gunparticle would be applied. For example, i added a springfield trapdoor, this thing usually shoots black powder cartridge which produces a big white smoke cloud but it can also fire smokeless powder cartridge which shouldn't produce white smoke cloud

It's still possible to do something with event handler but well....

[tropic jewel]

Could close MFD ticket now? =p https://feedback.bistudio.com/T167913

[fast lodge]

It would be super nice to have an script command which disables the zoom on right click

btw one thing which is super annoying is that if you dont have an weapon in your hand and you right click it will put you in a slow walk mode. (optics mode key)
There should be an option to turn it off since its on the mouse and you cant override the output of mouse actions or remove that system completly no point to have it if you also have the toggle slow walk option which you can use. A lot of people get into this state and they have no idea why they are slow walking

(isWalking player) dosent even return true because its not toggling the slow walk, it just puts you into the slow walk state

[unborn acorn]

https://feedback.bistudio.com/T192872#2827310 sраm

[heady thunder]

https://feedback.bistudio.com/T192853

Späm

[unborn acorn]

~~hmm diacritics, there 3 errors messages in last mission Apex Protocol campaign
open \a3\missions_f_exp\campaign\missions\exp_m07.tanoa\mission.sqm search for next text: ~~
init="doStop this; this setUnitPos ""ÜP"";";

[daring wagon]

Who's the German who typed Ü instead of U 😄

[gray wharf]

a Frenchman perhaps Ü

you would be… surprised! Ö

[daring wagon]

A french would have put a 🏳️ in there instead.

[solid marten]

it waz ü

[stoic idol]

certainly french. le üp

[oak apex]

Hello feedback team. the devs at Rotators Collective were kind enough to help me pinpoint the source of a bug thats occuring with vanilla arma. I though it was an issue with Reaction Forces but that doesnt seem to be the case. Any time i spool engines on a helicopter with a HMD, like the AH-99, MH-235, or stub wing blackfoot, the frame rate tanks to less than 1 FPS. outside the helo, or in 3rd person, frames are around 100. Didn't notice this happening before the latest update. Any advice or help with the issue would be greatly appreciated. The RPT files didnt seem to be much help but im more than happy to screen share and replicate the issue

should note that this occured running unmodded, vanilla Arma 3

[alpine tulip]

I actually checked the stream and it starts to lag the game after about a second upon he starts the engine

There is no script error spam so I am sure it's not our (CDLC) issue

[frosty osprey]

Hi there from your (probably) most retro player. I want to thank the devs for fixing the unsupported Intel instructions in the x64 executable, which now loads fine. And in game it feels pretty fluid. However two new issues have appeared after updating to v.2.20 (Windows 7 x64), which as per the changelog should be fully supported... The first one (T192972) is a combined launcher and game issue by which the Steam API is not working. This prevents the server list from loading, and even if I manually try to connect to an official server, it never loads. Essentially every feature involving Steam is not working... which is very bad. The second issue (T192977) is that I can't get into any vehicle, even in an offline single player game. Please have a look at these issues and see what can be done. If they are a launcher thing, maybe I can patch the v2.20 install with the previous v2.19 launcher?

[daring wagon]

Just be aware that the steam client also has no support for Win7. So that might play a role as well.

[agile trail]

Probably doesn't help you, apart from to suggest that it's not a general Arma bug, but more likely related to your system/settings, but I'm not seeing this issue. Having said that, I'm on the dev build 2.21, so you could check if that helps with your issue.
73 fps before start-up, 70 fps afterwards and no drop with time.

[gusty shell]

Hey there,

I’ve found a strange bug related to CDLCs (tested with both S.O.G. Prairie Fire and Spearhead 1944).

Steps to reproduce:

1) Start any single-player mission (tested on The Dentist – SPE, Temporary Duty – PF, and several others – it seems to happen across all missions and campaigns).

2) Press Save & Exit.

3) Load the mission and check your map.

Issue:
In the date section of the diary record, the day of the week (e.g., Sun, Mon, Tue...) is replaced with random words or sentences. It’s unclear to me where these come from.

This bug occurs even in BI Arma 3 missions when CDLC is loaded. It does never occur in pure vanilla Arma 3 (with no CDLCs or mods).

It doesn't seem possible to fix this through scripting. Would someone from the dev team be able to look into it, please?

[desert trench]

i can confirm this - for me its "unregistered string" tho

as far as i can tell this diay title/subtitle? is purely engine generated. so no clue how CDLCs could cause this

[crystal wraith]

Orange, uh, hyperlinks?, in briefing tabs do not underline when hovered for me as of 2.20. True for the default ones like vote kick, ban, in multiplayer, as well as marker links or other scripted links in mission briefings. The buttons still function when clicked, there just isn't a hover behavior

[oak apex]

Interesting, how do I switch to the dev build?

[agile trail]

Steam > Arma 3
RMB Properties
Betas
Beta Participation: development

[oak apex]

Holy shit that worked

[oak apex]

So it seems the only build of the game that was having that issue was the profiling build...which is a shame since thats the most high performance build for me

[oak apex]

Yeah confirmation. Regular and dev builds are fine. The profiling build is having the HMD issue. Is anybody else having this problem? Any known fixes?

[modest sonnet]

Open a ticket/report to dedmen/ #perf_prof_branch with a reproduction

You gotta keep an eye on perf changelogs if you run it, last build included some changes to MFD radar stuff
But the bulk of performance changes were already pushed with the 2.20 stable update, you shouldn't see that much difference between perf/stable at the moment

[trim acorn]

@deft marsh no need to startle rest of the bunch, I am onto it 😃

[devout wave]

Ticket needs moving to the Refrogger section (or maybe DayZ, can't tell) https://feedback.bistudio.com/T192962

[gray wharf]

thanks, I asked

[oak apex]

Anybody got a fix for the black screen issue? Had to switch off of profiling because there’s an issue with choppers an HMD, and now about 60% of the time I boot arma the screen just goes black. Can’t alt+tab, can’t ctrl+alt+delete. Whole screen goes black and I have to turn my PC off then turn it on again.

[regal nimbus]

No mods?

Not heard of that one.

[oak apex]

Thanks. Thats a lot of help

[hallow sun]

Do you start the game and then tab out by chance ?

If so, try to run the game in windowed full screen mode, that should get rid of the blackscreen freeze thingy.

[dreamy bane]

I wonder why this code doesnt trigger the ScriptError EH ```sqf
addMissionEventHandler ["ScriptError",
{
hint "Got error";
}];

[] spawn
{
sleep 1;
i = / / 0; // Creates "invalid number in expression" error
};

[gaunt depot]

Do you execute that code as whole? It probably fails to compile the whole thing without adding the event

[dreamy bane]

yes as whole in init.sqf. it may be its more of an parser error than runtime. that would explain the behaviour

[regal nimbus]

There are possibly multiple issues where the game thinks it has more threads than it does, and things grind to a halt. However ctrl+alt+del should eventually work there. Might be quite delayed because it's a CPU-thrashing case.

[oak apex]

When I was running profiling build I could tab out with abandon lol. Had to go back to stable for other reasons mentioned. Learned my lesson about tabbing out but sometimes when it’s starting up or just when I’m playing a scenario the screen will freeze, go black, then after a bit I can hear the game running but the screen remains black.

[untold sky]

We also raised the minimum system requirements.
The instructions will probably come back.

[untold sky]

FT ticket please, with the first image included.

[untold sky]

I need an RPT of that.

[oak apex]

I sent RPTs to the Rotators guys it didn’t help them

[cyan basin]

update your outdated steam client. nothing can be done on their end, win7 and 32 bits are deprecated with 2.20 and going to be removed in 2.22

[vague moss]

Gonna try the other issue real quick

[solid marten]

Zeus?

[vague moss]

Yeah

Remote control broken-AI thingy

[solid marten]

OK lemmie know zues result

[vague moss]

gonna try this one now, yeah

[vague moss]

As Zeus,Place a rifleman or other unit,place an unarmed enemy and he engages. Take control of him, then release control. Place a unarmed enemy and watch him simply stare at him instead of shooting.

Place a manned static mg and take control of it, then release. Place an enemy unit,he will engage,disembark him from the static and place another enemy,he will stare at it.

the Rifleman now correctly engages, as well as the dismounted static mg operator

So seems to be working!

[solid marten]

cool cool!

[vague moss]

Do I close the Issue?

[solid marten]

I just did

[sudden loom]

https://feedback.bistudio.com/T193098

Helicopter will not land on helipad that was placed on object over sea surface. He will fly round, but never land on that helipad.

[vapid cradle]

I think i found a solution for this thread https://feedback.bistudio.com/T181499. Just need to increase the values

                {
                    ambient[]={10000000,10000000,10000000,1};
                    diffuse[]={10000000,10000000,10000000,1};
                    emissive[]={10000000,10000000,10000000,1};
                };```
of the material of the MFD. The result is not as beautiful as when the values are at 1 but it's still readable and that's what i need. With these values i can go 1km deep and i still see the MFDs

[solid marten]

I'm going to look at it, in the mean time get rid of the waypont and add this to the helicopter init

this landAt [HELIPAD, "land"]

[solid marten]

OK fix in next profiling

[sudden loom]

There a small detail: landat command will work, but if you place helipad a bit above object's surface (like 1 m + above objet's surface), then helicopter will not land at object's surface and will hoven above object's surface.

[solid marten]

then use optional boolean

heli landAt [position, mode, waitTime, putOnSurface]

putonsurface

this landAt [HELIPAD, "land", -1, true]

[sudden loom]

It's seems that helicopter try to land exact at [0,0,0] coordinates of helipad and ignore object's surface at all.

[solid marten]

with waypoint? yes thats the problem at the moment, should be fixed in next profiling

[sudden loom]

with landat without putonsurface param

[solid marten]

Only not [0,0,0] but [x, y, 0]

[sudden loom]

I mean in object's model coordinates

[solid marten]

no it should still work it would just hover

anyway, before we get all more confused, wait for the next profiling

[sudden loom]

Ok

[solid marten]

Should be fixed in Revision: 153035

[oak apex]

i mean im not even making it into the game lately.

I don’t know what happened but on the last week I’ve been able to play maybe an hour. Game keeps crashing. It’s very frustrating

[unborn acorn]

https://community.bistudio.com/wiki/CfgWeapons_Config_Reference#reloadSound[]={"",1,1}
This parameter doesn't seem to work e.g. the mortar is supposed to make reloading sounds every time it fires, but it definitely doesn't
1) spawn B_Mortar_01_F
2) shot
3) listen → no reload sound
Although magazine reload sound works ok
Actually everything works, for some reason I was sure that this was not happening. whoopsie sorry

[daring wagon]

Can someone please hire Trager for BI QA. He seems to enjoy that 😄

[deft marsh]

hehe

[sleek scaffold]

Yeah and not only that but he also often even provides the potential fixes to the issues. How many QA testers do that?

[cyan basin]

most in serious jobs do, afaik

[gusty shell]

https://feedback.bistudio.com/T193129

Here it is.

[last jolt]

I dont know if this is the right place to ask, but has a fix for the shortened render distances of objects when using a custom FOV value in your arma profile ever been considered?

Maybe it's not a well know problem but when you set a custom FOV to be larger than default, you have to zoom in to get normal rendering distances of objects, but if you zoom out to normal view everything always results to around 2km. I have always just learned to deal with it but with recent fixes to some well known bugs I figured I would bring this up

[vapid cradle]

I created a ticket https://feedback.bistudio.com/T193149 because when i attach an object to an ai character and try to use playAction or playMove, the character plays the animation but does not move

It happens with any kind of object and any kind of ai character

[vapid cradle]

nevermind, i think i found a way around isong disableAI "MOVE" first

[outer spear]

Yarrr!

[vapid cradle]

Actually isong disableAI "MOVE3 is not a good idea, the character won't collide with objects anymore

[desert trench]

should the game actually crash on faulty config inheritance definitions? can it be made not to crash even?

ie with IFA3 and SPE we encounter this fairly often with other mods doing bad definitions and when loaded alongside, it breaks the game start like this and a "crash"

16:41:19 Error updating 'config.bin/CfgVehicles/SPE_ArmoredCar_base/DestructionEffects/Light2/' by 'WW2\SPE_Assets_c\Vehicles\ArmoredCar_U1_c\m20_auc\config.bin/CfgVehicles/SPE_ArmoredCar_base/DestructionEffects/Light2/', base 'Light1'
16:41:19 ErrorMessage: config.bin/CfgVehicles/SPE_ArmoredCar_base/DestructionEffects/Light2/: Cannot find base class 'Light1'
16:41:19  ➥ Context: WW2\SPE_Assets_c\Vehicles\ArmoredCar_U1_c\m20_auc\config.bin
16:41:19 Application terminated intentionally
16:41:19  ➥ Context: WW2\SPE_Assets_c\Vehicles\ArmoredCar_U1_c\m20_auc\config.bin
ErrorMessage: config.bin/CfgVehicles/SPE_ArmoredCar_base/DestructionEffects/Light2/: Cannot find base class 'Light1'```

even more annoying is that rpt wont log the responsible source mod - something like this would be needed:
16:41:19 Updating base class ->Launcher_Base_F, by KA_SRAW\config.bin/CfgWeapons/KA_SRAW/ (original KA_SRAW\config.bin)
however probably complex to impossible to track/identify in the config building stack

that said the game does build an origin tree with these
https://community.bistudio.com/wiki/configSourceMod
https://community.bistudio.com/wiki/configSourceAddonList
https://community.bistudio.com/wiki/configSourceModList

so maybe its feasible after all

the other problematic part to identify the source is that both the addon list (pbo/addon) as well as the full modlist is logged only later - so if the game "crashes" during config issues, its a lot harder to track down the source mod to cause it

what does this actually mean?
16:41:08 Warning, config 'rds_a2port_civ\config.bin' contained invalid CfgPatches entry with extra required Addon 'A3_Weapons_F'. Invalid CfgPatches entries are invalid. A temporary workaround is being applied, this mod will soon stop working

its apparently not from reyhards version but an older repack of this mod - it has this:
requiredAddons[] = {"A3_Weapons_F","A3_Weapons_F_beta","A3_Weapons_F_Acc","A3_Structures_F_EPA_Items_Electronics","A3_Structures_F_EPC","A3_Soft_F","A3_Soft_F_TruckHeavy","A3_CargoPoses_F","A3_CargoPoses_F_Heli"};

however thats perfectly valid, right?

[desert trench]

---

another interesting issue - despite using the par file, it cuts off the modline

it cuts off at that @TF Snow Storm Menu mod

from rpt

@tf snow storm menu doesnt seem to contain any non standard characters

any ideas what could be amiss here?

(trying to replicate another issue loading 300+ mods)

[desert trench]

seems not related to this particular mod - with a different setup it stops with another mod

it seems like even the -par file has a limit on number of mods?

~4050 characters till @TF

[agile trail]

Whenever I've encountered hard game crashes, it's generally been caused by animated class Reflector lights, such as spotlights on vehicle turrets.
Haven't been able to pin it down to any one specific thing though.

I'm fairly sure there was (still is?) a limit on the number of characters available for the mod list.

[dreamy bane]

is it possible we get longer stack traces for errors?

[regal nimbus]

longer?

If you need a longer stack trace than what you get in -debug then you must be doing something truly horrible :P

[dreamy bane]

its not quite long enough to see the origin of the error, sometimes.. sometimes theres like ten traces on same file and that doesnt help

[dreamy bane]

maybe command to set the maximum trace length . for developers so it wont effect other users

[regal nimbus]

For debugging a complex system I'd generally recommend manually logging the input parameters at the start of each function.

stack traces aren't supposed to cover everything.

[daring wagon]

I mean I'd you have 10 nested functions then it might be cut of.

Might also want to rethink the whole code structure at that point.

[dreamy bane]

typically programming languages show the whole stack

[dreamy bane]

i use lot of functions, thats not going to change 🙂

[regal nimbus]

Yeah but usually when they show you a stack it's because they crashed. I don't want a screenful of crap every time a vehicle is created.

[dreamy bane]

its for crashes that I need this

[foggy shuttle]

If you're talking about just reading script errors, could probably just utilize the "ScriptError" event handler to diag_log the parameters it gives. If that's not what you mean, then ignore me lmfao

[dreamy bane]

thats what i am using. but the trace cuts off..

[foggy shuttle]

That's probably because diag_log has a character limit

Might be able to copyToClipboard if you're testing locally

[dreamy bane]

nah it does it line by line. its the trace array thats short

[foggy shuttle]

Strange, never had an issue with that and I use it quite often

[dreamy bane]

well most of the time it works for me too but not today

it was useless 🙂

[alpine tulip]

https://feedback.bistudio.com/p/jeckyjeck/
Hello ChatGPT, please spam this feedback tracker

[desert trench]

faulty definition in Tractor_01_base_F

[untold sky]

It says XAudio,

Headphones (Intel® Smart Sound Technology for MIPI SoundWire® Audio) - Channels: 8, Freq: 48000
What a weird output. Setting frequency to 44100 should help with XAudio crashing

[untold sky]

Custom FOV is not supported

[untold sky]

need repro to know where it happens to see if I can find the mod that messes it up.
But I assuem its bad because a mod is removing that base class? I don't think we can know who did it after the fact. But potentially we can list some candidates

[untold sky]

yeah it does that on classes, not on non-existent classes

[untold sky]

invalid CfgPatches means the classname of the CfgPatches entry was invalid.
Probably duplicate

[untold sky]

That is a param text file, not a .par config file right?
Maybe the config file format would handle more data

I don't see a limit, there is a 4096 buffer but it gets reused, would have to repro that

[untold sky]

The length is not limited, it prints the whole stack

[dreamy bane]

ok in that case it must be remoteExec or something that cuts the trace

[desert trench]

-language=Chinesesimp
rpt

Unsupported language: Chinesesimp
and
Unsupported language Chinesesimp in stringtable
but it works

https://community.bohemia.net/wiki/Arma_3:_Startup_Parameters#language

Starts client with preferred language. See Stringtable.xml - Supported Languages for a list of supported languages. Case-sensitive! (className column)

https://community.bistudio.com/wiki/Stringtable.xml#Supported_Languages
className: Chinesesimp

is the check wrong for the warning?

[last jolt]

fair enough

[vapid cradle]

I'm wondering, why is the AI targeting the vehicle when the driver and the crew are visible and targetable ?

I mean, under a certain ratio in between the hit value for the current ammo and target vehicle armor value, the ai should target the crew when its visible and targetable

[untold sky]

stringtable one means that the translation was not found, it used one of the fallbacks (for exmaple Original or English)
The first one means it wasn't found in our engine internal language whitelist (all languages that can be enabled by languageList.bin) and indeed, Chinesesimp is not in there, only Chinese.

I ran into that too when we added the new languages. Our "Chinese" is actually "Chinese (Simplified)" and Chinesesimp is.. well not there.

I didn't want to deal with it.

[untold sky]

Because complexity and no-one added it.
Getting the vehicle is easier, than getting the vehicle and then checking all inhabitants

[devout wave]

The AI can target exposed crew members but it depends on how the vehicle is set up. For example, they'll target turned-out tank crew with small arms.

[regal nimbus]

Apparently the CUP M113 that I just tested is set up wrong then :P

[limpid rune]

it can be real difficult to get working

like you can set crewVulnerable and that will basiclaly make AI always try to shoot crew, but the problem is if the AI has an AT weapon (or they're a tank) they'd sooner use their MG/Gun instead

[devout wave]

You could ask the WS team about it as well, they ran into this stuff with their Marid HMG variant. Basically "turned out" is a fairly specific state for an armoured vehicle, and the AI crew will try to avoid using turned out positions when in combat, so for AFVs you don't necessarily want to treat combat positions as turned out - the tradeoff being that the AI will treat non-turned-out positions as being fully protected by the vehicle's armour, and won't try to use small arms on them.

[limpid rune]

the other issue is AI dont like mixing some positions being turned in and some turned out

[devout wave]

WS ended up working around it by attaching an invisible target dummy to the exposed gunner position to make the AI shoot at it, since the dummy's not actually "in" the vehicle

[limpid rune]

oof

[devout wave]

Full discussion of it on the Rotators Discord around here #901047391451484191 message
Apparently they asked BI for a better way to flag turrets as vulnerable, and the request was declined

[limpid rune]

generally thats sort of thing shouldnt be necessary but im guessing what I said with mixed turned in/out might be the issue

cuz like the M3A1 halftracks in spearhead the AI will stay on the turned out mg, and get shot at etc without much issue but I spent a lot of time getting those to that point

[tropic jewel]

Vanilla offroad as well with the standup-turnaround-turnout. When turned out enemy will shoot the unit. when turned in it will switch to targeting vehicle center.

[vapid cradle]

Complex ? I'm not a specialist of Arma engine but you can easily check if a line intersects a convex shape without intersecting other convex shapes. But yes it would probably consumes way more resources than just targeting vehicle center and apply a random small deviation

The ai doesn't have to find the appropriate position, we just need something simple. The ai can check from its current position if it has an angle to shoot the crew, if yes it targets the crew instead of the vehicle. If it doesn't have an angle it just keeps targeting the vehicle

[scarlet sage]

One time I was playing Extraction and those dummies were not so invisible.

[desert trench]

this setup (from Eden) apparently adds the appId into class mission of the mission.sqm and thus makes the mission require the CDLC (and not allow non owners to join) - this seems to be a bug, right? aka this should only happen if "Require DLC" is ticked, right?

[uncut briar]

IIRC yes, if require DLC is not checked it should be visible that the mission uses DLC on the select screen but you should be able to play without owning

[desert trench]

above setup apparently results in:

class Mission
{
    appId = 1175380;
    assetType = "Free";```

[on that note thanks for your responses Dedmen - quite busy this week, yet I will respond/make FT tickets with repros for said issues]

[daring wagon]

Yeah, iirc it was only supposed to show a frame around the scenario image in the scenario menu.

[vapid cradle]

My bad, actually there is an option for it: https://community.bistudio.com/wiki/Config_Properties_Megalist#crewVulnerable

[vapid cradle]

More info has been provided: https://feedback.bistudio.com/T193149

[stoic idol]

if this is the right place to report, there's some spam bots here: https://feedback.bistudio.com/project/view/79/

[untold sky]

@sinful kettle https://feedback.bistudio.com/T163593 laser designator eventhandlers

_unit addEventHandler ["LaserTargetChanged", {
  params ["_unit", "_laserTarget", "_designatedTarget"];
}];

When laser is turned off, then [_unit, objNull, objNull] or we need a separate event just for laser on/off?
You wanted 3 EH's but I think we can all do it in one.

An EH that triggers when the laser is activated (using Fired/FiredMan perhaps)
An EH that triggers when the target is being lased (perhaps both for the "firer" and the "target"?)
An EH when laser target is lost

Laser activated, will be _laserTarget becoming non-null, and laser deactivated will be _laserTarget becoming null (that's the entity thats spawned at laser impact point)
Target being lased, that happens whenever the target position changes?
target being lost, target is just null.

Tho when it says "changed" it should pass both current and prev target, no?
It will basically always transition through objNull when you laze at terrain between two targets

[gaunt depot]

Damn this is cool!

How does it work for vehicles though? Unit can lase with vehicle turret AND personal laser designator at the same time

Start lasing as commander seat, turn out and use laser designator

Even more, turrets can keep lasing targets without a unit in them @untold sky

https://feedback.bistudio.com/T167317#2351958

My repro steps to have 2 lasers at once from 1 unit

[untold sky]

Very good question 🤔
The lazing here is done by the entity (the soldier)
A vehicle would also be an entity, but a vehicle can have multiple lasers?

The laser is actually per .... Oh I see

I actually get the turret that is doing the lazing.
I have access to turretpath, gunner inside the turret, weapons of the turret.

So I should probably pass some of that

[gaunt depot]

Send two entities into EH perhaps? Vehicle and instigator unit?

And turret path perhaps?

[untold sky]

I think vehicle (the one EH was added to), the operator of the turret/weapon, the turret path

[gaunt depot]

params ["_unit", "_vehicle", "_laserTarget", "_designatedTarget", "_laserWeapon", "_turretPath"];
```maybe?

Can you have two different lasers on one turret? I dont remember if I ever tested that.

_laserWeapon in case you can have two lasers, if its hardcoded to be impossible no need for that and you can just cycle weapons until you find laser on that turret if you need it

[untold sky]

Turning out already is not a "problem"

[gaunt depot]

Not sure what it should return in case of empty turret still lasing. I know its possible but not sure how it works internally and what's accessible

[untold sky]

Because the vehicle laser is triggered as event on the vehicle.
The turned out units laser, is triggered as event on the unit

But the vehicle itself can have multiple lasers (gunner/commander turrets)

[gaunt depot]

Hmm, leave it to scripter to find which turret fired with VEHICLE laserTarget TURRET command?

Or pass turret path in the EH to make it easier

[] for personal lasers, [-1], [0], etc. for vehicle lasers?

[untold sky]

If turret is empty and target changes (for example driver is driving around), the operator/instrigator will just be null.

[gaunt depot]

Oh yeah, path would be more reliable then

[untold sky]

There are no personal lasers when you add the eventhandler to a vehicle.
There are also no vehicle lasers if you add the eventhandler to a unit

[gaunt depot]

Ah, alright

Same event name though?

[untold sky]

yeah

[gaunt depot]

_unit addEventHandler ["LaserTargetChanged", {
  params ["_unit", "_laserTarget", "_designatedTarget", "_turretPath"];
  // On unit: Unit, Laser, Target, []
  // On vehicle: Vehicle, Laser, Target, [0]
}];
```?

[untold sky]

It looks like you cannot have multiple lasers in one turret.
The turrets weapon state only has one laser target, but has multiple weapons in it

[gaunt depot]

Maybe swap unit with turret path array?

so you can get it yourself with turretUnit

[untold sky]

_unit is the vehicle the EH was added to. You still need that to use turretPath on it

[gaunt depot]

^ Nevermind, I just reliazed unit would be vehicle

Yeah

Providing laser weapon class name could make things easier for newbie scripters too

[untold sky]

Actually I think I don't even know which weapon started the laser.
And you may switch to a different weapon after starting it. And you may have multiple that can start one

[gaunt depot]

Not needed then, its a them problem

[untold sky]

You could check it yourself by detecting the first event when the laser is started up, grabbing the turret, and finding the currently selected weapon

I should probably provide some special flag to identify first laser startup.
Problem is that would be a useless parameter during the actual target-changing while laser is active

[gaunt depot]

Leave it to scripter to setVariable save last laser target to check if it was null, if it was then its a new laser firing and its time to find the weapon

_unit addEventHandler ["LaserTargetChanged", {
  params ["_unit", "_laserTarget", "_designatedTarget", "_turretPath"];
  if(alive _laserTarget && isNull(_unit getVariable ["last_laser", objNull])) then {
    private _laserWeapon = if(count _turretPath > 0) then {_unit currentWeaponTurret _turretPath} else {currentWeapon _unit};
    systemChat format ["Lasing with: %1", _laserWeapon];
  };
  _unit setVariable ["last_laser", _laserTarget];
}];

Guess this should work

[untold sky]

For some reason the player designating the tower, interrupts the gunner laser's designation of the tower 🤔

If both player and tank aim close to the same spot, both get objNull designated targets.
If I move the player aim higher up, after a bit of distance, both get the tower 😄

So two laser targets close to eachother, disable eachother??

The LaserTarget entity, has no shape/model, intentionally so.
When scanning for target it uses nearestObject (not nearestObjects), and it ignores objects that have no shape.
So it finds the LaserTarget, and ignores it, setting the target to null.

That is quite annoying, and also sounds problematic in multiplayer if two players are designating the same vehicle, and their lasers get too close

[vocal abyss]

When scanning for target it uses nearestObject
It being the EH?

[gaunt depot]

Hidden feature to disable enemy laser largets, lase it yourself

[untold sky]

no the laser updating code

[vocal abyss]

Another quirk: Laser targets spawned by script just vanish after a while, unless they are attached to something. What's up with that?
From my pet project:

private _laserTarget = "LaserTargetC" createVehicle [0, 0, 0];
_laserTarget setPosASL AGLToASL _position;
_laserTarget attachTo [IFSI_laserTargetAnchor]; //Keeps _laserTarget from despawning.

[gaunt depot]

Guess laser targets are hardcoded to some TTL which refreshes if its being laser by something in the engine

[sinful kettle]

yeah I think one event is good enough.

Target being lased, that happens whenever the target position changes?
you mean it keeps triggering for moving targets?

[untold sky]

No I decided not to. I'm passing the laserTarget object, and getting position is just getPos on it.
Can do that eachFrame if someone wanted to track the position itself

[untold sky]

Yeah I found it, if its not being "actvated/moved" for 10 seconds, it deletes itself

but I think setPos counts as move

Yes, setPos updates it

[untold sky]

Oh also objectParent laserTarget returns owner. I forgot who wanted that

[dusky tendon]

bruh...
so many lights make black zone

[dusky tendon]

"dedicated server with mods"
We had a player shooting missiles from a plane and these missiles were hanging in the world and it is impossible to even delete them using deletevehicle

These missiles don't even respond to setdamage commands. I called the code globally in the console with the assignment of damage 1. After that I checked the damage and the missiles had 0 damage.

[untold sky]

ah yeah ive seen that happen before

[vapid cradle]

Yes i've noticed that too when i was making naval cruise missile for my submarines. When you spawn a missile class ammo it doesn't blow up until it hits something or runs out of timer

SetDammage doesn't work on it, ithought it was intentionnal back then but when i think about it now, it looks like a bug

[dusky tendon]

I have a different error. As you can see in the screenshots, I did not do them at the same time, and the rocket remained in one place as if it had simulation disabled. But in fact, there is no code anywhere that says that the simulation was disabled for rockets. In addition, it looks more like a desynchronization. Because the players who were next to me saw rockets in the air that I did not see.

[vapid cradle]

Usually deleteVehicle works on missiles. That's how i blow up my missiles before they reach their target, i delete it, replace it with a custom explosive charge with appropriate hit values and instantly detonate the explosive charge with SetDammage 1

[dreamy bane]

is there really no way to prevent action menu from fading away? Its a problem with BIS_fnc_holdActionAdd because if you have too long duration the progress icon will fade away before the action completes

[dusky tendon]

Bruh and where problem...

14:36:56 No alive in 10000 ms, exceeded timeout of 10000 ms
14:37:09 No alive in 22531 ms, exceeded timeout of 10000 ms
14:37:20 No alive in 33578 ms, exceeded timeout of 10000 ms
14:37:30 No alive in 43578 ms, exceeded timeout of 10000 ms
14:37:40 No alive in 53578 ms, exceeded timeout of 10000 ms
14:37:50 No alive in 63578 ms, exceeded timeout of 10000 ms
14:38:00 No alive in 73578 ms, exceeded timeout of 10000 ms
14:38:10 No alive in 83578 ms, exceeded timeout of 10000 ms
14:38:20 No alive in 93578 ms, exceeded timeout of 10000 ms
14:38:30 No alive in 103593 ms, exceeded timeout of 10000 ms
14:38:40 No alive in 113593 ms, exceeded timeout of 10000 ms
14:38:50 No alive in 123593 ms, exceeded timeout of 10000 ms
14:39:00 No alive in 133593 ms, exceeded timeout of 10000 ms
14:39:10 No alive in 143593 ms, exceeded timeout of 10000 ms

I again try do this and game freeze...

14:45:49 No alive in 10000 ms, exceeded timeout of 10000 ms
note: Minidump has been generated into the file C:\Users\Andrey\AppData\Local\Arma 3\Arma3_x64_2025-07-21_14-39-54FROZEN0.mdmp
14:46:02 No alive in 23641 ms, exceeded timeout of 10000 ms
14:46:14 No alive in 35469 ms, exceeded timeout of 10000 ms
14:46:24 No alive in 45469 ms, exceeded timeout of 10000 ms
14:46:34 No alive in 55469 ms, exceeded timeout of 10000 ms
14:46:44 No alive in 65469 ms, exceeded timeout of 10000 ms

i have file
Arma3_x64_2025-07-21_14-39-54FROZEN0.mdmp
he can help find problem?

I figured out what the problem was. I had a search code for selecting a player in onLBSelChanged

    _size=(lbsize _ControlListboxPlayers) -1;
    for "_i" from 0 to _size do {
        _iddata = _ControlListboxPlayers lbdata _i ;
        if (_iddata == _selectedplayer) then {
            _ControlListboxPlayers lbsetcursel _i;
        };
    };

which caused an infinite loop of calling the list item selection. ((((

[unborn acorn]

Magazine reloading sound does not follow player's weapon, still stay in position when unit start reload https://feedback.bistudio.com/T193506

[vapid cradle]

I added a video to explain why i need this issue to be solved. It's not to bother you, its just i really need it to be fixed ^^ https://feedback.bistudio.com/T193149

[ripe crystal]

Wanna preface this with that this could be an issue within our mission and not an actual profiling issue. More background though can be found in #arma3_scripting message

[regal nimbus]

Uh, where exactly are these diag_logs running from?

[ripe crystal]

Okay credit to @foggy shuttle for finding the issue but essentially this does not work as it previously used to anymore:

someFinishedScript isNotEqualTo scriptNull

Now if a script is <NULL-script> it will not equal scriptNull. I could be butchering this a little bit but I think that is because it still stores information about the script in memory now like whether it was terminated or not. Anyways the solution is simply just conventional null checking using isNull:

isNull someFinishedScript

[ripe crystal]

Also apologies for not getting back to you on this and seeing this late

[idle quarry]

Are fixes to the models something that's still being done ? I noticed you cannot retexture buttons on US parade uniform from art of war most likely due to model missing a selection. Went to feedback tracker to see if anyone reported it and it turns out someone did, so guess I'm mostly bumping this up.
https://feedback.bistudio.com/T167361

[gaunt depot]

Or traditional https://community.bistudio.com/wiki/scriptDone

[lunar star]

@untold sky WeaponChanged EH broken when added via config
https://feedback.bistudio.com/T193588

[daring wagon]

Also Spam in the same ticket

[hallow sun]

Can we get getVehicleLoadout & setVehicleLoadout, instead of having to mess with dozens of commands for turrets, pylons etc?

[summer crow]

Zeus placed AT mine and M6 SLAM mine will detonated by vehicle on original position after move the mine
https://imgur.com/a/WRsVsvq
https://feedback.bistudio.com/T193669

[unborn acorn]

It could be related with https://feedback.bistudio.com/T171095 but marked 'fixed' since 2.14

Fixed: Anti-Tank mines did not sync their position and damage in multiplayer (https://dev.arma3.com/post/spotrep-00111)
Can be reproed in Eden if setPosATL mine somewhere - mine's trigger area will stay in initial position

[vapid cradle]

Ive noticed something wrong with jpg tp paa conversion: EVERYTIME i try to convert a smdi file with correct naming ( ***_smdi.jpg) from JPG to PAA, the glossiness is applied on the entire model. Something breaks the conversion, yet when i open the paa with texView, the smdi looks good. I have to use PNG to have it working properly (and only with smdi, if i use png with _co or _nohq it doesn't work properly sometimes)

[limpid rune]

why on earth are you using jpg for smdi

you definitely should be using png or tga and then converting to paa for any real texture

SMDI relies on specific colour channel usage, god knows what jpg compression would do to that

[vapid cradle]

At the very beginning i was using png but some textures were sometimes transparent so i decided to use jpg to be sure there is no alpha. I'm not a developper of Arma 3 but it is veeeeery easy to read a jpg or a png in c+ or c# and vet each pixel RGBA values

[devout wave]

setParticleCircle causes inconsistent Z placement of created particles, probably an AGLS issue https://feedback.bistudio.com/T162981

[gaunt depot]

setTerrainHeight data is still randomly not arriving to JIP. Pictured, I joined the server, saw wrong terrain, then went back to lobby and its fine on next join

No repro of course but its been an issue since it was introduced

[heady thunder]

spüm
https://feedback.bistudio.com/p/Reda333/

[topaz flax]

I don't know if this is the right channel, but I have a little feature request for the say3D command. Would it be possible to give this command the ability to change audio volume midplay? Also a param to change the refresh rate of location check?

[tranquil harbor]

mostly an option to change/overwrite the volume would be enough.
playsound3d is a thing, yes but it is not really good for moving objects

[vapid cradle]

I found a bug today, units animation speed is being multiplied when they're attached to an already attached object https://feedback.bistudio.com/T193780

[gaunt depot]

    class FlareWhite_Illumination_F: FlareWhite_F
    {
        author = "Bohemia Interactive";
        ammo = "F_20mm_White_Illumination";
        displayName = "Illumination Flare (White)";
        displayNameShort = "White Illumination";
        descriptionShort = "Type: Illumination Round - White<br />Rounds: 1<br />Used in: Flare Gun";
    };
```gib flare gun 😠

[untold sky]

Thanks, fixed next prof. Didn't consider people would do such weird things, but we need to cover everything of course 😄

[regal nimbus]

Wiki did specifically say you could do it :P

[gray wharf]

sorry!

[limpid rune]

idk how much of a horrible mess it'd be but would it be possible for https://community.bistudio.com/wiki/playAction and it's 'now' variant to return the resulting animation instead of nothing? I know that sounds odd but the way the action lists work you can have for example a different gesture/animation prone or standing for example

I'm currently working around it by running gestureState immediately after but having the command return the resulting animation would be 'cleaner'

[gray wharf]

the wiki is always right, it's the code that's wrong 😄

[kindred halo]

Are we allowed to provide Arma III feedback on this channel?

[scarlet sage]

Use https://feedback.bistudio.com/
But don't spam impossible wishlist suggestions, it's more for bugs or small things.

[spice temple]

Hello, would it be possible to add something like client side/server side/client+server side status for mods to Steam Workshop? Similar to required for addons.

[cyan basin]

that's already a thing in workshop filters

i'm imagining it's costly to add ingame

[spice temple]

Where ?

[cyan basin]

mod and server

[spice temple]

Okay, so it's not actually there... because the mod clearly says that it's a mod, and it doesn't matter whether it's just client-side or client-side and server-side...

[kindred halo]

Define 'small things'.

[scarlet sage]

Look at https://discord.com/channels/105462288051380224/1372557207446618202 and see which things were suggestions.
What are you thinking of?

[kindred halo]

I dunno, something like...
A Battlefield-style Conquest mode (Sector Control is too big without vehicles at the spawn), or a 'control vehicle from any seat' mode like Easy Red 2. Stuff like that.

[scarlet sage]

The first is way to big and there's not really a point considering the amount of mission makers/modders on the workshop and how would the second be different from what we have with W/A/S/D driving orders and manual fire?

[kindred halo]

That's what 'manual fire' does? I see.
It wouldn't span the whole map, just a smaller area. Like one of the cities.

[scarlet sage]

It's the scripting that's hard. And the fact that there's nobody left to do it.

[kindred halo]

I can do it. I learned how to code in college.

[vapid cradle]

It would be really great if we could have an objIntersectObj and a terrainIntersectObj functions to make additional collisions for very large (multipart) vehicles. Of course we can do something with lineIntersectObj but its not very effective and consumes way more resources than a simple objIntersectObj function

An additional option to attachTo to activate geometry collisions to attached object could be a more simple solution

It doesn't have to apply mass, i think we can calculate that with getMass and getCenterOfMass

With these functions, Arma 3 could become a swiss knife

[vapid cradle]

It would also be a very useful way to determine what did hit the player or his vehicle

Wh"en the player drives a car for example and crashes on something, HitPart event always returns ObjNull so we cannot determine what did the player hit

IT would be great to be able to determine what the player hits with vehicles. Nowadays, if i hire a foreign army to make some military operations on my territory, if the soldiers drives like drunken addicts and destroy everything, i want to be able to identify what they destroyed or dammaged so i can make a bill for them

[gaunt depot]

You can get which object they collided with using EPEStart

It is quite tricky though and there is no realiable way to tell if crew inside got damaged because of collision to certain object or not if they're remote client crew

You can make it fairly reliable when dealing with local entities, but for remote crew there is no way as you get no HitPart not EPEStart, only HandleDamage with nulls and even if you RE the event it still arrives after units are already dead.

Damage system needs major overhaul to be able to properly track the damage, I'm afraid we're 12 years too late for it

[vapid cradle]

I just checked this solution, its a very good way to track collisions but unfortunately this event doesn't tell the resulting amount of damage for the collision. I need to know how much damage have been dealt so i can properlycalculate the price of the damages

[gaunt depot]

You'll have to sum the damage from null sources or self sources and then see if there was EPEStart event within some time frame (I used 1s)

Its complicated yes

[vapid cradle]

Hmm i see but that would mean i need to add a damage event to all map objects. That a lot of objects. I don't know if the game will handle it

[gaunt depot]

Wait you want damage to the objects, not to the vehicle?

Hmm, I wonder if EPEStart is called before HandleDamage, I think it should

[vapid cradle]

I want to check when the vehicle damages map objects

[gaunt depot]

You can add EPEStart to all vehicles, then on the event check if that object was hit for the first time, if yes then add HandleDamage to it

But HandleDamage only works on objects with HitParts iirc, not all objects

[vapid cradle]

Yes, that's a way

[gaunt depot]

Is there HitPart even on collisions? I thought it was only from projectile hits?

I don't remember details anymore, been a year since I did a huge deep dive into it, but I believe it should be possible to do what you want with current events, just very tricky to do it right and reliably

[vapid cradle]

I'm not sure i don't remember if hitPart or Damage triggers but i can get the damage from collisions from one of them

It will be a bit tricky but i think i can do it

[gaunt depot]

General logic:

1. addMissionEventHandler ["EntityCreated", {...}] with check if object is of carx/tankx/airplanex/helicopterrtd simulations
2. If its right object, create JIP RE so all clients add EpeContactStart event to it
3. In EpeContactStart check if hit object is a new one, if it is then add HandleDamage to it, also save vehicle that collided with that object and timestap
4. In that HandleDamage collect all null or same entity damages by whatever logic you want, mark the object to be checked at the end of frame or next frame
5. On next frame check if there was vehicle colliding with it and it wasn't too long ago (Since its EpeContactStart it only triggers at start of collision but it can continue for a bit dealing more damage frames later), at this point you know sum of collision damage and which vehicle caused it, you can act however you like from there

HandleDamage might not work on lots of map entities, you'll need some other event, I don't remember which will work, you'll need to add all damage events to different entities to see which trigger

[vapid cradle]

The problem is EPEContactStart fires after damage event so at firts EPEContactStart if i add a handledammage event i will miss the first damages

[gaunt depot]

Some basic stuff like fences and such bit not trigger any damage events, just death events

[gaunt depot]

That's why you need to do your decisions at the end of the frame or next frame

HandleDamage is multi call event anyway so you'll need some other place to act on damage sum anyway

Oh but HandleDamage might not be there in the first frame 🤔

as we're adding it on EpeContactStart

[gaunt depot]

https://feedback.bistudio.com/T177724

I also encountered this issue and it was apparently changed since 2.18

[vapid cradle]

No it won't be therre. When EPECOntactStart fires, all damages has already been dealt

[gaunt depot]

So EPEContactStart should fire before HandleDamage so you might be able to add it there before its even fired

[vapid cradle]

I also wish it was ^^

[gaunt depot]

I haven't checked if this change really occured as I adjusted my code to work without it

Can you check and report in that ticket?

[vapid cradle]

Here is the code i tested:

 params ["_object1", "_object2", "_selection1", "_selection2", "_force", "_reactForce", "_worldPos"]; systemChat str _this; systemChat str (getDammage _object2); 
}];```
When _object2 damage is displayed, the collision damage has already been dealt

Yes i'll report in this ticket

[gaunt depot]

Are you sure about it? Add both events with frame number logs

[vapid cradle]

Even without frame number logs, according the logic, if i type cursorObject addEventHandler ["EpeContactStart", { params ["_object1", "_object2", "_selection1", "_selection2", "_force", "_reactForce", "_worldPos"]; systemChat str (getDammage _object2); }]; I'm supposed to see the damage of _object2 before the collision

I just tested and i get the damage after collision

[gaunt depot]

No, it should first report the collision, then fire damage event

That's right, you need to add damage event on collision event

It was the other way around before 2.18, damage before collision event, but now its in proper order

[vapid cradle]

Ok, i'll check with damage event

[gaunt depot]

15:33:35 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||
15:33:35 "Frame 31221, EpeContactStart: C_Hatchback_01_F"
15:33:35 ["1b461430080# 1816214: hatchback_01_f.p3d C_Hatchback_01_F (C_Hatchback_01_F)","B Alpha 1-2:1 (Sa-Matra) (B_G_Offroad_01_F)","","",485.491,[-240.677,414.597,76.7167],[23555.3,18041.7,3.93274]]

15:33:35 =========================================================
15:33:35 "Frame 31221: HandleDamageVehicle: C_Hatchback_01_F (ALIVE=true)"
15:33:35 ["1b461430080# 1816214: hatchback_01_f.p3d C_Hatchback_01_F (C_Hatchback_01_F)","",0,"<NULL-object> ()","",-1,"<NULL-object> ()","",true,0]

15:33:35 |||||||||||||||||||||||||||||||||||||||||||||||||||||||||
15:33:35 "Frame 31221, EpeContactStart: B_G_Offroad_01_F"
15:33:35 ["B Alpha 1-2:1 (Sa-Matra) (B_G_Offroad_01_F)","1b461430080# 1816214: hatchback_01_f.p3d C_Hatchback_01_F (C_Hatchback_01_F)","","",485.491,[240.677,-414.597,-76.7167],[23555.3,18041.7,3.93274]]

15:33:35 =========================================================
15:33:35 "Frame 31221: HandleDamageVehicle: B_G_Offroad_01_F (ALIVE=true)"
15:33:35 ["B Alpha 1-2:1 (Sa-Matra) (B_G_Offroad_01_F)","",0,"<NULL-object> ()","",-1,"<NULL-object> ()","",true,0]

Got curious and went to check, I drove offroad into hatchback and indeed it is still a problem if you have an event only on one vehicle, you need events on both to reliably connect EpeContactStart to HandleDamage

[vapid cradle]

Aaaaah i knew something was going wrong

I just tested this code cursorObject addEventHandler ["EpeContactStart", { params ["_object1", "_object2", "_selection1", "_selection2", "_force", "_reactForce", "_worldPos"]; _object2 addEventHandler ["Dammaged", { params ["_unit", "_hitSelection", "_damage", "_hitPartIndex", "_hitPoint", "_shooter", "_projectile"]; systemChat str _this; _unit removeEventHandler [_thisEvent, _thisEventHandler]; }]; }]; And it doesn't throw anything which means Dammaged event doesn't fire after EPECOntactStart

[gaunt depot]

I guess your only option is to scan for objects near relevant vehicles to check if they're new objects that need events setup for them

Which objects do you care about anyway? Trees? Fences?

[vapid cradle]

Most likely all dammageable objects

The other option would be to apply the vent to all object on mission preinit, maybe i could be fast enough

[gaunt depot]

Lots of world objects like fences and trees are simple objects, you can't add any events to them at all

Okay, here is another approach, one EpeContactStart add hit objects into an array with damage at start of collision, then check if damage increased at the end of the frame for all objects in that array

If it did, means your collision did it

You probably can even used EpeContact for it, no time stamps needed

We should probably move to #arma3_scripting

[vapid cradle]

The thing is at the start of event EPEContactStart, damage has already been dealt

Ok let's move

[vapid cradle]

New ticket added for EPECOntactStart firing after Damaged event on terrain objects: https://feedback.bistudio.com/T194024

[gaunt depot]

You can't add any event handler to simple objects at all, most terrain objects are such

Otherwise not sure how this can be solved, have all EPE events fire before HandleDamage? Might need large engine modifications

[vapid cradle]

I just realised that i actually works with terrain objects but only with those declared in the config (those who can have events). You are right, it would require to make a config for each terrain object

[gaunt depot]

Yes but it still wont be reliable because of unreliable EPE events order

[vapid cradle]

Actually even order probably fires properly on terrain objects too but we just cant be sure because we can't add events to them

[vapid cradle]

Added new feedback https://feedback.bistudio.com/T194104

[untold sky]

Okey but what actual impact has that? Is that actually a problem worth solving?

[vapid cradle]

I wanted to apply a system similar to Advanced Train Simulator mod in order to make horse being able to transport carriages. Some carriages require multiple horses to be attached one to another in order to carry them.

But its not mandatory, i can find a way around

Most people will probably not have to use playAction or playMove this way i think

[desert trench]

under what circumstances is a spawned function to live/continue beyond mission end on a DS? [Server: SetServerState to MISSION ABORTED]
is there any, would this be a bug?

1. execution line

// init.sqf
[] call SPE_CampaignUtilityFunctions_fnc_MissionStart;
[] spawn SPE_fnc_animatedBriefing;
[...] spawn spe_missionUtilityFunctions_fnc_animatedBriefing_init;
[...] spawn spe_missionUtilityFunctions_fnc_animatedBriefing_local;
[...] spawn spe_missionUtilityFunctions_fnc_eventTimeline;

2. CODE gets executed even when mission was ended by #missions or #reassign

// Main loop.
while {spe_missionUtilityFunctions_eventTimeline_pointer_local <= _frames && (missionNamespace getVariable "BIS_fnc_eventTimeline_play")} do {
    if (_syncToMusic) then {
        ...
    } else {
        waitUntil {
            if (missionNamespace getVariable ["spe_missionUtilityFunctions_eventTimeline_skip", false]) then {
                ...
                while {(missionNamespace getVariable ["spe_missionUtilityFunctions_eventTimeline_skip", false])} do {
                    ...
                };
            };

            CODE
        };
    };
};

is this to be expected/intended?

[untold sky]

It can run during debriefing screen i guess?
I think the servers cleanup only runs when actually starting a new mission

[desert trench]

Debriefing we use actually for the audio voice over.
I would have expected the "mission end" would halt the simulation, but its seems indeed until the next mission gets selected.

Usually this should be a none issue - i only noticed as due to diag_log debug code ran on each frame and thus the rpt grew even after "mission end" a great deal.

[surreal bough]

I found this error: if you change the size of the hud from the main menu of the game (settings, video), then the game will automatically restart... but if you do this from the game itself, then the game will simply turn off - the restart will not occur

[wary ferry]

I have scenario where i can consistently reproduce a hang, but im not sure if it is a issue with the modded terrain im using or if that specific location just happens to provide some edge case that was not present on Altis. Or perhaps you want the issue on the tracker regardless?

I am calling a custom function (spawn) that is creating arty shells at random positions around the player object. On Altis i can spawn a lot of these loops (100+) and still not notice any issues, but on the location on the modded terrain i always get the error after 5-20 instances (a bit random due to random sleeps in the script).

Any way to figure out if it is the terrains "fault" or if i should do a bug report? (i got crash dumps, a scaled down mission, and steps to reproduce)

Exception Code: 0xc0000005
Exception Info: The thread tried to read from or write to a virtual address for which it does not have the appropriate access.
Heap Info: Not Present

[regal nimbus]

Send mdmp to Dedmen and then see if he needs anything else. Not sure if the automatic reporting is working at the moment.

[wary ferry]

alright, will do!

[vapid cradle]

New feedback added so we can have an option to see backpacks in vehicles https://feedback.bistudio.com/T194376

[azure crane]

This has been an "issue" for so long I doubt it can actually be added.

[vapid cradle]

I dont think its complicated thing to add the backpack to the character in vehicles

New feedback added for attached objects sliding away https://feedback.bistudio.com/T194382

[limpid rune]

at one point there was a bug where dead units would show them 🙃

[untold sky]

I remember fixing that 😄

[slim echo]

I don't know if the feedback tracker would be appropriate for this; more of a feature request than a bug.

Who do I have to pester for the launcher to tell me how much storage the current mod list uses? That would be super useful information to have.

[surreal bough]

I want to remind you that the AI ​​skill is broken, we play with an AI skill of 0
here is a small example

[alpine tulip]

How this can be a example of any in-game issue when you have massive amount of Mods, and likely an AI Mod as well?

[surreal bough]

we don't have mods on ai and in another topic i provided screenshots that skillfinal is zero

[untold sky]

I need a vanilla repro in a ticket

[surreal bough]

well it's unrealistic to do this, we conducted regular tests in the desert on vanilla - everything is clear there, you can do it all yourself... I should rewrite the project on vanilla, in order to prove that it's not even a fact that you will edit... such a thing
my job is to talk about the problem - your right to see it
it's easier for me to prove that I use simple popular mods and provide a mission, I can provide access to the server
but I know perfectly well that if I have a problem, it doesn't mean that everyone has a problem)
the AI ​​skill is broken - I claim this, there is no evidence, unfortunately I have no desire to prove it either
so just a video fact from the player

[regal nimbus]

I'm not sure what you expect. In that situation the guy outside has visibility of you long before you have visibility of him, so he's not reacting that quickly. And then they can always hit with the first shot if they're lucky.

[surreal bough]

Didn't you see that he was working in the other direction and instantly turned around with the skill at zero?

[regal nimbus]

Turn speed doesn't seem to have any dependence on skill.

At least they'll do 45 degrees pretty quick on zero.

hmm

There might be an inversion in here though

They're killing me faster on 0 skill than on 0.6 :P

never mind, probably just because I changed them to combat mode and they dropped prone.

[proper grotto]

Then what does that mean?

[regal nimbus]

That's under precision not skill.

Also not sure if that one changes visible turn speed either. Haven't checked yet.

[surreal bough]

I find it very strange that I was the only one who noticed that the skill changed at some point... it turns out that I'm tilting at windmills I have the ability to run a mission on ancient hardware, but I no longer have the ability to run it on 2.16

[proper grotto]

Well, precision in this instance is also 0.

[surreal bough]

I did tests on 2.18 just in the editor placing vanilla units - the result is the same: at 500 meters in the desert on Altis they oneshot the player at the spawn

[regal nimbus]

uh, that's a very different scenario

Maybe actually pin down a situation that you can replicate and then I'll test it.

[proper grotto]

The wiki says, that its Skill though

[regal nimbus]

Can't see any difference with either. They just turn pretty fast through 45 degrees.

[proper grotto]

I'd say it's not only about turning.
The AI was just shooting someone, so it has to spot a new target, switch to it, aim at it, fire at it, hit it. And all that he does in less then half a second with 0 skill.

[regal nimbus]

Spotting is waaay slower at low skill. But in that example you might already be spotted and then it's a target switch.

Amusingly they seem to be relatively likely to hit with their first few shots at low skill. After that they're waving all over the place.

It's shitty logic but I doubt it changed.

Checked 180 degrees, can't see any skill difference in turn rate there either.

[mint raptor]

server browser only works once

btw

[verbal harness]

could you be more specific?

[mint raptor]

Sorry. I just did some testing so I can provide more accurate information.

It's not as simple as I previously stated. It seems like it might be related to rate-limiting or anti-spam/ddos measures.

Steps to reproduce:

1. Launch A3 and go to server browser in-game.
2. Either join a join a server quickly and return to the browser or refresh the browser a couple times.
3. The browser now either only loads a few servers or none at all.

Browser functionality does not return unless you restart your game.

[unborn acorn]

FT ticklet about squad.xml picture glitch in shadows https://feedback.bistudio.com/T194406

[twin hornet]

Bug with Projectile impact position deviation between different players in multiplayer is back in the game.
https://feedback.bistudio.com/T194410

[regal nimbus]

I don't normally use it, but curiously the "internet" tab doesn't show anything at all for me at the moment. However, the official servers tab is behaving similarly to what you describe.

First click, full list. Switch to internet and back again, half list. Switch again, nothing.

Launcher will happily grind through several thousand servers.

Perf branch, for what it's worth.

[untold sky]

Some routers or firewalls do that. They detect the replies coming back from the servers as a DDoS attack and block it. Not our fault

[regal nimbus]

I don't think that's what's happening here. Launcher is doing the same thing, isn't it?

[untold sky]

Launcher and in-game are doing the same thing yeah

[regal nimbus]

Launcher gives me 4k internet servers and >200 officials each time. In-game currently gives me no internet servers and an official list that reduces each time I switch to it. Feels more like a logic bug.

[mint raptor]

Yes, the launcher server browser has different behavior. I can refresh that all day and not get locked out, but it does load the servers slower from what I can tell...

[regal nimbus]

Also restarting the game resets the behaviour but waiting doesn't, which suggests that it's not router or ISP level.

[untold sky]

On where? dev/profiling branch only?

[regal nimbus]

2.20 stable behaves the same as perf. I don't have the download bandwidth to try 2.18 and I'm not sure it'd be meaningful anyway given that the servers are all 2.20.

[untold sky]

Ayaa.
My suspicion would be the netcode stuff, but thats not in 2.20

[regal nimbus]

I don't normally use the server browser in-game so I'm not sure how long it was broken for.

[surreal bough]

it doesn't let me access my own server and says this

18:54:10 Client: Dimon UA - Kicked off because of invalid ticket: Ok - Ticket invalid

a very strange situation: recently one problem appeared - my server is displayed in the launcher twice, this has never happened before... I think it is somehow connected with the fact that the server is located at my home, ok (but this has never happened before and I have not changed anything in this direction).
if I go to the server that I went to - I start to get chains - yellow>>>red>>>break
I go to the server that appeared new - I played for a couple of days normally, but today it kicked me out and that's it - I can't go anywhere anymore

Client Dimon UA - client's ticket has become invalid. Code: 6

[stone tree]

https://feedback.bistudio.com/p/wsmhjqq123/

Spam account

[vapid cradle]

I wonder what's their objective doing this on this website. Most A3 players dont even know about this website. So what's the objective ?

Sharing the info with other bots ?

[devout wave]

It's not targeted at the FT specifically. They just detect a system where you can create a post, so they create a post to try to promote whatever dubious websites are on their list.

[untold sky]

google results

[vapid cradle]

Hmmm

[snow cairn]

the amount of errors in BIS_fnc_3DENModuleDescription is the reason why its been broken since 3DEN was released...

line 37
line 52
line 53

[daring wagon]

What errors?

[surreal bough]

What do these lines mean and the reason for their appearance?

9:16:45 Skipping mods and signatures due to overflow flag being set.
 9:16:49 Skipping mods and signatures due to overflow flag being set.

[regal nimbus]

Means that you have too many mods for the steam protocol packet, so for example the launcher doesn't get a complete listing of what mods the server is running.

The steam protocol packet size can be increased in server.cfg but it requires everyone to have UDP packet fragmentation support.

[vapid cradle]

I wouldn't like to give you developers of Bohemia too much work, but do you guys think making a createUnitLocal function could be possible ?
Because right now, there is absolutely no way to have a dialog with a unit on a server. I mean if two or more players try to talk to a unit at the same time, the unit can answer only one time because each computer cannot play animations locally for the unit. The only way would be to create 3 units, make one of them visible only for each computer then play dialogs and animations. But that's a huge charge for the server in my opinion even with headless clients.
It would be great if we could add some sort of RPG experience to current Arma multiplayer experience

[untold sky]

Possible sure. But we won't do that.

"that's a huge charge for the server" having 3 units, that you can disableSimulation on every machine except the player that sees it. Doesn't sound like much of a problem
I don't see how headless clients are relevant, you wouldn't put the units on the headless client. You spawn it on the client that uses it, and delete it after

[vapid cradle]

Hmm, i didn't realize that if i spawn a unit on a client machine, the client machine will handle the unit instead of the server

[solid marten]

Not to mention you can hide unit on all other machines but one

[vapid cradle]

I intended to hide units on other machines but i thought it would still require resources from the server

[tepid fractal]

Spam
https://feedback.bistudio.com/T194480#2859521

[vital junco]

When submitting a bug report, is it possible to embed an image without linking to an external website?

[scarlet sage]

Yeah

[vital junco]

well how

[devout wave]

Click the Upload File button - the rightmost one on the left side of the formatting toolbar, shaped like a cloud with an up arrow

[vital junco]

found it, thanks

[gaunt depot]

https://feedback.bistudio.com/T178958#2860975 🤖

[untold sky]

I never wanted to touch AI driving because it would be a pain and backwards compat issue.
So now that I'm trying to touch AI driving, its even more of a pain that I thought.
The problem is they try to turn in place, which they can't do in water. Giving forward and backward thrust at the same time and thus not making any progress.
So I add a water check, and now they just drive up the river (instead of across) constantly wobbling left/right. And if they manage to touch the coast, they do a 180° in-place turn, and flop back into the water, instead of driving up the river bank...

[devout wave]

Wheeled vehicle AI must have some solution, because they are "OK" in water and can already handle not being able to turn in place. Could you have tracked vehicles switch to using wheeled AI while in water?

[untold sky]

They already do

[devout wave]

Oh, okay

[untold sky]

Tank Steering component, is Car Steering component but with "can turn in-place" turned on

And some differences in how the turn input is converted to thrust :/

[devout wave]

About the turn-and-flop-back-into-water behaviour, I'm just curious which vics you're testing with? There are certain modded amphibs that have some kind of scripted(?) system that forces this behaviour even for players, e.g. 3CB AAVP-7A1

[untold sky]

The GM M113. I have no problem as a player when I just give thrust forward.
I can see in diagnostics that the AI vehicle is also giving thrust forward, but making much worse progress 🤔

Also not fun to test when every test run has it behaving differently 😄
I just got it to work. It zigzag'ed across the river for 2 minutes, then it hit the coast with so much inertia that it slipped all the way out of the water, and then it pulled itself the rest of the way up and made it to the waypoint.
But that was still more of an accident.
This transition region between water and not water is nasty

Cars have it easy, they give forward thrust and steering wheel. But Tracked vehicles have no steering wheel to steer. I can force them into forward thrust (Thats what I tried where it made it) without steering, and it'll somewhat work but it still too bad to be usable

[untold sky]

Huh. I think I got it.
Just keep randomly guess-swapping conditions and numbers and tadaa, suddenly it can do a 180° turn in water aiming straight towards the waypoint and drive up the river bank.

But not entirely sure if I actually got it or I'm just lucky. There seems to be a chance that after driving up the bank, they suddenly decide to turn back around, drive away from the waypoint and land back in the water

And that all is IF they don't decide to do a backflip when touching the water and sink upside-down underwater and eject all crew

The thrust behavior is very weird and I don't understand it.
They want to make a right turn, so they give 100% foward thrust on the right tracks, and 100% backwards on the left tracks..
But, that does a left turn not a right?

But somehow, when above water, this math does work out, somehow?
Maybe the code is just mislabeled..

Okey I changed nothing, and tried to record a video of it working nicely.

1st did a backflip into the water, but ended upright, nicely drove to coast, then.. just got stuck at the cost, somehow not enough thrust to get up?
2nd backflip upside down in water, crew ejected
3rd normal into water, drive to coast, stuck cannot get up
4th normal into water, drive to coast, get up the coast, do a 180° back into water.
5th normal into water, steer to other coast, get up, slightly stuck but pulls itself up slowly, successful. But.. I didn't record because I expected this to fail again

Darn it

[limpid rune]

well they almost cleared the river with that backflip

[untold sky]

Ah! Flawless river navigation. Bug fixed! Ship it! (Hopefully QA is lucky when testing it)

The water navigation is pretty good now I would say. If only it would actually keep going when it touches land, then we'd be perfect.

[untold sky]

Do note how in a right-turn, if gives full thrust on right but reduced thrust on left. That seems like nonsense, but it works and it does a right turn?

[uncut briar]

@lavish cargo ⏫

[lavish cargo]

thats just how wheeled vehicles work, if you want to turn your car around you have to drive forward, the tank can spin in place, which you cant do in water because the rudder wont work

[devout wave]

Yes, I know. My point was that the AI intended for wheeled vehicles must be able to understand that, so if it was possible to make the tracked AI think like wheeled AI while in the water, it might work better. But Dedmen says it already tries to do that, so it's obviously not that easy 🤷

[untold sky]

I'll put it into prof tomorrow and you can see if thats good enough. I only tested one vehicle on one river 🤞

[regal nimbus]

For what it's worth I think there's a step in the 3-point-turn logic that also steers backwards, so they repeatedly turn back into the same wall rather than making progress.

[untold sky]

I saw that once it getting stuck at that.
Forward a bit, backward a bit, forward a bit, backward a bit. Forever

[regal nimbus]

once :D

wheeled vehicles IIRC

[untold sky]

On the topic of mine simulation performance.
Mines stream-in terrain objects when checking for possible triggers (When the mine is far away from player)
Streaming-in terrain objects alone, is already 20% of the simulation cost.
And.. Mines cannot even be triggered by terrain objects, so why load them in the first place.

[verbal harness]

Maybe to prevent a mine from going off if there is a terrain object (wall/corner of a house) between the mine and a vehicle?
or is it just a dumb distance check?

[untold sky]

There is no line intersect for radius trigger mines, only the wire trigger mine uses line intersect
Can't see anything that would make a wall block detonation, it should see through it

[regal nimbus]

Hmm. Doesn't that terrain data persist between frames?

Not like the mines are going anywhere

[untold sky]

I don't know when it gets unloaded.
But the locking of the cell alone is taking time, and as far as I can see the unloading happens atleast alot. The cells are far away

[regal nimbus]

Huh. I figured it only unloaded terrain when it ran out of RAM.

and then it'd be some LRU system

[untold sky]

2000 mines on Altis. Before/After. Seems good.

[gray wharf]

what was that big red loading time there :o

[daring wagon]

That's the mines?

[alpine tulip]

https://tenor.com/view/mine-finding-nemo-mine-mine-mine-thats-mine-lulugifs-gif-20827668

[steady panther]

Oh wow! @agile trail did you see this too?
Something to wait for 😄

I guess the river bank there in the test case is a bit harsh for any vehicle

[untold sky]

Please do testing of that and report if its good, its available on prof branch

[agile trail]

I've always assumed there is a change of simulation type, or at least physx interaction system, at the boundary between water and land, that has caused the typical "turn-around 180 and go back out to sea" behaviour. Almost like one side of the vehicle gets traction on the ground before the other side, and so spins it back out.

[surreal bough]

Well, landing infantry from landing ships is also a problem... I won't even mention the veiponts, but at least here you can stick the ship on land and give the order to disembark.

while {(alive _vehicle)} do
        { 
            if (speed _vehicle < 10 && {(surfaceIsWater position _vehicle)}) then {
                private _vel = velocityModelSpace _vehicle;
                private _vel1 = _vel select 1;
                _vel set [1,5 + _vel1];
                _vehicle setVelocityModelSpace _vel;
            };
            sleep 0.2;
            if (speed _vehicle == 0 /*  && {!(surfaceIsWater position _vehicle)} */) exitwith {
                _time = time;
                waituntil {(speed _vehicle == 0 && {(time - _time) > 20}) || !alive _vehicle}; 
                if (!alive _vehicle) exitwith {};
                private _doors = (configProperties [configFile >> "CfgVehicles" >> typeof _vehicle >> "AnimationSources", "isClass _x"] select {getnumber (_x >> "animPeriod") >=0.1}) apply {configname _x};
                {_vehicle animateDoor [_x,1]} forEach _doors;
                {_vehicle animateSource [_x, 1, false]} forEach _doors;
                {
                    moveOut _x;
                    unAssignVehicle _x;
                    sleep 0.5;
                } forEach (crew _vehicle);
                {_vehicle animateDoor [_x,0]} forEach _doors;
                {_vehicle animateSource [_x, 0, false]} forEach _doors;

because the boat can just stand still and do nothing, and you have to push it towards the shore

And it's also very strange that if you set the waypoint to disembark on the island, let's say in the centre, all boats will go to one point on the island to disembark.

[surreal bough]

https://community.bistudio.com/wiki/setCustomAimCoef
How does this command work, and does it work at all? No matter how I change it, there is no effect.

[daring wagon]

Works as expected

[regal nimbus]

Usually works fine? Like setting a value of two doubles the sway.

Main issue is that there's no stacking, so if you have two or more mods that try to use it then they'll fight.

[clear glen]

@untold sky toying around seeing if I could make some unmodifiable hashmaps using createhashmapobject using ["sealed","noCopy"] in combo with compileFinal and ran into this bug. In case it's an easy thing or if it's even worth your time. https://feedback.bistudio.com/T194855

[cosmic field]

Is it suppose to work like that ? No mods, just vanilla. The problem is that 40mm grenades doesn't hit the mark

[modest sonnet]

You're shooting 40mm over 1km, it's probably normal

[foggy shuttle]

Yes, that's normal.

[static imp]

Uh, sorry if posting wrong but this URL does not exist.

[gray wharf]

yep, should be https://feedback.bistudio.com/. @untold sky?

it should either be changed or redirected ^^

[static imp]

same on the BI website, outdated link under "Arma 3 Feedback Tracker"

https://support.bohemia.net/

[gray wharf]

@jaunty narwhal ?

[jaunty narwhal]

That re-direct should just work, but it doesn't anymore, so we'll report it to the web team.

[jaunty narwhal]

Should be fixed.

[gray wharf]

@static imp ↑! 🙂

[gleaming plume]

Interesting issue I've discovered earlier today - tracked vehicles that are completely empty on fuel can still be neutral-turned to the left and right, using the mouse (with vehicle freelook disabled).

I get that wheels on wheeled vehicles can be turned when vehicles are static and engines off - this should not happen for tracked armoured vehicles.

[regal nimbus]

hmm, I wonder if that's related to the desync you often get from disabled AI-driven tracked vehicles.

[vale maple]

The Issue

1. It starts when a player crashes/disconnects in an ungraceful manner and then tries to rejoin immediately.
2. After this happens and they rejoin, they'll get kicked by Battleye (Query Timeout) after exactly 2 minutes.
3. This will persist until they either change their IP address (modem/router restart for some users and VPN/connection change for other users) or the server restarts.

Solutions

• Changing IPs always resolves the issue.
• A server restart will always resolve the issue.
• If you wait ~90 seconds before rejoining after the ungraceful disconnect, you can avoid the issue from starting.

Attempts to resolve

I've taken some steps to attempt to resolve this including:

• Changing datacenters.
• Changing machines.
• Changing hosting provider.
• Allowing traffic through the BE ports in windows firewall in a clean install of Windows Server 2022.
• Updating from Windows Server 2016 to Windows Server 2022.
• Updating the server directory via steamCMD to make sure BE is up to date.
• Adjusting the kickTimeout in server.cfg to 90 seconds.
• Creating a playerUID based kick at the first client state picked up by clientStateChanged hoping it was before the BE cache gets corrupted.

Best Guesses

Seeing that the IP change or server restart always resolves it and allowing your player to fully exit the game in a crash/ungraceful disconnect situation stops it from happening, my guess is that there's some sort of IP address based Battleye cache that's running and it's failing to register the new player after they have an ungraceful disconnect and rejoin before they're removed.

[regal nimbus]

Two minutes after the rejoin or two minutes after the disconnect?

[vale maple]

Two minutes after the rejoin, every time they rejoin until they either change their IP or the server is restarted

[modest sonnet]

Long-standing problem

[vale maple]

Oh I know, we've been dealing with it for 5 years about 25 incidences a day. We've been able to narrow it down to at least the IP keyed cache being the problem with consistent reproduction/avoidance/solution so I'm hoping that's enough for a fix. Adding a check to see if a player is already registered before they get "out of sync" in the BE cache (or a way to kick them if they're already in without "poisoning them") would likely solve it but that's assuming it's even in their hands (and not Battleye's).

[modest sonnet]

Raise a ticket if there isn't one already I guess, there is a battleye category. I'm not sure on who's end it is to fix but maybe it will poke someone

I was trying to write one but I couldn't really reproduce it easily, it sounds like you have some data though

[regal nimbus]

saveProfileNamespace and saveMissionProfileNamespace mangle nils in hashmaps. Nils in arrays are fine. Example:

_testHM = ["key"] createHashMapfromArray [];
profileNamespace setVariable ["testNil", _testHM];
saveProfileNamespace;
profileNamespace getVariable "testNil";

[clear glen]

confirmed. But ```sqf
_testHM = createHashMapfromArray [["key",nil]];
profileNamespace setVariable ["testNil", _testHM];
saveProfileNamespace;
profileNamespace getVariable "testNil";

works fine

or at least it is returning 'any' (as opposed to initial <null>)

[regal nimbus]

Reminds me of VB's seven different types of nulls :P

[clear glen]

I suspect it has more to do with using empty array for values vs explicitly defining as nil

and yeah.. VB was.... something

[regal nimbus]

IIRC my initial failure case was with something like _hashmap set ["key", _hashmap2 get "key"], where the latter didn't exist.

you'd think that was equivalent to ["key", nil] but apparently not.

[clear glen]

_hashmap set ["key", _hashmap2 getOrDefault ["key",nil]] maybe??

I saw your correction... though I kinda knew what you meant

you can also merge

if you have many keys to set

[regal nimbus]

yeah that's not really the point. Arma's full of bugs you can workaround, but I feel like if a hashmap is valid in SQF it's still supposed to be valid once bounced through profilenamespace.

Maybe a sign of a worse problem somewhere.

[clear glen]

Yeah that is odd a non-existant key value is not nil.

[gaunt depot]

A new type of nil discovered! nil-nil!

[ripe crystal]

Huh, I never knew nil in hash maps caused issues in the profileNamespace. Pretty sure we did have an issue where players were losing hashmap data all together on the profileNamespace. I chalked it to a procedure of mine I had which sanitized keys and values for code, and perhaps that was still it but seeing this maybe not.

I have always thought it odd that nil is able to be stored for a keys value. In any case where we would use nil with a variable, it would delete. Perhaps because nil is still technically a value represented by the engine as a lack of one that's why it can be assigned to a key? Not sure I am smart enough to understand the why's of that but very interesting nonetheless

[gleaming plume]

Sorry gang - did anyone at the dev team see this?

[devout wave]

if you can reproduce it, make sure to make a ticket https://feedback.bistudio.com/project/view/1/

[nimble veldt]

See error in logs. Seems to be something is broken in 'HideTerrainObject' module. Windows server, stable branch

xx:47:57 Error in expression <_pivot","_a","_b"];

private _radius = (_a max _b) * 1.42;
private _objects = []>
xx:47:57   Error position: <_a max _b) * 1.42;
private _objects = []>
xx:47:57   Error Undefined variable in expression: _a
xx:47:57 File A3\Modules_F\Environment\HideTerrainObjects\init.sqf..., line 58

[eager trout]

Dumb question for the group. I found that any attachment to the top rail on the SPAR-16 (so basically any 3rd party laser/light combo) is misplaced and sits too far left on the rail. I made a bug report on the tracker, but now I'm wondering if this is a CBA joint rails issue or a BI issue? All of the BI PEQs sit on the side rail, so there's no way to test this except using 3rd party attachments.

[scarlet sage]

I think it's either a CBA issue or an issue from whatever attachment you're putting on.

[eager trout]

It's all attachments (RHS, CUP, T1, TOTT, etc). That's why I wasn't sure if it's the SPAR model or CBA that's causing the issue.

[scarlet sage]

I think it's that those mods all have decided on a singular size the make their guns and attachments, so those attachments only work on those guns.

[deft marsh]

This was a much used feature for terrain peoples that's been broken for a while, if you'd be so kind as to pass it to the tools dev's that'd be sweet, thanks ! - http://feedback.arma3.com/view.php?id=27613

On another note, im getting some impressive results with terrain processor +1 😃

[surreal bough]

The standard Evolution mission, like ours, is all tied to blowing up the antenna conditionally... here are a couple of very old problems: if a tree has fallen next to an antenna or an antenna in the wall of a building, at least by some edge, or just in the building, then the tree or the building takes all the damage from explosives - the antenna cannot be blown up

[daring wagon]

So what exactly is the issue?

[regal nimbus]

Could do with a CfgPatches attribute with a list of addons that should block loading this one. Example use case is building on top of both ACE compat and non-ACE versions of mods. Otherwise some cases currently need __has_include and non-binarized config.

[rapid badge]

@deft marsh Sounds like Terrain Processor was a good use of tools dev time/prioritizing in terms of what it enabled for community terrain makers...

[deft marsh]

well there are plugins for GIS software out there that do a bunch of stuff, but it definately wasnt a waste of dev time polishing it a bit for release

just need better vegetation to make the most of it now 😄

[ripe crystal]

Not sure how possible or feasible this is but we are kind of running into a unique problem. Ever since HTTP downloading, we've added a lot more to our servers mission file. Right now the current size is ~120MB. Every time we push any kind of mission update (content or hotfix), we have always generally used a new mission name for version histories sake.

The problem is now those old mission files are probably adding up a lot on people's hard drives and I doubt people are aware that is space being taken up that they should probably be clearing. Is there a way that they could be notified upon old missions taking up old space? Alternatively I suppose older missions could be auto-deleted but that sounds kind of risky.

Worst case we could probably just rework our CICD to use a static mission name but figured still might be something worthwhile bringing up

[unborn acorn]

Portable Flagpole - script error after use undo in editor and issue with "FlagTexture" control attribute (default value not detecting) https://feedback.bistudio.com/T195033

[mortal tendon]

Okay, I've done my job of reporting, I hope it's worth it 🙂

[regal nimbus]

Could do with a way to semi-optimally render polyline user markers onto map controls. drawPolygon doesn't work atm because it completes the line. Spamming drawLine is slow.

[snow cairn]

reporting what?

[dusky tendon]

_vehicle setVehicleAmmoDef 1 can remove unit weapon in cargo position...

My EH WaypointComplete patrol for the first waypoint involved restoring ammo for vehicles. I tested the situation and determined that all passengers were losing all their equipment except for clothing and first aid kits. As a result, I had enemies running around the combat zone without weapons.

commander same lost weapon

[faint fossil]

https://feedback.bistudio.com/T195065

[unborn acorn]

spaм https://feedback.bistudio.com/T195033#2867709

[vale maple]

Made a new feedback tracker post for the Query Timeout since all the existing ones had almost no actual information on the bug or were just people ranting about it. https://feedback.bistudio.com/T195080

Anyone who has any further related information please chime in there to help aggregate the info although we have a pretty good understanding of it already. Hopefully this is enough for a fix/workaround for one of the longest standing BE related issues plaguing the community for as long as I can remember. The problem itself seems pretty straight forward but I do worry that BI doesn't have access to fix the BE side of it so perhaps one of the proposed workarounds will fix it. All we really need is a way to kick players before they get registered by BE and plagued by the timeouts.

[sacred river]

Check https://feedback.bistudio.com/T180455

[vale maple]

I read this one when searching. It was very incomplete as far as information about the actual problem goes. Most of what we've learned about the issue wasn't even mentioned in it. This isn't a random issue related to hardware like many people seem to assume (or at least, the variant we've been dealing with for years isn't). It's a very specific engine/BE issue with consistent reproduction steps related to ungraceful disconnects and order of internal events when joining a server (something possibly in BI's hands and not locked away by BE)

[sacred river]

I cannot really confirm the steps to reproduce the issue. Around 90% (if not more) of BE Query Timeout issues are related to player hardware. The rest might be connected to connection issues. But most of these issues we had were solved when players actually used better drives. But as Dedmen said in the ticket that he won't do any changes, I actually just made a tick on the issue as resolved in my brain.

[vale maple]

Our variant is very reproduceable and certainly has nothing to do with any individuals hardware or connections. It's very clearly directly related to rejoining too quickly before BE updates it's own player hashmap when the player disconnects ungracefully. This means that assuming they have some form of control over the order of operations when joining a server, specifically when in that process BE registers the player, this could be solved by BI without needing BE.

For example, by updating the kickTimeout config option to kick before BE registration. This may not be so simple in such a legacy engine but I haven't seen anywhere near the amount of information we've collected on this issue anywhere else and our data suggests what the problem is quite clearly and potential solutions.

[desert trench]

Mission cycle in combo autoselectmission=true; [as best as we could tell] can also lead to BE Query Timeout in some way.
Its very reproducible on the official BI servers for SPE.

Dedmen tried to find the root of the issue last year, and fixed some potential causes, yet unfortunately it didnt help in the end.

Overall I'd assume as the message implies some handshaking between BE and DS and/or client in certain conditions doesnt work out.

Its a shame it hasnt gotten more attention, ie more extensive logging may give more clues, as going by FT/BIF/reddit/etc post, the issue seems to have been quite prevalent for many years with A3.

[hallow sun]

These BE issues are a big turn off, don't wanna know how many people rage quitted because of that alone over the years.

[fast lodge]

Do you have to turn off battleye so nobody will get kicked or is it enough to turn off the verify Signature feature?

[untold sky]

compileFinal creates a copy of the hashmap, and the copy is then final-ed.
The bug is that during copying, only some flags are copied over, not all

Specifically the NoCopy and ReadOnly flags are skipped.
Sadly I didn't leave a comment, and don't remember why I skipped the NoCopy one.
I don't know why you would want a copy of a non-copyable hashmap to become copyable 🤣
So its probably fine for me to just fix this.

Also compileFinal copying the hashmap is inefficient. I can edit the existing one in-place if no-one else has a reference to it.

[untold sky]

No because the redirect goes directly into the Arma 3 project on FT

[gray wharf]

ye solved ^^

[untold sky]

There are two issues.
There is one where even right at mission start, the server is somehow stuck and every player gets kicked with BE timeout and cannot get in.
I investigated that in 2023, made a couple fixes but end result was seemingly not fixed.

The other is the one after kick. As far as I know that is on BattlEye's side. The game does not have a timeout, but BattlEye must have some internally which is why you can join after waiting.

We actually CANNOT tell BattlEye that a player has disconnected. I assume that is the problem. It doesn't handle a new player trying to connect on a IP where a player is already supposed to be on.
I cannot fix BattlEye

[untold sky]

Indeed..... What a curious little thing.
It replaces the "no value" with "nil value"

Don't think I'd agree on the term mangle. Both are a undefined variable.

The problem is during saving, it replaces the value. It should not do that.
But also fixing that doesn't make that much sense, because on loading it will place a nil value there anyway.
So after a save+load the result would be the same.

The bug is in createHashMapfromArray not allocating nil value instances.
But if I do that, that makes the use of hashmaps as a "set" container with no values, less efficient because it will allocate needless values.

So can we just consider that an intentional quirk?

Mh but it must be a bug?

_testHM = [[] select 0]; 
profileNamespace setVariable ["testNil", _testHM];
saveProfileNamespace;
profileNamespace getVariable "testNil";

Also creates a "no value" so should become the same, but doesn't.

Oh I see.
The normal serialization creates a temporary nothing instance.
Then serializes. And if the serialization result is nothing, it sets it back to no value.

The hashmap serialization is optimized to cut out one layer of it, exactly the layer that has that wrapper.

But won't that turn Nil into..
Aw gawd man

"any" == nil
"<null>" == nothing
"<null>" == no value

Cursed be ye

[untold sky]

I reported that a couple times in the past. There are different types of nothingness.
3 to be precise.
Typed nil ("Any" or other type), Nothing ("nothing"), no value at all ("<null>")

[untold sky]

It doesn't.
nil doesn't cause an issue. The problem is that there is not a nil here.

Perhaps because nil is still technically a value represented by the engine as a lack of one
It is not. The command nil returns a value. A nil of type Any

[untold sky]

Yes and I don't care

[untold sky]

The "objectarea" variable on the module does not have 2 elements.
It is set by engine. The code wasn't touched.
Do you have repro steps? just place the module? Or start the game with one, or start a MP game.
What is it?

If the object has a ContainsArea flag on it, it will be set at mission start when the entity is spawned.
I can't see why it would be broken

[untold sky]

Sounds good, please FT ticket with example mod.
Probably "skipLoadIfAddonPresent" something like that

[untold sky]

Yeah I had that on my mind too.
For RPT folder we do a deletion based on number of files and then oldest.
I guess we need same too.

Probably file size based though, instead of count.
Like.. 512MB folder size or so?

But I also don't like doing the deletion, because I like having a history and backup of all missions I've played in my folder. That's why I didn't do anything about it.

Maybe a combo of "if folder is too large, delete the oldest files that are larger than 50MB, but keep all the small files around"

FT ticket please

The hashmap mess is fixed

[ripe crystal]

I think your combo idea is great. I'd hate for people who commonly play other servers or missions other than ours to lose those just because we tend to overstuff lol. I'll create a FT ticket- thank you!!

[nimble veldt]

Could be some fluke, I don't see it anymore despite this module being used in several submissions. I've added more logging to at least know where did that happen and will come back with more info and/or repro steps if it ever happens again. Thanks.

[ripe crystal]

https://feedback.bistudio.com/T195101

[daring wagon]

Awesome if it happens

[clear glen]

Ah, I see... but no I was trying to prevent copying. A set of read-only non-copiable hashmaps to act as like an enum with tamper prevention. Makes it useful to equal by ref. It truly is just a frosting on the cake need. So if it is possible, I can find a use for it.

[vale maple]

I figured you guys couldn't edit Battleye but do you have control over when Battleye registers the player as part of the server join process? All we'd really need to solve it is to force the player to wait the 2 minutes before rejoining. Currently kickTimeout runs too late (the kick works but the player will already be broken in the BE player cache/map). I've also tried kicking on client state 1 but that seemed to be too late as well (unsurprisingly).

[gleaming plume]

Thank you for your time.

[onyx mica]

would it be big deal instead of kicking someone of server when he has soundfiles in documents\Arma 3\sound to only not loading those sounds and let him join?

[untold sky]

Crash at game exit on Stable version, long been fixed on profiling branch

[untold sky]

I have never heard of that command till today.
The code of it is weird, it says it would create a new dummy vehicle, but then never does.

Then its supposed to copy the turret data from that dummy vehicle.
But instead of the dummy vehicle, it passes null to the copy function.
And in the copy function, 90% of the code is just commented out.
It then just calls setVehicleAmmo.

I would say that command is just completely borked, and has been for a while. Don't expect a fix

[dusky tendon]

LoL... classic - "Don't expect a fix"

Maybe someone could at least add some text saying that the command is broken and shouldn't be used?
https://community.bistudio.com/wiki/setVehicleAmmoDef

I also wonder how long it will take to fix setUnitTrait ["loadCoef"], which breaks the stamina calculation logic when moving items in the inventory?

1. If the player is 90% loaded and has 10% stamina with the default loadCoef value.

2. If the player is 90% loaded and has 10% stamina with a loadCoef value of 0.

3. When adding or removing any item from the inventory, stamina is distributed correctly in any slot.

4. When adding or removing any item from the inventory, stamina is distributed relative to loadCoef 1 instead of loadCoef 0, which causes the player to instantly lose stamina if they had 100% stamina at that moment.

This is especially annoying when reloading.

In the end, someone forgot to add the values ​​of the getUnitTrait "loadCoef" parameter to the calculation.

[daring wagon]

I can do that if dedmen tells me what I should write.

[untold sky]

But it already contains the solution

If there is a timeout, its in battleye config. If there isn't then it won't be added

[untold sky]

And what would that accomplish?
So you get kicked by the game, instead of kicked by battleye timeout? Is that not the same end result?

[untold sky]

BattlEye and Signature verification are completely separate systems

[untold sky]

Yeah just needs someone to make a ticket now 😉

I will do feedback tracker tickets probably all day tomorrow, so if its there by then..

[untold sky]

Profiling branch on monday will have it

[daring wagon]

Thought there was one already. Will do

[untold sky]

The server.cfg eventhandler for player connection attempt, that you can use to force a player to wait.
That is the earliest you can get, don't know if that is before BE.

If I get right your proposal is to essentially, when a player disconnects, temp ban them for 90 seconds so they cannot reconnect, until BattlEye has timeouted?

Atleast I didn't see it in my last pass

[untold sky]

Probably not, but I need repro and FT ticket.

[untold sky]

From reading that text, I don't understand what it means.
So probably FT ticket with repro steps and observed vs expected behavior

[untold sky]

Warning! Does not function as expected in Arma 3

[daring wagon]

And to use setVehicleAmmo?

I understand they are pretty much the same?

https://feedback.bistudio.com/T195122 Ticket about skipLoadIfAddonPresent @regal nimbus

[untold sky]

If the vehicles ammo is full, they are the same.
Otherwise the Def one fully rearms every turret, and then runs setVehicleAmmo.
Maybe "veh setAmmo 1" would also do the full rearm?

[onyx mica]

ticket and repro here -> https://feedback.bistudio.com/T193138

[sleek scaffold]

Aren't completely borked stuff what needs fixing the most?

[limpid rune]

the few times I tried using it it seemed to work as intended? but I only tried it on like tanks etc

[vale maple]

Yes, exactly. Basically force them to wait the 90-120 seconds when they ungracefully disconnect so BE has the time it needs to clean them up before they rejoin. That way, they'll only get kicked the once instead of every 2 minutes until the server is restarted.

[dusky tendon]

Who know why currentweapon on player can be Put? How fix this?
(Moded server)
in logs error while he fire
Cannot find cartridge pos for Put, wrong weaponType

BRUH... ALAMUT = AK

[regal nimbus]

Put is the grenade throwing weapon.

[scarlet sage]

No, it's the put mine weapon. Throw is the throw grenade weapon.

[regal nimbus]

oh, yeah

[dusky tendon]

no know why but command broken with mod... how... how engine command can be broken with mods..

currentweapon result
primary = put
secondary = primary
handgun = put

How ....

[scarlet sage]

Even on yourself?
And are you on the profiling branch?

[dusky tendon]

bruh... on server(server call) he have primary Put but my unit have corrent gun

And are you on the profiling branch?
No

Even on yourself?
How you see on screenshot server and me see normal

[scarlet sage]

I wonder if he's normal on his side.

[dusky tendon]

[player,(currentWeapon player)]remoteExec ["groupchat",0];

global call... on he's side this normal

[scarlet sage]

Weird, weird mods

[dusky tendon]

weird weird

[scarlet sage]

That's it?

[dusky tendon]

current mods

[scarlet sage]

Actually, wait:
currentWeapon

Lou Montana - Posted on 2009-07-15 - 04:00 (CEST)
It seems it is not working in 1.02 for other unit than player.

[dusky tendon]

lol

2009...

[regal nimbus]

Could have just marked it as local argument and claimed it was intentional :P

[foggy shuttle]

I do know this is a common bug that happens. We haven't tracked down exactly what causes it, but the only way to "fix" it is to force the unit/player to update their weapon state. Firemode, muzzle, weapon etc.

Not just a mod thing, it does happen in vanilla with no mods as well

Basically the player can see his own weapon state properly, but over network it doesnt sync some times.
One way we verified a replication method was with exiting a parachute will always screw with the initial weapon state, and be incorrect for an unknown time frame.