#arma3_feedback_tracker
1 messages · Page 40 of 1
because sooner or later your script execution time becomes longer and longer
wiki says scripts get 3ms time to run before paused
startLoadingScreen
yep but you cant be in loading screen all the time
technically…
i’m not following, if you allow longer execution of a script then other scripts have to wait longer for their turn
yes and other scripts will also be executed longer
which of course might effect the FPS
but you get more scripts done
prioritising scripts may (or may not) be a good idea
increasing the duration spent on a script, not so much 
ah well on one script maybe, on all scripts no
More problems with vehicle randomisation i guess.
"O_LSV_02_unarmed_black_F" - Shows the turret even tho it has no turret (not usable)
https://gyazo.com/bfec7b73d3be7c6608f3a63432038835
https://gyazo.com/977c8919647c6abd822b9f5b423e8781
This made the turret disappear
cursorObject animate ["unarmed_main_turret_hide", 1];
https://gyazo.com/bc7fec622076e4dce8e3c00bf354b6a2
Same thing for the "B_CTRG_LSV_01_light_F"
See my previous message about the "same" thing
#arma3_feedback_tracker message
Can we get the old behaviour back ? Never had any problems with missing textures in years and now we have to deal with this 
is it worth reporting freezes/crashes from changing audio input device during A3 runtime?
There are some improvements inbound made by @solid marten . I guess you might be able to check it in some way next week
Let me get this straight, the problem is with disableRandomisation not working on animations? If this is the case then we have addressed this recently. FYI this was the case even pre 2.06
Texture overwriting has been also addressed. Try next dev or profiling with @untold sky ‘s mod
Can already do it now without dev/profiling right?
Just mod in #perf_prof_branch
I think so, but bad case might just be caused by windows and nothing we can do
Now that we have freeze dumps that also makes it easier to find the cause
there you go: https://feedback.bistudio.com/T161571
I have been unable to assign alternate textures to the Caesar plane and Taru helo via the virtual garage. It seems broken and either adds a random texture or just the default. I managed to add the appropriate texture to the caesar with setobjecttexture, but Im not sure how to do it with the Taru....
is this the same issue with skin randomization?
is it worth to mention that when you empty a group with joinSilent the empty group gets deleted?
that's… normal behaviour?
IDK if you deleteVehicle the group members then the empty group stays
that's to be reported
the group should be deleted when empty, unless "deleteWhenEmpty" parameter is set to false
right
and should never be deleted when deleteWhenEmpty is false?
This code instantly deletes the group: ```
_testGroup = createGroup [west,false];
_unit = _testGroup createUnit ["B_medic_F", markerpos "spa", [], 0, "NONE"];
[_unit] joinSilent (group player);
systemchat format["Group: %1 - %2", _testGroup, units _testGroup];
even when auto deletion is false
yep, there are some (many?) instances where this happens
"should", but IMO not a priority
yea I had to use workaround by adding logic to the group before joinSilent then deleting the logic
yes I know, ty
Well I created the ticket for this: https://feedback.bistudio.com/T161588
"Explosion" EH is not recognising the fact that there was an explosion, when the player who placed the explosive has disconnected.
Fixable or some smart way ? 🤔
can you construct a simple repro?
Would it be possible to extend parseSimpleArray to also parse hashmaps?
As, according to the Biki, it's "primarily intended for use with callExtension to parse the String output into Array" , this could be handy for extensions that would like to return hashmaps to SQF-land.
Like, something that could look like {key: value, key2: value2} or {[key, value], [key2, value2]} for example (or any other way to easily denote hashmaps)
toArray already does that?
But it doesn't do the other things that parseSimpleArray does 😛 (otherwise you wouldn't need parseSimpleArray at all)
On another note, it's too bad that when using str (the fastest way to serialize objects that I've found) a hashmap is serialized to just arrays of keys and values because you can't figure out if what has been passed - whether it was a hashmap or just a lot of arrays 😦
It would have been better to serialize to {...} or even hashmap(...), but I'm afraid that it's too late to change that
if you check the value of a hashmap ingame you actually can see that it returns an array like structure like so: [["key1", 0],["key2",1],...] so why not return an array to sqf and then
_hashmap = createHashMapFromArray parseSimpleArray _return;
``` ?
Because the arguments to parseSimpleArray can contain nested structures, one of which would be a hash array (and the others could be strings, numbers, booleans, etc...)
and my extension, for example, by definition, cannot limit the extension user in the types of arguments passed in and out (since it's a wrapper for python code and I'm letting the downstream mod user write their python code).
So unless a function like parseSimpleArray handles hashmaps, I have to tell my users "well, just handle this case yourselves"
And letting them just return python dicts that would then transparently be converted to hashmaps would be a nice QoL improvement 🙂
Not sure if making parseSimpleArray less simple is a good idea
Would be helpful though
tbh from my understanding it actually is, because the spirit of parseSimpleArray is really "call compile for extensions but without code execution as a side-effect" and if it truly is the case, then it's literally my use-case where I want to parse stuff effectively, for my extension, and if someone wants to create a hashmap then by all means let them do so - especially if it turns out other functions, like toArray, are already hashMap-aware, anyway.
Unless doing so will incur a huge speed penalty for the function, obviously
Cannot create objectwith type[AISubgroup], param[subgroup], NMT code[111]:
Cannot create objectwith type[AIUnit], param[unit], NMT code[116]:
is to have many of these "concerning"/sign of a bug or "normal"?
Forgot to post this the other day:
https://media.discordapp.net/attachments/897642382877687809/898049837583441991/unknown.png?width=748&height=517
EmptyPositions has "freePositions" in the in-game example instead of EmptyPositions
Simple typo error 😛
@untold sky I had a question about setPosXXX stuff.
You said that it has to update the quadtree or something right? But does it matter which lods the object has?
or if the object is simple or not?
But does it matter which lods the object has?
No, at most it will take the bounding sphere from geometry lod
what if it's 0?
(no GEOM LOD)
does it ignore it? or update it in one tree node only?
then it might use that bounding sphere.
I don't know if it uses bounding sphere at all.
ok 
I was trying to see if I can optimize one of my codes which does a few thousand setPosWorlds on a helper object... 
with attachTo the quadtree updating is the bottleneck.
With scripted stuff, the script itself might be so slow that it doesn't matter much
808x Client: Object(type Type_112) not found.
314x Client: Object(type Type_117) not found.
93x Client: Object(type Type_118) not found.334x Client: Object(type Type_112) not found.
200x Client: Object(type Type_117) not found.
129x Client: Object(type Type_118) not found.
are these problematic or no need to review the mission/feature code?
there are engine bugs related to this kinda stuff, hard to say if its caused by you or by us
we are working on a retreat-flee-surrender system. so it involves a fair amount of AI checks and behavior changes.
the code itself is probably not as strict as it should be with local checks
beyond that the test mission has nothing special - so ~20 small AI patrols, basic WPs/taskPatrol used
would it be possible to get sqf commands to list and remove guard points please
even the "guarded by" triggers dont list them. ref: #arma3_scripting message
if you could get their reference, you can update the position or remove them when not longer necessary - both would add a lot of flexibility and make guard WP even more powerful 🙏
FT request: https://feedback.bistudio.com/T161625
@solid marten / other devs. Hi, just wanted to ask if you saw my comment in the ticket https://feedback.bistudio.com/T161588 ?
done
replied to you
thx replied back 🙂
replied back
ok thx a lot just wanted to make sure you got my comment 🙂
the first time round
great, it just confused me how the commands dealt with empty groups (as they both can be used to empty a group)
Any idea why a ticket I made would display this screen?
hoping that someone didnt come in and yeet it for some reason
did you make a private ticket?
I don’t think so, I was able to access it yesterday. I also can no longer see it on my account
Spam:
https://feedback.bistudio.com/T161596
https://feedback.bistudio.com/T161613
https://feedback.bistudio.com/T161326
https://feedback.bistudio.com/T161183
oh and this one...about...bench presses??? https://feedback.bistudio.com/T161638
about the last one, some spam bots leave unsuspecting messages and then edit them later on (this is not one of them though, too obvious). I am also waiting for this message to be edited: https://forums.bohemia.net/forums/topic/236193-play-sounds-for-specific-players/?tab=comments#comment-3448165. marsk61 created the account and immediately made that single comment. but if he is not a bot then sorry and that comment brightened my day a little :)
Since most doors in arma 3 cause the player's models to get stuck in them for a few seconds when opening... could it be possible to have their collisions disabled during those "opening seconds"? You know to reduce player frustration... If so I can make a ticket for it
speaking of doors, does it make sense to have an event hander that triggers for object animations?
Is getLighting supposed to return the moon direction too? My return seems wrong if it is, when a night time
Or, maybe I'm using it wrong?
180+((getLighting#2#0) atan2 (getLighting#2#1))```Compare the return with the compass. In daytime seems fine. Nighttime, nope
it seems to me it only uses the sun position
I gave it a test, and at night, the vector points outwards from the terrain
which is where the sun would be... 
the lines you see are:
[eyePos player, eyePos player vectorAdd getLighting#2]
at night, the line is pointing towards the sky
(because the sun is on the other side of the earth...)
is multiline for drawIcon3D feasible?
the \n or br not working?
nope
did you rename them?
according to Dedmen 2 per unique freeze
but when that freeze happens in a code it just keeps spamming I think 
i mean the rpt got cleaned up but the frozen mdmp not
they may have their own count (vs rpt), but might be better to just delete them alongside rpt (once over the limit)
Well if its a human then I'd be surprised. According to google that guy registered accounts in dozens of different forums all on the same day and either posted no messages, or posted contextless garbage like he did on BIF.
And english coontextless garbage on german-only forums too 
That sounds like its gonna be a bye bye
I don't know if Arma deletes mdmp's at all, does it?
Could we get event to allow executing code on server when the mission has not yet started (lobby)? Or is that a bad idea? 🙂
we already have, check OnUserxxxxx
yes but dont those only work when the mission starts?
Did you read the documentation? Let me know if it is unclear
hmm where should I run the OnUserConnected EH?
seriously?
they are missioneventhandlers
so they only work after lobby?
no they work when events are triggered
ie some one connects to server BOOMSHAJALAK onuserconnected fires!!!!
oh yes I get what you mean now, I know they work like that
my confusion is where do you run the "addMissionEventHandler" command
on server
yes but at which file?
they are mission event handlers so they need mission running. you can let your dedicated server start mission before anyone joined the server and set all handlers then let people in
aah thats what I thought
which is why I hoped for event handlers that work even the mission isn't running
yes..
What he meant was having some way of server-client communication from lobby, without starting the mission.
yes ty veteran29
you can run code in lobby or server browser but you need mods
Server does not respond to client remoteExec at that stage
yeah I got the Client lobby modded but cant do anything about server
you can also run code on user connected to the server when no mission is running, but it will run in server namespace only, the commands you can execute are limited. the event handlers for those are in server.cfg file also named onuserxxxxx i believe
I tried that one already, wasn't able to call any functions or anything at all
diag_log should work in there and you can later access variables you store in server namespace from mission
but as I said command set is extremely limited and that is not going to change because a lot of game functionality is not available there
no servernamespace is server namespace
ooh I dont think wiki says that
but I guess that still dont help me to get the data from server to client in lobby
link plz?
Can I ask you, do you actually look up wiki sometimes?
lol of course I do
thx the reason I didnt find that is because its not listed in here: https://community.bistudio.com/wiki/missionNamespace
I would like to take your advice and try starting the server early. which switch allowed that again?
added it to see also
great ty
persistent and init something, check server params
maybe not persistent unless you want it
too bad there's a problem with that one. wiki says: Warning: This will break the Arma_3_Mission_Parameters function, so do not use it when you work with mission parameters, only default values are returned!
-autoinit
then don’t use it
when the world is not loaded the game has extremely limited functionality designed mostly to get people from the lobby into the mission, then when you have mission running you have more tools. You want custom lobby, make it on briefing screen like Apex did.
yea thx for the advice , will think about it
if I could even get a chat message back from the server I could use that 😄
thanks, dealt with
hey I don't know if this is a bug or not but I noticed that the client can't receive any remoteExec calls from the server while in briefing screen. The RE calls seem to be in queue and come through only when the mission is started. is this the intent?
Lots of networking code is limited by clientState being >= briefing, which is after role selection.
This very likely is one of them
ok, thx
so I had a weird request.
is it possible to have a command that allows us to draw 3d shapes in the game? 
e.g.: drawTriangle3D? 
go play a puzzle on the highway oh, interesting idea! 😁
I personally wanted it for debugging purposes... 😅
draw image something or draw polyline something no? because otherwise a triangle could be many things
I figured I should ask the simplest thing you can draw first! 😄
equilateral triangle? what orientation? etc 😛
I am a man of many questions, some do dub me as the PITA man
it's a 3D triangle 
given 3 coordinates, your triangle will only have 1 orientation
because 3 points always make a single unique surface (assuming they're not collinear of course)
oh, 3D in this way
IIRC Ded is planning it?
IYRC Ded is planning everything
void draw3DTriangle_method(vector a, vector b, vector c, Color color)
{
drawLine3D(a, b, color);
drawLine3D(b, c, color);
drawLine3D(a, c, color);
}
```😄
I wish I knew it sooner... 
I'm already half way done with my "3D Debugger" thingy...
I don't remember planning draw triangle, maybe i forgot
you also planned to giveMoreFPS and provide walking in moving vehicles, next patch plz kthxbai
I recall something about drawPolygon3d
in-game info says attachedTo returns array
but it returns OBJECT according to the wiki
@solid marten should I make a ticket or...?
btw while you're at it, could you also plz change the description for addForce to [forceVec, relPos] or something instead of [[x,y,z], [x,y,z]] ? 
(also it's x,z,y in the game...)
No.. 😅
oh jesus you almost gave me heart attack
A null object passed as a target to RemoteExec(Call) 'bis_fnc_objectvar'
//register everyone localy to player
[MODE_ADD] call bis_fnc_sharedObjectives;
//handle disconnect
if (isServer) then
{
addMissionEventHandler ["HandleDisconnect",
{
private["_player"];
_player = _this select 0;
_player setVariable["@","",true];
//run localy on server (as trigger won't fire on server)
[MODE_UPDATE,"",_player] call bis_fnc_sharedObjectives;
[[MODE_REMOVE,_player call bis_fnc_objectVar],"bis_fnc_sharedObjectives"] call bis_fnc_mp;
}];
};```
p:\a3\functions_f\tasks\fn_sharedobjectives.sqf
BIS_fnc_MP 👀
this is the only code context i can find using RE with bis_fnc_objectVar
maybe this needs some isNull check?
Attempt to override final function - bis_fnc_storeparamsvalues_data
p:\a3\functions_f\params\fn_storeparamsvalues.sqf
missionNamespace setVariable [
"BIS_fnc_storeParamsValues_data",
compileFinal ([_values, " param [", _names, " find toLower _this]"] joinString ""),
true
];```
this probably should have a check if already defined, no?
Didn't we have an isFinal command? 
Yeah we do
https://community.bistudio.com/wiki/isFinal
rev 148434
should forgetTarget lead to getAttackTarget returning objNull? (aka worth making a ticket?)
no idea what you just said
forgetTarget simply resets the target knowledge
it affects the entire group afaik
while getAttackTarget is per unit
he means should forgetTarget also set getAttackTarget to null?
getter for: allowCrewInImmobile
https://feedback.bistudio.com/T161783
BIKI on say3d
For some unknown reason if at the moment of command execution the player is in first person view and is inside a vehicle, the sound created is greatly attenuated.
is this still true?
you can pass 2 as the parameter for attenuation thing
you can already set that to true to get radio attenuation thing
IIRC, it was more specifically if you were the driver of a vehicle and that vehicle was the sound source for say3D. But a quick test suggests it has been fixed.
the stacked EH problem as discussed in #arma3_scripting :
https://feedback.bistudio.com/T161871
Dedmen, since you touched pip code not so long ago is there any chance to get this fixed? Or there is some kind of stoppers?
chance, ill put it on todo
Oh man, do you realise what can of worms you just opened? I've been at it all day already
The repro is just tip of the iceberg

btw there is a ticket somewhere on feedback tracker describing more bugs
I know it is somewhere there because I made it years ago
~~https://feedback.bistudio.com/T161693~~
last comment spam
yes, otherwise what would be the point? 
🤷♂️ you never know 😄
which section did you make it in and when?
Arma 3 - Category: Dedicated Server
Sun, Nov 7
sure, you will not like what you see 😅
does it happen if you dont use kickClientsOnSlowNetwork?
ill set to to logging for testing
kickClientsOnSlowNetwork[] = { 0, 0, 0, 1 };
I think it tested it a while ago and it didn't make any difference, but i will test again to be sure.
allright, thanks. i will reply again after a day or two 🙏
ok
Lou steps on a Mine
would it be a good idea to store all the script errors in an array or something and display them in a new window? (which user can open via button)
I made this arma .rpt parser once but that's really slow
can we maybe get a command like:
enable3rdperson false;
switch to "INTERNAL" view and disable "EXTERNAL" view
writing a loop for it everytime is really annoying.
you can disable the view change action with a displayAddEventHandler
there is a difficulty setting for this btw
a simple" first person allowed" on/off command would be handy tho, maybe you want to force first person in some cases and others not.
so one could toggle it on command without the need for loops or displayEH's 🤔
as far as I know this option is not possible for sp or mp missions, its only for server admins idk
exactly, this is why i asked
We are aware of that , it was kickClientsOnSlowNetwork last time and there was no work done on it
Then make your parser better and faster? 😄
or make arma not spam the .rpt with same error 😉
Got better things to do
or find error and fix it
I'm sure you guys are busy but would it be possible to have OnScriptError event handler so that we could make that error log window?
is possible
nice 🙂
should I make a ticket for that?
you could
eh, change seat on ViV cargo object? 😄
Yes 😛
Well 🤣
yes
cargo shotgunning wen? 
I'll do some other ViV fixes this week so thats perfect timing
Also, am I crazy or is it not possible to spawn Taru Bench Cargo without create3DENEntity?
https://feedback.bistudio.com/T161940
Ticket time
is there any chance these can be looked into? rather old bugs that have been around for some time:
https://feedback.bistudio.com/T83382 and by extension https://feedback.bistudio.com/T160377
it makes addUserActionEventHandler ["optics", "Activate", {}]; rather useless because it doesn't trigger using the default mouse controls, only when a keyboard bind is used
Many Taru pods are missing. There are quite a few tickets about that on the trackers.
Perhaps they are. Some were missing though at some point.
is the ghost hotel also missing? i cant find it in the editor 😩
Intended
I already looked into it and decided no fix
can i then also add furniture to it or is it just a simple object
once placed you can add anything to it.
any alternative coming to that in the future?
probably not
Also a suggestion about the skies, which can affect to the gameplay:
- the visible daytime moon
- colorized star map (the red-white-blue colors)
Both can be related to the navigation IMO
Double crash (two people crashed at the same time) just for you: https://feedback.bistudio.com/T161962
...now that the mdumps are created correctly 😏
shouldn’t this be in #perf_prof_branch ?
corrupted ogg file 
Doesn't seem corrupted, very weird very interesting
No, because the other friend was playing on the main branch, at least that's what he told me
hi, can we please get a new option for "BIS_fnc_dynamicText" which will set position/size of image and text at the same position/size for every interface size and resolution.
(a boolean : true will take user interface and resulotion in account)
something like:
[text, x, y, duration, fadeInTime, deltaY, rscLayer, userInterfaceResulotion] spawn BIS_fnc_dynamicText
i love BIS_fnc_dynamicText because it is so easy to use but only difficult part is to get the same results for everyone especially if you create sp/mp mission, i had to delete a few mission just because of this and creating dialogs and such is not easy for everyone but "BIS_fnc_dynamicText" is really easy to use and understand.
sorry for the long text.
you should make a ticket, not post it here 🙂
but I think it is unlikely to happen at this stage I am afraid
okay, i will make a ticket, thank you
dont say that xD
Actually I think Lou's right. You're better off just copying the function and modifying its code to achieve what you want
if i only knew.. but this wouldent be a definitive solution, rather a workaround
Dreaming of a new cool error dialog 🙂 https://feedback.bistudio.com/file/data/i2oh6kw5c7qizmyt2rqq/PHID-FILE-o2pyydaqigtogblmy3tp/errorDlg.png
I disabled all these and it still happens.
//maxping = 400;
//maxpacketloss = 30;
//maxdesync = 40;
//kickClientsOnSlowNetwork[] = { 0, 0, 0, 1 };
kickTimeout[] = { {0, 60},{1, 60},{2, 60},{3, 60} };
27 players out of 60 got kicked after around two hours mission runtime.
3.8k pending messages slightly before the kick
could you log jip queue next time it happens?
would it be possible to get an EH (per unit/group) for danger causes?
ref: https://community.bistudio.com/wiki/Arma_2:_FSM_Danger_Causes
like the engine already forwards the events to the danger FSM, so should be fairly easy to put an event trigger to this
one can do it via modifying the danger FSM, but this breaks compatibility with AI mods and you may not need it for all AI
Hi, this has been years and really affect gameplay, is there any intends to fix it? https://github.com/acemod/ACE3/issues/6819
not on plan
Lou Montana — Today at 20:09
it might be worth a feedback_tracker bump!
This old FT ticket contains a couple of suggestions for commands related to vehicle optics control that would be useful to add: https://feedback.bistudio.com/T85659
you little…! 😄
hmm https://community.bistudio.com/wiki/weaponState seems to report wrong firemode for vehicles 🤔
Get in a BTR Kamysh, run the code:
vehicle player addEventHandler ["Fired", {
params ["_vehicle", "", "_muzzle", "_mode"];
weaponState [_vehicle, _vehicle unitTurret gunner _vehicle, _muzzle] params ["", "", "_firemode"];
systemChat str [_mode, _firemode];
}];
switch to titan AT > topdown mode, fire, you get a message with ["TopDown", "Player"], fired EH correctly reports TopDown used, while weaponState does not, it seems like it always returns the first firemode of weapon's config
Related: https://feedback.bistudio.com/T162056
and https://feedback.bistudio.com/T162061 after I discovered enableDirectionStabilization doesn't do what I thought it did
Dynamic stabilizations with EHs would be nice to have. Optics thing sounds bad. If you change fov for optics with reticles you need to redraw them. Turn on/off zoom levels sounds not worth to implement. Maybe currentVisionMode and EH could return all modes and you can just use black screen on desire.
Is that by intention or bug that the music volume played via "PlayMusic" is affected by other sounds? (for example firing, rain, etc)
???
https://www.youtube.com/watch?v=0UDsNquqnnc
If I remember right, I reported this thing already last year when I noticed for the first time. Someone fixed it, and few months back I noticed that again.
What do you expect to happen? weaponstate returns state of either current weapon, in which case it will return firemode correctly or the given weapon, but it needs to know weapon muzzle and firemode in order to find the correct weapon. if firemode is not given it will find the first muzzle on the list. I am adding 2 extra params, though it wouldnt make any difference as the muzzle information is identical for both firemodes on the same muzzle. Maybe it will be useful for modding or the future changes
I'd say thats intentional, something like that doesn't happen by accident
I don't think so 😄
But it's most likely a not really fixable game issue. There is a defined volume limit that can't be exceeded in digital audio.
A3 is mixed way too loud, so when the music plays and shots are fired, the shots are running into what is called a limiter, which then tries to tone the overall volume down in order to not distort the mix
I like it 👀
there is a bug in some house doors in tanoa where the selection is invalid for the addaction command
the selection seems to cover only the door handle while all other doors cover the whole door
alright, thx for making it clear 👍
please make a ticket with house classnames (or getModelInfo), and screenshots
ok
screenshots optimally with arrows pointing at the problem
maybe I just zip mission with test script
Please still list the buildings, that makes it easier, no need to start game and find that info before being able to start fixing
I included the building model name: https://feedback.bistudio.com/T162067
is it a bug that in HitPart EH the projectile is sometimes null?
yes. Can detect and throw error @vocal abyss
Cyclic include-macros crash the game. Worth a ticket or is a note on the wiki enough?
script1.sqf:
#include "script2.sqf"
````script2.sqf`:
```sqf
#include "script1.sqf"
```Somewhere else:
```sqf
execVM "script1.sqf";
(you may consider these messages proof that
can see into the future)
allowFunctionsRecompile = 1;
```has to be added to previewed missions in the editor; otherwise one cannot recompile, is it wanted? the wiki says otherwise
https://community.bistudio.com/wiki/Description.ext#allowFunctionsRecompile
You can recompile in the editor but the button to do that is greyed out until this is added to mission config.
ah yep
I was indeed referring to that button 🙂
That could use a FT ticket, could be fixed now with is3DENPreview command.
just wanted to say, thank you all for all your effort and time you spent to improve arma ❤️
Devs, any comment on this? 👀
If I shoot missiles from pawnee to house the projectile is always null in HitPart EH
Don't know, needs investigation, thus ticket
ok here you go: https://feedback.bistudio.com/T162138
Could be because the part is taking damage from the explosion, not the projectile 
Maybe but could still use the projectile object
HitPart EH can tell the difference between direct and indirect damage. Using the test mission in the ticket, it appears to occur on some indirect hits, but not all. In all cases it can correctly identify the ammo type used (i.e. it's not going "well that's a generic explosion I guess").
yea but without _projectile variable its not possible to detect the hit position as the _position variable always points to each door.... (In case of splash damage)
I'm just providing more details about the problem.
glad to hear others think its a problem as well 😄
maybe it's a question if _projectile is invalid or is the _position bugged
https://feedback.bistudio.com/T161807
I collected some more logs and uploaded them to the feedback tracker ticket, tried different basic.cfg values and no difference.
20.11.2021 - 22:21:24 - JIP Que log included
Spam comment: https://feedback.bistudio.com/T126275#2259936
Each time an AI is killed I start to have Object WeaponHolderSimulated can't have zero or negative mass! a3\weapons_f\dummyweapon.p3d spamming
Is it normal?
do you create it yourself? wasn't the weapon holder that holds dead unit's weapons called "GroundWeaponHolder"? 
WeaponHolderSimulated sounds about right as dropped weapons have physx and stuff
this feature is from ACE?
WeaponHolderSimulated is a vanilla item (https://community.bistudio.com/wiki/Arma_3:_CfgVehicles_Equipment)
no
well I'm pretty sure I've never seen that error in the vanilla game when that happens
I just tried and I don't see the error
ground holders are doing weird things sometimes, maybe engine is not updating their mass properly for whatever reason.
Ever since BI added that functionality that "unstucks" underground items I've seen them jumping every then and now.
yes, you are right
well it's indeed called weaponHolderSimulated, so I was wrong about that one 
The engine create them apparently
I could not reproduce on a my local dedicated with CBA and ACE
Quickly tested with ACRE also on host but no message
May be it is due to the linux dedicated server I was using
could be
I was looking at the output of the
"#logEntities"server commands and i noticed this.
For me it looks like are are loads of possibly empty/inactive entities of these SupplyXX chillin at [0,0,0]
My guess it could be empty "GroundWeaponHolder" / "WeaponHolderSimulated" / "LootWeaponHolder" trashing the server instead of getting deleted ? (We spawn loot in buildings etc).8k Supply40 entities after 2h 50m server runtime.
17k entities at pos[0.000000][0.000000][0.000000]*if you can trust the position output of the function.
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:68:1305, Pos: [0.000000][0.000000][0.000000], N:Supply30
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:68:1306, Pos: [0.000000][0.000000][0.000000], N:Supply30
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:68:1307, Pos: [0.000000][0.000000][0.000000], N:Supply40
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:68:1308, Pos: [0.000000][0.000000][0.000000], N:Supply40
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:68:1309, Pos: [0.000000][0.000000][0.000000], N:Supply40
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:68:1310, Pos: [0.000000][0.000000][0.000000], N:Supply40
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:68:1311, Pos: [0.000000][0.000000][0.000000], N:Supply40
I also noticed this increase with longer runtime.
https://gyazo.com/7707b7a5c97fb18d4e4c129474faa27e (~2h 55m server runtime left / ~70 minutes on the right)
https://gyazo.com/90ff46c49579ca74e3cd4b1bac403ffd ("Vehicles" after 2h 50m server runtime)
https://gyazo.com/3685326e16cde915bbbe6b613feaf2c6 ("OUT VEHICLES" after 2h 50m server runtime ) *whatever that means
Any thoughts?
These are most likely backpack containers, backpack containers being left behind and calling "object not found" kinda spam in logs is a long time issue
I don't know if uniform/vest does something similar. Supply30/40 is a bit small for a backpack
not ground holders, ground holders have gound holder class
id numbers are sequential it means some script must have been running spawning those at high rate
68 is some client spawned them and when left they got transferred to the server and left as junk
wait you run loot spawner on client?
another possibility is that position wanted is null because object is null like building that hasn’t been streamed yet so position will be [o,o,o]. long story short, can you choke server with badly written script? ABSOLUTELY
loot is spawned by the server via containers and groundholders.
i can get some more logs and ask other admins if they have something similar, if that's useful 🤷♂️
Edit: I think i tracked it down to players creating it, so maybe vest/uniform/backpack like you said
Some players create between 1k-3k of these in the 3hr "mission runtime", if I'm not mistaken these are clients with bad internet (have to verify).
A LOT of modification to the Eden editor is possible using SQF currently, but it seems to be missing an editor side "pre-previewMission" that would fire before do3DENAction "MissionPreview" actually exits the editor.
Would such an event be possible and with the given lower popularity of editor customization, be likely to be added?
also sorry if this is better suited on the actual feedback tracker, first time for me.
Just use a preInit function? 
Should be doable, can you give me a usecase?
Yes please
Give usecase!
(And make ticket
)
And name.
prePreView? OnBeforePreview/BeforePreview?
I hate prePreview 😮
preInit of an eden previewmission runs after the eden editor has closed. This event should runs before that and still give you acces to the eden editor commands.
I currently have scripted customConnections in the editor. With a saving functionality that runs onMissionSave. But if a player goes into preview there isn`t anyway to store the connections before the editor closes. In general, with OnMissionSaved and OnEditorPreview you could script water tight saving functionallity for any custom editor features/data.
The event that fires when the previewMission is loaded is OnMissionPreview. OnEditorPreview would seems like the logical name for an editor side event.
that sounds cool 
the customConnections or the eventname?
I`ll create the tracker asap btw
Buuut?
Engine support would still be cool, even if it invalidates the work I did :)
2 Hour runtime and i got almost 5k of Empty Supply20,Supply30,Supply40,Supply60,Supply80 etc, just from one client (The particular client was only online for 1 hour tho).
https://gyazo.com/986d6b0c0ab82a587256e89eb33bf0a7
How bad is this for the server performance ?
And if it's bad for performance, is there any way to mass detect & delete them ?
I tried _supply40 = entities "Supply40"; on the server but it doesn't return any
Have you tried allMissionObjects too?
Yes
I asked one player what he was doing while he caused that many (2-3k) of these and he said he was just looting, transferring loot to vehicles etc.
So taking vests,uniforms,backpacks from dead AI/vehicles/containers and "transferring" them around with the intention to sell them at the trader and to keep some of them.
Some game mode specific information might help to track it down 🤷♂️ .
of course it is bad, they are global objects, they need to be synced among all clients + jip. The network is just get clogged so server cannot serve all customers so it kicks them out
I am still waiting for that 😛
same
If we can get reproduction steps to cause this (and script to test that it's going on), ideally with a single backpack already reproable.
We'll look at it
Sounds good.
I think it should be reproducible in a heavier mission with many players (~30-40+) and some more activity than usual involving gear changes.
(vests/uniforms/backpacks from other players, containers, vehicles or dead ai/players etc)
Maybe someone with less mods but similar activities ingame could verify if they can observe the same behaviour.
Ill ask some other Exile server admins if they have the same behaviour (i guess so because of the object not found spam you mentioned that might be related to that).
"Heavy mission with 30-40 players"
Is not something we can do.
And even if we could, there would be so much other stuff going on that we would have a really hard time tracing out the problem
[2021-12-04 03:31:23] [info] AEOG-ServerMetrics: Build:148581 | Runtime:30min | Players:16 | AI:98 | allMissionObjects:6978 | AllVehicles:456 | ObjSimulated:1299 | AllAliveEnt:1261 | activeSQFScripts:32 | FPS:101 | Last10mAvgFPS: 92.2)
[2021-12-04 04:20:26] [info] AEOG-ServerMetrics: Build:148581 | Runtime:79min | Players:17 | AI:117 | allMissionObjects:8043 | AllVehicles:497 | ObjSimulated:2240 | AllAliveEnt:1314 | activeSQFScripts:70 | FPS:86 | Last10mAvgFPS: 74)
[2021-12-04 05:20:00] [info] AEOG-ServerMetrics: Build:148581 | Runtime:139min | Players:18 | AI:104 | allMissionObjects:7064 | AllVehicles:463 | ObjSimulated:1158 | AllAliveEnt:1290 | activeSQFScripts:38 | FPS:52 | Last10mAvgFPS: 58.6)
[2021-12-04 05:52:09] [info] AEOG-ServerMetrics: Build:148581 | Runtime:171min | Players:16 | AI:91 | allMissionObjects:7147 | AllVehicles:466 | ObjSimulated:1305 | AllAliveEnt:1302 | activeSQFScripts:46 | FPS:73 | Last10mAvgFPS: 65)
#logEntities after 2h 52m runtime and i found 6.400 "empty supplyXX" just from two clients 
Maybe less players needed...
Yea i thought so and the mods. I get some more info and maybe even try to replicate in a "vanilla as possible" environment.
And now i stop clogging up this channel, so thanks for the response so far 👍
I would love to fix it. I just can't if I don't know why it happens.
And I cannot find out why it happens, if I don't know how to make it happen
woah, they are working overtime now before xmas, even on days off
also spam https://feedback.bistudio.com/T162215
https://community.bistudio.com/wiki/weaponState seems to return the correct fire mode if the first syntax is used on a unit, but not for syntax 3?
Test code when any weapon is set to full auto:
weaponState player select 2 // "FullAuto"
(player weaponState primaryWeapon player) select 2 // "Single"
(player weaponState currentMuzzle player) select 2 // "Single"
thanks, cleared
too 👍
muzzle with single mode has the same name as muzzle with full auto, it is impossible to distinguish which one you are looking for by the muzzle name alone, so the result is the 1st matched muzzle. I can make it compare name with current weapon and return current weapon if matches
if you could make a ticket so I wont forget as this is kinda low prio
dealt w/, thx
no, no more spam!
argh
j/k, let's nuke it 😉
its like 3 posts
i wonder if anyone ever just a link posted like that to buy or use a service XD
it improves their Google visibility the more websites aim at theirs
they don't (really) expect it to be a working advertisement, more like another ladder step towards Google's first page
Hm random idea.
Both AI and players sticking their gun through a wall/objects in order to shoot at objects behind the wall, wall glitching.
We could just when you fire detect if there is a wall inside the gun, and if yes, have the bullet directly impact the wall instead of just spawning behind it 
Thoughts?
And do you think that's worth the effort?
Sounds good but wouldn't it be performance intensive?
performance-wise, maybe just the players?
if you do a trace or something alike that is expensive
Totally worth it for Ai... but for players the wall is already blocking their view... I see more complains in COOP missions about the Ai being able to shoot through walls contributing to the Arma famous "where was I shot from?" Now again wouldn't a check on the way the Ai gets its targets do the same thing without having to do a check on each weapon fire? You can add a visibility line intersect to the system in which the Ai gets its targets and if there is a wall blocking the Ai from its potential target then it is not considered a "viable target" and so the Ai would not even fire... whatever solution is less expensive...
If I got a dollar for each time I got players complaining about Ai shooting through those HQ Cargo Houses' walls I would be rich already...
Yeah, the cargo houses and similar buildings with thin walls are the worst ;D
We could just when you fire detect if there is a wall inside the gun, and if yes, have the bullet directly impact the wall instead of just spawning behind it :bird_woah: it would need special routine for bushes, since player can traverse through their fire geometry freely
in any case, I guess something like that would also fix issue when players are crawling under the vehicle to kill their crew
the bullet would cross anyway so it should be ok?
though it might change the ballistics a very teeny tiny little bit, with the additional distance to travel
we can say it's the branches to hit, that's why 😀
it might be bit more tricky with HE ammo though 😄
naaah, it's o— boom
Was the pilotcamera class brought in with jets dlc
a bit earlier;
Jets is 1.68 and commands are 1.64
Is there a counter part to pilotcamera that can be used in turrets
I don't know, but why, would you make a ticket for such feature?
I don't think it can make it that late in dev
A bullet does a trace every frame.
Doing one when you shoot basically doesn't cost anything
I would just spawn the bullet a bit further back in the barrel, instead of at barrel tip.
But needs to ignore weapon collision then
That was what sparked my idea yeah :D
I’m not looking for it to be rushed in , especially if it’s to late for this dev version
no I meant it as "most likely too late ever"
it would imply new commands (versions at least) to use turret path, it would only be useful to modders, so maybe it's not a big chance
again, I don't take decisions in this matter so you may still take a shot! 🙂
Well I hope it can get added Evan without the scripted commands, having the turret lock and stab would be huge for turrets
you can do scripted turret lock and stab with directionStabilized
You can but it’s not the same, the stab dosent stabilise aircraft movement, and turret point lock is messy to do good
point lock is messy to do good huh? I did something like that in RHS at it is working quite well
What command did you use?
in any case, I was thinking to add ability to toggle on/off directionStabilized since UAVs already have that option
I think iv tried a basic version with that before. Getting the improved camera stabiliser from pilotcamera is the bigger thing of my thread
So i collect some more info and i think the "Out vehicles" entitie issue might cause the mass kick problem we currently have.
"Vehicle Entities" (Left low pop server period, right high pop primetime with 50+ players)
https://gyazo.com/c76a0fc29ec602d7b205941041de8312
#logEntities "Out vehicles" after the mass kick and 2h 52m runtime
IsLocal: 22200
IsMarkedToDelete: 1
IsDestroyed: 1
IsDamageDestroyed: 1
IsDamageDead: 1
IsSimulated: 23362
IsInvisible: 0
IsAnyPlayer: 9
IsOutsideMap: 0
12k "Supply40" elements alone.
I connected ~15k "OUT VEHICLE" SupplyXXX entities via netID & ClientOwner just to four users 
Server Performance Metrics log (Mass kick was at 20:47:33)
https://gyazo.com/bd360734ce23b6b6b8e15acd253d43cb
Assumptions don't help me much.
We need to know how to repro these stray entities
I know. Iam thinking isn't there any way to delete them? They are detected as "empty', so one could detect them somehow and delete them without causing any more severe problems than we already have (degrading perf and the mass kick due server overload) ?!
I tried to delete one via and it seemed to work.
_obj = objectfromNetID "X:XXX";
deleteVehicle _obj;
I try to pinpoint the cause over the next week
try to add some logging when these are created (by players and/or server, their position, etc) and try to trace to specific events. finally try to reproduce it in a reliable way
Maybe i found the reason, i just had a player creating a low amount (30 stray entities) in the time-span of 11 minutes.
I asked him / and checked what he was doing, he was just driving around in a vehicle and he switched uniform/vests.
Looks like it's happening for players in vehicles 🤔 I get some more data and reply back if iam 99 % sure
player numbers are super lower (2-8) and server perf shouldn't be a factor since it's running between 80 - 90 (FPS Limit = 90)
https://gyazo.com/fd8def081d96ed713e188d95f5ff3204
there is nothing we have on the client side that should create this (as far as i know)
only some clients causing them and others not ^^
Just had another player with a seemingly bad connection (got kicked for signature missmatch two times) causing 700 stray entities in 35 minutes while the server is running at same low player numbers and high FPS 
not speaking of triggered by scripts necessarily. can be from engine behaviour too. say item interaction, putting items into vehicles or out
what Dedmen is saying they need a specific reproduction by one person best case. so everything you can do to narrow it down to what actually triggers it, will help
sorry Lou, spam https://feedback.bistudio.com/T162300
don't apologise, that's on them 😉
dealt with! 👍
Sorry spam, Lou
Exotic Spam https://feedback.bistudio.com/T162309
gone
KH-3A Fenghuang (Pacific CSAT drone) does not have configurable pylons
https://feedback.bistudio.com/T162334
I only discovered this recently when trying to implement this asset for a mission. I'm guessing it was accidently overlooked when the devs were creating new "dynamic loadout" versions of the existing aircraft at the time of Jets DLC development.
@shrewd meadow - it was actually decided it's not worth upgrading it to dynamic loadout since it's not supposed to be able to mount any other missiles. It could be potentially still useful for community but adding new variant is bit of hassle, since it has to be a new class for compatibility reasons.
Hi everyone.
How does one add screenshots and an error report to a feedback report? Can't believe I'm asking for support for making a support ticket 🙄
make a ticket 🤣
I believe a drag & drop is enough though 🙂
Yes I did that
This seems to have worked. Thanks Lou. Oy vey an extra line of text from the designer of this tracker would have been appreciated.
Can someone explain to me what Wulf means by "exception" in this ticket? Is he saying it's a deliberate design choice? Because otherwise I'd assume he's saying it's unfixable, which makes absolutely no sense to me.
https://feedback.bistudio.com/T139722
It's almost as if BI are deadset on denying milsim players the ability to script in survivable aircraft crashes, but I don't see why that would be the case.
Well he says
Closing as this most likely won't be changed.
I don't understand why, though? Is it a deliberate design choice?
Sounds like, just due to how that stuff was implemented it's not viable to fix it.
Not impossible but also simply won't be done
Hmm, now I wish I'd replied to the ticket sooner so Wulf would be able to answer me. I'm guessing he probably doesn't remember anymore.
Handlehit maybe gets better results?
Yeah, eventHandler, my mistake. I haven't touched the script in over a year now. I'll have to look into everything again, I sort of feel like the handleHit EH might actually provide more information that I need for the script as well.
Is this the one you're talking about?
https://community.bistudio.com/wiki/Arma_3:_Event_Handlers#Hit
Yes
Alright, thanks.
why does setVelocityTransformation use ASL? 
it should use world...
I'm simply rotating the object using setVelocityTransformation
now imagine moving an object in a straight line 
start = getPosASL player vectorAdd [0,0,10];
end = start vectorAdd [0,1000,0];
spd = 5;
duration = 1000 / spd;
heli allowDamage false;
timer = 0;
onEachFrame {
if (isGamePaused) exitWith {};
if (timer > duration) exitWith {onEachFrame ""};
timer = timer + diag_deltaTime * accTime;
_dir = [0,1,0];
_up = [sin (500 * timer), 0, cos ( 500 * timer)];
heli setVelocityTransformation [start, end, [0,spd,0], [0,spd,0], _dir, _dir, _up, _up, timer / duration];
}
I don't know about you guys, but that's not a straight line
(I mean the movement of the object center)
this is because getPosASL uses the land contact of the object, which changes according to orientation, so technically it's still a relative position, not absolute, which nullifies the point of using ASL format
would it be possible to add an additional optional bool parameter to make it use world instead?
made a ticket:
https://feedback.bistudio.com/T162411
I've never noticed. That makes modelling aircraft trajectories even worse 
Spam comment https://feedback.bistudio.com/T162294#2264956
gone, thx
This got me thinking: is there any chance to implement a workaround that would prevent AIs disembarking into rocks before you implement this fix? 😔
The fix is obviously needed, but that was also my way of dealing with these AIs, who otherwise would remain in the rock forever
Although I also have no idea what to do to prevent that, that would not possibly break the game in another way
disembark oof. What do they do instead, randomly teleport to nearest available position, or just not get out? 
both are kinda crap and have the chance to end badly
Maybe a semi-decent solution would be to cycle over all the exits a vehicle has and try all of them (following the human logic of "driver's door is blocked, let's try to exit using the passenger door") and if all fail to disembark correctly, then just use the original exit.
How about that?
This would also prevent some cheats in the form of driving over to a base, parking just next to the base's wall and then exiting "into" the base, through the wall.
Sounds nice.
But i don't wanna do that effort 
How about some plushies sent to your company address, eh❔
That's bribery!
I think I'm not allowed to accept that 
Lies!
What Lou said!
Also, I have completed an anti-bribery training at my company, lately, so I would know if it was bribery!
Is there some way to access old Feedback Tracker tickets? http://feedback.arma3.com/view.php?id=12340 is mentioned in ACE's codebase and I'm getting 404'd.
Ah, wayback machine has it.
not anymore 😎
Gottem
, that spam was the closest to a actual message ive seen so far.. they are learning 
What they do now they just copy sentences from description or other comments so it looks legit and cover a link in punctuation so you wont see it clearly.
Thats what they've been doing on the forums for years
Spam comment https://feedback.bistudio.com/T162484#2265569
removeded
any one know how to make a ticket on here cause im in need of help and no one will
people cannot help you if you don't ask 🙃
Top right, the star, new Bug (off the top of my head)
Hey man thanks for the reply and yeah I've been asking since yesterday for help in the troubleshooting and ask a question here and just as normal it got pushed to the side with some one eals problem so I thought might as well open a ticket to see if anyone on there more help than half the guys in the chat because this problem a new problem to me aint seen it before
what's the problem?
Heya sorry just seen the ping im sadly on my phone so I dont have the photos but if you go to troubleshooting and scroll up a little bit you'll see the problem im running in to
Found another weird thing with the help of #logentities
Looks like the "Can't change owner from 0 to 2" error happens if a client accesses the inventory of a container spawned by the server and then something is trying to change the ownership of the supplyXXX entities inside (backpack,vest,uniform).
"AeoG_B_CargoNet_01_ammo_F" is a crate we only use for a "heli loot drop & crash script".
3:34:51 Server: Object info 11:11 not found.
3:34:51 Can't change owner from 0 to 2
3:34:51 Server: Object info 11:12 not found.
3:34:51 Can't change owner from 0 to 2
3:34:51 Server: Object info 11:13 not found.
> Loc: 1, ForDel: 1, Dstr: 0, DmgDstr: 0, DmgDead: 0, Sim: 1, Inv: 0, AnyPl: 0, Out: 0, NetID: 11:11, Pos: [0.000000][0.000000][0.000000], N: Supply120, SupplyInfo: (W: 0, M: 0, I: 0, B: 0), OwnerN: AeoG_B_CargoNet_01_ammo_F, OwnerSupplyInfo: (W: 0, M: 0, I: 4, B: 2)
> Loc: 1, ForDel: 1, Dstr: 0, DmgDstr: 0, DmgDead: 0, Sim: 1, Inv: 0, AnyPl: 0, Out: 0, NetID: 11:12, Pos: [0.000000][0.000000][0.000000], N: Supply140, SupplyInfo: (W: 0, M: 0, I: 0, B: 0), OwnerN: AeoG_B_CargoNet_01_ammo_F, OwnerSupplyInfo: (W: 0, M: 0, I: 4, B: 2)
> Loc: 1, ForDel: 1, Dstr: 0, DmgDstr: 0, DmgDead: 0, Sim: 1, Inv: 0, AnyPl: 0, Out: 0, NetID: 11:13, Pos: [0.000000][0.000000][0.000000], N: Supply40, SupplyInfo: (W: 0, M: 0, I: 0, B: 0), OwnerN: AeoG_B_CargoNet_01_ammo_F, OwnerSupplyInfo: (W: 0, M: 0, I: 4, B: 2)
Could be that the missing uniform/vest bug (naked bug) is also related to this somehow ?! 🤔
Can you get me a repro?
The naked bug i don't think so, but Arma works in mysterious ways
Ill try, first i have to fix this damn mass kick issue. Iam at a point where i simply wanna quit this game because of all its bugs and problems.
I thought it's the stray entities, but i run scripts to delete them and instead of 4-5mb logentities logs its not bigger than 2,8mb (worst case) and only a 1-3k stray entities instead of up to 20k.
Server is running fine and out of the blue you see this and mass kick, and there is no indicator of whats going wrong 
https://pastebin.com/UsuCrV9A
Traffic drops to zero and after everyone is getting kicked the server runs fine again
Arma Build: 148592
22:33:44 Server load: FPS 28, memory used: 3792 MB, out: 0 Kbps, in: 49 Kbps, NG:0, G:5832, BE-NG:0, BE-G:0, Players: 54 (L:0, R:0, B:0, G:54, D:0)
fps also drops to zero? or does it keep ticking?
vTune trace of that happening would definitely be useful
but I imagine thats pretty hard to capture
Fps drops a bit but not to zero.
I have one vtune profiler run where it happened
i thought we established jip queue is too large so server purges everyone and clears it???
that's new to me 😆
gotta compare the size before and after kick
that was my next plan to get some logs about that and captureframe
if that's the case i can remove loads of containers (persistent game mode, so people accumulate loot),
to get the size of the jip queue down ?! Or is that not something in the jip queue ?
Edit: Indeed the case, deleted container table in our DB and compared the Jip log size (11mb vs 3mb) 
-> https://gyazo.com/a65249f84e5f8bf368ddf9b8fca70ae1
there is a way to have unlimited amount of stash on the server with 0 impact on performance
also hacker free
no one would be able to get to your stash unless you told them about it, so theoretically you could come back in a year and get your stuff
Iam gonna reduce the amount and add a container limit for each base and see if it changes anything.
It makes sense for me at the moment, because last time it stopped happening randomly and didn't happen for 7 month's.
probably because less hoarders / less bases / less containers were active 🤷♂️
That's a different issue
This thing happens mid game, not at JIP
is ctrlSetAngle warping the RscPicture a known issue when the client is using a custom FOV?
been looking into an annoying glitch for the better part of a year and found it to warp the texture when using a FOV other than the default
ah yeah there is a ticket https://feedback.bistudio.com/T136844
Not sure if it's due my "Entities Garbage cleaner scripts" or other bad code.
Just had 5k #SoundOnVehicle entities accumulating on the server
55+ players & server FPS around 10
Deleted them via
{ deleteVehicle _x; } forEach (allMissionObjects "#soundonvehicle");```
and boom server runs with 25 FPS after 160m mission runtime and same amount of players 😆
Gonna debug that a bit more (with stable, without the frequent use of say3D, without "Entities Garbage cleaner scripts") and will report back.
Server Build:148592
Perf Metrics: https://gyazo.com/ae14469b4e6b3af5aa587345cb42c252
VTune Profiler: https://gyazo.com/dcea0f352308d84d341e2ef87f49ab4d
I need the vtune capture files, not a screenshot of it.
It will be hard to find out where these sound entities come from.
I could add debug logging for you to dump script calltrace when they are created by script
I'll have time to look at all your stuff in two weeks
Have you created a ticket on this?
I know dedmen, i just included the screenshot for a quick look.
I have the VTune run for this period (#soundOnVehicle) if you need it.
Iam watching if it happens again, i think its not that common. (First time i saw it like that)
Maybe it's just the same useless script kiddy sabotaging our server 🤷♂️
No, iam not too sure if it's that common and FBT worthy ^^
Just wanted to raise awareness so maybe other admins/server can confirm/verify it.
Would be a thing to check for other big game modes when they have a sudden server FPS drop.
Log this on the server side
diag_log format ["#SoundOnVehicle Counter: %1", count allMissionObjects "#soundonvehicle"];
Is there any technical reason the AI can't walk with lowered pistols?
In PistolLowStandActions it looks like it was intentionaly disabled for AI or somebody forgot to change non-player actions.
the AI variants are not overwritten and are using parent class raised weapon variants:
also PistolStandActions >> stopRelaxed animation is not a one with transition to PistolLowStandActions like it is a case for rifles. So AI will not even try to use the lowered weapon state due to that.
valve @untold sky pls fix
https://feedback.bistudio.com/T134374
Happened again with the #SounOnVehicle entities
Looks like there might be also a problem.
Got 8k of these and server FPS went down to 8 😵
Build:148592
#logEntities: https://gyazo.com/7999a32ece0df37aa99268ad8f8fa53c
Perf Metrics: https://gyazo.com/1b93d118f89690b2b35c5971038624d1
*last line in the perf metrics you see that after everyone got kicked before server restart all the #sounOnVehicles entities are gone and server runs normal again.
[2022-01-05 17:59:18] [info] AEOG-ServerMetrics: Build:148592 | Runtime:178min | Players:0 | AI:0 | allMissionObjects:7669 | AllVehicles:298 | ObjSimulated:1502 | AllAliveEnt:1123 | soundonvehicle:0 | activeSQFScripts:16 | FPS:87 | Last10mAvgFPS: 22.4)
Same day with much more players, it looks like it's working like intended. 
https://gyazo.com/94e93d73a668573ebfd620a7aa16e1ac
Gonna switch to stable build next restart and check there.
=================== Out vehicles ====================
.....
.....
Total objects: 60828
Statistics objects: 60828
IsLocal: 3215
IsMarkedToDelete: 0
IsDestroyed: 0
IsDamageDestroyed: 0
IsDamageDead: 0
IsSimulated: 60828
IsInvisible: 0
IsAnyPlayer: 8
IsOutsideMap: 0
30.740 supplyXXX entities from one user caused in 70 minutes playtime
https://gyazo.com/4cb3fda661b6d689da71db518fd34bec
25.079 supplyXXX entities from a second client caused in 70 minutes
https://gyazo.com/1d2fa198d492c4a3def3ea7ca4c1ecca
Perf Metrics - https://gyazo.com/3616e7db45758327cade6d85dabe1152
- starting where this two clients joined the server.
Arma Server Build - Stable:148470
Both connected/disconnected at the same time, both are from the same country and use the same internet provider (Whois IP data) 🤔
Add +10k for each client like the netID indicates, since i delete stray entities every 10 minutes but only up to netID "OwnerID:10000".
Surprisingly server didn't crash nor was there a mass kick.
After a server restart they caused it again, first one caused 20.000 in 120 minutes and the other guy 10.000+.
What is this ??? 
People use cheats to spawn items?
What?
As much as I understand message entities are from specific users.
No, no they don't
spam comment, kill pls
https://feedback.bistudio.com/T162775
thx - removed
nukeded, thanks
I don't have permissions for these, so it's cool 😄
looks like the new "optimized" expression for configClasses and configProperties doesn't work if you simply invert the condition 
e.g.:
configProperties [configFile >> "CfgVehicles", "isClass _x", false] -> ~2 ms
configProperties [configFile >> "CfgVehicles", "!isClass _x", false] -> ~8 ms
yep, because it's text-based I think
as in "true", "isClass _x", etc
yeah it is. I was hoping the inverse variants could be added as well (or at least just "!isClass")
so Y U NO HAPPY 😄
also another question. does the game store the count of classes and properties separately?
if so I would really like a command for that (e.g. instead of count _config we could do _config count "classes"/"properties"/"all
)
@solid marten btw I get an error in the map related to getElevationOffset the first time I open the map. it was something like this:
Error missing ]
is it a known issue?
oh nvm it was a typo. but anyway needs to be fixed. this was the error:
_mapPos + getElevationOffset);
];
Error missing ]
It is known, the relevant code has been reverted since last weeks Dev Branch update. The reverted version should hit the Dev Branch in this weeks update.
Spam: https://feedback.bistudio.com/T162857
https://feedback.bistudio.com/T162850 (2 comments, OP is not spam)
https://feedback.bistudio.com/T162843
dealt with, thanks
nukeded
Semi-cleverly-disguised spam: https://feedback.bistudio.com/T162869
https://feedback.bistudio.com/w/ft_a3_howto/gamecrash/ spam in comments 
shhhhhh don't worry, it is gone now. it cannot harm you ❤️
gone, thanks
any thoughts on this?
https://feedback.bistudio.com/T162811
thanks, removed
seriously tho. it's a lot better than the current nearestObjects, where you have to pick a large radius to find objects 
and the game already has that info at hand anyway, so shouldn't be hard to implement (in fact, I'm sure it's 99% the same as the current nearestObjects; only the criterion has to change)
Sounds useful.
Thanks for reporting this, im looking into getting all that spam removed. Atleast i can remove anyone who made the Dashboards👍
cleaned, thanks
@solid marten btw remember when I asked for the roadway lod support for lineIntersectsSurfaces? and you said engine doesn't support it?
https://feedback.bistudio.com/T150513
doesn't 3den already use the roadway LOD for placing objects on top of one another?
and so does the getPos command
so I was hoping if you could check again. roadway lod is extremely useful. and constantly doing getPos (which needs a setPosXXX command + helper object) is so wasteful
@gray wharf spam:
https://feedback.bistudio.com/T128842
all removed (but plz no ping)
sorry I'd pinged KK before so I sorta wanted to separate the messages 
just post the links, I frequently track this channel and deal with spam 🙂
You imply that other uses/commands you mentioned all use the same method as lineIntersectsSurfaces, but this is not like that
To me the changes required for it look rather simple.
But I don't know if I'm missing something.
Our intersection method in engine doesn't support roadway LOD as KK said, but we can add that
I'll look at it later, but might still be that there's more to it and its not viable. But maybe its as simple as it looks
thanks. btw could you look at this too? not saying accept it or anything, but just wondering if it's actually possible as I think
https://feedback.bistudio.com/T162811
the name nearest implies that they are sorted by nearest. So adding a "don't sort" parameter doesn't make much sense for me.
Also sorting is pretty cheap I think so doesn't matter much if you want it for performance
"large radius can be slow" since.. eh.. since some time ago that command is multithreaded, so it scales much better with large areas now. So is that really still required?
_obj distance _pos < _radius + _obj.BoundingRadius;
not possible.
It works by finding the fields in the QuadTree that are inside radius, and iterating through all objects in these fields.
if you set radius to 0, you won't find objects that are not in the field in the center
fields in the QuadTree
basically this is what I want. is the minimum size of the quad tree the same as 1 terrain cell?
in other words, what I need that parameter for is:
- all objects in a single terrain cell
- all objects that can potentially collide with one another (due to their bounding spheres colliding)
- all objects in a single terrain cell
I don't see how that would work for your usecase.
If you are 68m from the center of an object with bounding sphere size 70m. And cell size is 5m, all objects in current cell won't return you that object
I thought that's how the game stores objects 
basically I thought an object can occupy multiple cells
no
so how does the game do the collisions?
e.g. when a unit is moving around an object that's 50m wide, the game instantly knows they can collide right?
or as another example, how does lineIntersectsSurfaces work? it should still know what object it can collide with 
the game searches for nearby objects around it.
Thats why we have the max size limit
so it's kind of similar to nearestObjects still?
btw what is the max size limit? 100m?
lineIntersectsSurfaces has a hardcoded max radius, that it adds onto the search area
lineIntersectsSurfaces adds 25m radius.
People in modelling say the max collision size is around 70m diameter
I guess that makes sense, 25m*2 and then round up to cell size*2
Go ahead if you could figure it out. This was almost one and a half years ago. I don’t remember why exactly I said what I said, but not because I didn’t want to do this for sure.
so I gave this a bit of thought and I don't think using the current nearestObjects is gonna be fast enough for what I want to do.
because even if I choose a radius as large as the max limit, I could still miss the object by as little as 1 mm.
so that means I have to choose a radius at least as large as 2x the max limit, and once I reach 1x max limit, do the search again 
I tested this on a terrain where the biggest object in the terrain was ~50m, and nearestObjects with a ~100m radius was almost 0.8ms near dense areas (lots of trees, buildings, etc.). it's relatively fast but not fast enough for what I have in mind.
is there any engine solution for this?
can you just do it with line intersects? to check if you're inside the object?
well the thing is I want to use intercept to do line intersections in another thread
so I want to cache all objects LODs (rn only GEOM), and grab objects that I might intersect with as I move along a line, and check intersections with those in another thread
the reason is that I want to do "body-surface" intersection, not line-surface
the thing with lines is that you need many, so you have to repeat part of the calculations, which is a waste
also you need a small spacing between these intersections. otherwise you could miss the surface
The stuff I thought about was indeed as easy as I expected.
But, our intersect does intersect on convex components.
Which all the geometry LODs have, but Roadway does not.
So we need to implement intersection with planes/faces for that
Do you need to detect roadways when the intersect ray comes from the bottom? as Roadway's are one-sided faces
no
I mean walking on upside down roadway lods
Roadway is only top-down
yeah then it's good
arrow = "Sign_Arrow_F" createVehicle [0,0,0];
onEachFrame {
_ins = lineIntersectsSurfaces [
AGLToASL positionCameraToWorld [0,0,0],
AGLToASL positionCameraToWorld [0,0,1000],
player, objNull, true, 1,
"ROADWAY",
"ROADWAY"
];
if (count _ins == 0) exitWith { arrow setPosASL [0,0,0] };
arrow setPosASL (_ins select 0 select 0);
arrow setVectorUp (_ins select 0 select 1);
hintSilent str _ins;
};

Well that.... somewhat mostly works.. kinda.. 😄
go home Ded, you're drunk
btw I know I'm probably pushing my luck here 😅
but do you suppose you could do a debugging check later to see why lineIntersectsSurfaces fails in some cases? 
example:
blep
The green thingy works.
Only when you look upside->down through the roadway. So not bottom-up. But I guess thats fine. I can do bottom-up but effort and also hurts perf a bit
I can do bottom-up but effort and also hurts perf a bit
out of curiosity, why does it hurt the performance? don't you just use a surface intersection equation? that's independent of surface normal direction

Also roadway intersection is supposed to be a lot faster than GEOM right?
It doesn't use convex components, it checks all faces
If roadway has less faces than the convex components on intersect path, then yes
Is it faster than getPos?! 😅
bla bla one-sided face bla bla only intersect from one direction bla bla.
I can write code for both sides, but we don't have that now and I don't care to add it if one side is enough
There you go
Now be happy 😠
nukeded
I updated this ticket as while investigating I noticed it was also effecting vanilla planes wheels in multiplayer
https://feedback.bistudio.com/T150837
it seems like these config sources just don't sync in multiplayer
I'm gonna test the damper source too as it may also be bugged, which would suck as any sort of complex suspension animations usually require it
ah nope damper is fine
Btw can you plz push it to today's update? (If there's one) 😅 I've been waiting for this for so long and I can't wait to test it... 
Also multiple intersections and sorting (reverse vs normal) work as usual right?
Btw can you plz push it to today's update?
Thats not how that works.
Todays update went into staging yesterday morning. I can't push stuff to it anymore
Also multiple intersections
No, just fixed that
sorting is done in the command itself, so should be fine
Result:
0.0062 ms
Code:
lineIntersectsSurfaces [
[4475.14,1470.14,6.46562],
[5440.8,1495,-252.156],
player, objNull, false, 1,
"ROADWAY",
"ROADWAY", false
];
Result:
0.0061 ms
Code:
lineIntersectsSurfaces [
[4475.14,1470.14,6.46562],
[5440.8,1495,-252.156],
player, objNull, false, 1,
"GEOM",
"GEOM", false
Well. Its "slower" for sure 😄
I don't suppose you have one of those "retail exes" lying around right?! 😅 (one of those that you sent me once to test something that wasn't yet released...don't remember what it was)
I have just the perfect code for testing it performance wise and accuracy wise!
I do actually
can I have one plz?! 😄
The dog is like "Yeah, knock her out, she's not feeding me enough anyway"
"bump" about the issue with steam compositions being removed which makes user stuck permanently with annoying error popup every time they start editor/zeus.
https://feedback.bistudio.com/T160334
Perhaps change the popup error to a RPT log or handle it gracefully without any error/log?
so there's something that kind of bugs me a bit.
when you get the position of a unit inside a vehicle (the unit is a proxy) it takes 10x longer than when they're not in the vehicle.
is there any chance there might be some room for optimizations here?
example:
eyePos player
'on foot': 0.0011 ms
'in vehicle': 0.0125 ms
I guess there could be a lot of performance cost when there are lots of units in vehicles... 
I personally expected at most 2-3x the performance cost of a modelToWorld
also another question: why do selectionPosition and selectionVectorDirAndUp take 0.200ms longer when you use number instead of lod names? 
🤔 wuht
_this selectionVectorDirAndUp ["proxy:\a3\characters_f\proxies\weapon.001", 1]; //0.250 ms
_this selectionVectorDirAndUp ["proxy:\a3\characters_f\proxies\weapon.001", "memory"]; //0.0025
sorry meant 0.200 ms 😅
wait I got this wrong. it was 0.0025 ms
so it's ~100x slower 
actually the first one seems to have a variable performance as you said
sometimes 0.250 ms, sometimes 0.140 ms, sometimes 0.070 ms
ok so here's something creepier:
selectionPosition [_this, "proxy:\a3\characters_f\proxies\weapon.001", 0]
0.0035 ms
_this selectionPosition ["proxy:\a3\characters_f\proxies\weapon.001", 1];
0.02 ms
different lods taking different amounts of time?
they're the same lods tho
the unary one's lod index was a bit weird
if you recall I asked you in your channel before
basically LOD 1 in binary version is the same as LOD 0 in unary version
LOD 0 doesn't exist in binary 
actually I figured it out:
https://community.bistudio.com/wiki/allLODs
allLODs player;
[0, "1" , 1 , 91],
| |
lodIndex lodResolution
binary version uses the "lodResolution"
unary uses the "lodIndex"
but anyway, the point is that the lods are the same
yet one taking significantly longer
also the unary version is visual by default
so I don't see why the binary one is so slow
should I make a ticket?
at least I can use the unary one for selectionPosition, but can't do that with selectionVectorDirAndUp. that one's gonna be so slow
I don't want that ticket
well a simple unary version of selectionVectorDirAndUp similar to selectionPosition that works just as fast will do 
right now I'm doing that every frame for several AI
it's slow 
that takes lod resolution, not index.
but anyway, the point is that the lods are the same
no vis lod 1 is not the same as memory
no
I see no difference between lod resolution and string name. Atleast none that would explain +200ms
what about the unary version? why's it much faster?
no vis lod 1 is not the same as memory
I didn't say it's the same as memory
read this part of my message
I'm not using memory here
lod res 1 and lod index 0 are the same lods
so thats invalid
yeah not talking about that anymore
talking about this
the unary one looks up the selection twice
I don't follow. the unary one is faster
The unary one looks up the selection twice, besides that they are equal
so much copy-paste equal even that i should clean this up but eh
ok. well in any case, can you make something similar for vectorDirAndUp?
that's my point here
no
well that's sad. I guess this command is useless for me then 
a 10x slowdown every frame (for several AI) could mean the difference between 0.9ms and 9ms. which I can't afford
yep. useless.
not totally. I still use it for other stuff
but won't do for every frame application
e.g. finding the muzzle end of a unit's weapon, which needs the vector dir and up of the weapon proxy
well looks like vectorDir and vectorUp don't work on proxy objs.
that was the only other solution
so I guess this is a lost cause
(╯°□°)╯︵ ┻━┻
for some reason it can get even worse ¯_(ツ)_/¯
for reference:
player weaponDirection "" // 0.0022 ms
300x slower. yet giving the same thing (except weaponDirection won't give vectorUp)
looks like weaponDirection is the way to go. the hell with accuracy...
yup
actually there's one other thing I can do. "memory" still has a good performance. the only problems were a minor inaccuacy and that it wasn't aligned. but the second problem is solvable
yep, it works
player selectionVectorDirAndUp ["proxy:\a3\characters_f\proxies\weapon.001", "memory"] params ["_vy", "_vz"];
_vx = _vy vectorCrossProduct _vz;
_mat = matrixTranspose [_vx, _vy, _vz];
_m = [
[8.16497e-11,0.99992,-7.55699e-09],
[-0.99996,7.84329e-11,1.59162e-12],
[6.92773e-09,7.52425e-09,0.99996]
];
_mat = _mat matrixMultiply _m;
for pistol:
[[-0.682674,0.729006,0.0505476],
[0.0821446,0.00782176,0.996602],
[0.726116,0.68449,-0.0652224]]
launcher:
[[4.76837e-07,-1.86265e-08,0.999999],
[-0.999999,0,4.47035e-07],
[5.21541e-07,-1,0]];
binocs:
[[0.152574,-0.0469733,-0.987174],
[0.0464975,-0.997421,0.0546474],
[-0.987197,-0.0542392,-0.149997]];
looks like the accuracy diff was not as large as I thought
it's not even noticeable
Execution Time: 0.0112 ms
but will be fine once I write that in C++. but even in SQF it's still 10-60x faster than selectionVectorDirAndUp with lod res 1
Sometimes I think I can script decently, then I see anything involving math and get that idea out of my head really quickly.
dynamicSimulationEnabled for an unit always returns false
for groups it works fine and as expected
something worth to report on the FT?
i think to recall someone mention there is a ticket on the FT about AI firing early even as infantry - cant find it though. anyone knows?
early
you mean while still turning/adjusting aim?
yep
so they fire into the ground near them or airshots
[BUG] AI tank gunner fires (under player command only?) first shot too early and thus always misses first shot: https://feedback.bistudio.com/T152879
i made this for tanks
but i guess its the same or similar for infantry
i think usually its not as obvious as knowsAbout needs to be high enough before they engage
what if you increase the accuracy?
I think it's accuracy related
yet with a info share system (reveal between groups) i think its more prevalent
they have full skill (1)
the normal logic is the AI only fires when on target - can see this with switchCamera "gunner" quite well
but for some reason the first shot seems not to follow that condition
probably yes
can you test something?
onEachFrame {
_target = assignedTarget ai;
ai forgetTarget _target;
ai reveal [_target, 4];
}
see if they aim better this way
or maybe use getAttackTarget instead of assignedTarget
I think the problem is related to how A3 simulates AI spotting. If that's so that code should "fix" it
still firing early when not yet aimed on target
were they assigned a target?
if not instead of _target put the actual target variable, e.g.:
onEachFrame {
ai forgetTarget tank;
ai reveal [tank, 4];
}
onEachFrame {
{
_target = getAttackTarget _x;
if (!isNull _target) then
{
_x forgetTarget _target;
_x reveal [_target, 4];
};
} forEach SPE_CDA_RevivableUnitsToTrack;
};```
pretty sure this is not helping and not the root cause
well I just checked and getAttackTarget is always null for player vehicle gunners
so that code isn't doing anything
i am testing with infantry and AI GL
(that said firing early also happens under player GL)
i think it is already: https://feedback.bistudio.com/T123930
enableDynamicSimulation not working when passed Object
yeah just tested and it's not related
trim command is buggy:
"x" trim [""" '", 0] returns ""
basically any string with length 1 becomes empty
same result for "x "
but " x" works fine
on dev branch?
I remember a report of that. thought it was already fixed
nope. want a ticket?
I think I reported that a while ago, or was it another issue with trim?!?
Can I ask how often are the FT tickets given some love, like for example https://feedback.bistudio.com/T162138 ?
I recently learned about the Show/Hide Object Modifier module and put it to use in a couple of my recent releases. Primarily, I saw it as a way to optimize CPU performance while having a lot of unit activity and scenes for effects (helos flying over, masses of enemy appearing without having to ma...
is this worth reporting in the FT ?
nuked, thanks
Hello everybody,
9 months ago i asked if there is a chance if a certain bug could get fixed:
https://feedback.bistudio.com/T155018
The answer was that a person called reyhard might be able to fix it.
Nothing happend since then.
Does that mean it cant get fixed or that it got forgotten?
The Ticket didn't change so i got the feeling it got forgotten.
it might be fixable but none of the guesses in the ticket were correct so some additional tinkering is necessary
since it's bit more tricky issue, it might to wait till I have some more spare time for debugging
Feature request - "noPauseAudio" parameter
https://feedback.bistudio.com/T163147
Thank you very much for your reply. Would of course be nice if it was fixable. If not AAF at least still has the Gorgon-ifv 🙂
eee i planned to do that last week
Would it be possible to get some more information in MpKilled ?
Like died to collision, died to script damage from script XXX, died to explosion ?
I wanted to log some additional information but it looks likes it only possible with perf heavy detections that are not really reliable
For example this could be used to detect car crashes on the client side but it doesn't trigger for crashes with helicopter or planes.
Killed EH
if(isNull _instigator && isNull _killer) then {systemchat "CAR CRASH!"};
So would be nice to have a easy detection of the death reason in MpKilled.
But i guess that's not something that is easily done ?!
FT for command suggestion in #dev_rc_branch: https://feedback.bistudio.com/T163193
Here something that might be interesting 
That is the reason for the stray entities i reported here a couple of times.
#perf_prof_branch message
#arma3_feedback_tracker message
#arma3_feedback_tracker message
etc
Repro *done on an almost empty server (1 player) (100 "frozen" ai) running at 90 Server FPS
Get a entities log via server exec
logEntities
Execute this in MP local on a client that has a backpack.
for "_i" from 1 to 100 do
{
backpackContainer player canAdd "U_I_CombatUniform_shortsleeve"
};
Get another entities log via server exec and watch the stray entities pile up with every execution of the loop above.
logEntities
So you end up with thousands of these just because you use canAdd or canAddItemToBackpack which also had the same behaviour (probably more).
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:6:23402, Pos: [0.000000][0.000000][0.000000], N:Supply40
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:6:23403, Pos: [0.000000][0.000000][0.000000], N:Supply40
@untold sky sry for the ping but i thought you might wanna know.
Yes
Do you have FT ticket, I can get onto it next week
Ill make one
I hope thats the only cause for this bug. If yes then its gonna be huge
So the client executes canAdd, and that causes the entities to appear on the server right?
Exactly
As seen here in the entitielog (From my personal debugging test/no other players on the server just "frozen" AI)
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:6:23402, Pos: [0.000000][0.000000][0.000000], N:Supply40
> Loc:0, ForDel:0, Dstr:0, DmgDstr:0, DmgDead:0, Sim:1, Inv: 0, AnyPl:0, Out: 0, NetID:6:23403, Pos: [0.000000][0.000000][0.000000], N:Supply40
oof, can't wait to see what dedmen will find in here 
It was the same behaviour on the live server with different scripts using canAdd. (Trader, Player storage system or by just executing canAdd often)
pain and desolation I believe 😄
We still wanna try something different in order to see if it could actually cause the server to masskick all the players
you are looking at it?
most likely a safety feature, since it is supposed to return string
i think i could open image with it
that's also what I tried. wanted to see if I can send the texture to an extension. but loadFile returned a very small string to me (like only ~10 chars) 
Can I give feedback here?
depends. better use the Feedback Tracker site if you want to report bugs
It's a bug. Radio subtitles and server messages are often not displayed at all
well yeah, I want it to load a texture as a string. but I guess it doesn't let you do that as a safety measure to prevent ripping textures from ebos or something 
Just found a data bug:
[1945,9,31,18,27] is on my server 🙂
most likely it hit null when displaying content of your image and interpreted it as end of string
it returns a string. Strings end on null byte.
Images often have null bytes in them, most binary file formats do.
Yes. It's fine. It's all okay. We're totally fine with you OMFG-ing here without telling us what you've found! Carry on 😛
I expect a perfect memory leak now 😄
Well aren't we happy that we have HACK_CREATE_ONLY_LOCAL_CONTAINER_ENTITIES 🤣
maxLoad does the same but it sets it correctly.
Prefixing something with HACK_ is the programmer's equivalent of saying "With all due respect, sir". Anything goes after that 😄
So to get the size of the object. The engine creates a instance of the object.
But if you create a uniform, it creates a hidden uniform container entity in the background.
And because uniform inventory is global, that container entity is global by default.
So it creates a temporary uniform, creates a container entity which it sends to the server and everyone else, but the uniform itself was never sent because its a local only temporary.
So people get the hidden container entity, but never the uniform it belongs to
shouldn't the container be deleted after some time though?
if you do createVehicle, does it delete itself after some time? no
containers do ground weapon holders do
(which is why you have "GroundWeaponHolder**_Scripted**" and "WeaponHolderSimulated**_Scripted**" version)
specifically weapon holders have special code for that yeah
ah, I thought every holder
but yeah it's not fixing the root issue indeed
return false...return false...return NOTHING 
Mh yeah actually I don't know how these container entities are cleaned up 🤔
I thought the inventory item itself takes care of deleting it when it goes away, but I can't find that.
Now I wonder what happens when you spawn a thousand uniforms, and delete them again
I hope for a gentlemen, we got the bug moment
Ah I found how they are deleted..
And it makes me very unhappy.
They only get deleted when you use a RemoveEntityHelper while you are deleting it...
Now what are the chances of the code everywhere reliably having that....
Okey script commands that remove uniform/vest have it.
And when supply's are deleted (vehicle inventory, weapon holder inventory, items inside uniform/vest/backpack) has it too.
And a units inventory getting deleted (when unit itself is deleted) also has it.
I can't think of anything else 🤞
But its always bound to the actual item that owns the container, being somewhere.
In the bug that El'Rabito/KK found the container isn't assigned to any item. It isn't "somewhere"
Only found it because I wanted to add a dlc indicator to our trader and while browsing long through trader interfaces I had internet problems and got kicked. Thought it's a game bug so I started a logentities and I saw all the trash entities. Couple of hours later I had it nailed down to canAdd.
This has been a problem for years and noone had a clue where it came from
And because we now know whats going on here, I found a way to detect this happening.
So we'll now get a popup alert if something similar happens again.
Aaand it works
KK implemented the fix, I implemented the detection for future (though its diag exe only). Gonna be on profiling branch early next week.
Nice!
Might check setUnitLoadout if you didn't already. (Especially if your unit already has other gear *just a brain fart while reading it again)
Edit: Just tried, no trash entities for me.
I'll try to remember later
Good call!
But it seems like its actually cleaned up. But later than it should've been
yep it is. Not actually a leak
Right ofc 
I have replied to you on FT but here it is again https://community.bistudio.com/wiki/showChat will hide all messages, MP interrupt is using it AFAIR, a mod can be using it, someone might remote exec it. Otherwise there is no reason messages just stop showing
Question. Is it possible to disable Arma3 Units at 3den MP preview? It is frustrating to wait when service is dead
no
Could he reduce the timer that was introduced recently?
Could he...
By is it possible I mean: is there any chance, that developers will implement condition to skip this check.
Not planned yet
Found another command with incorrect ret type: ropeSegments
says returns OBJECT; should be ARRAY
thx,But why was it modified by others
And I couldn't find a place to enter the script...
ticket plz
thanks
Guys, so what about that date bug above? Should I create a ticket or something?
yes
nukeded
I think the rendering of CT_SHORTCUTBUTTON is broken because it renders half of the image when using small buttons . Screenshot: https://imgur.com/a/5qv7Nbh
with bigger buttons it renders them fully
ticket time?
KK, was there any answer from config guy?
https://feedback.bistudio.com/T160483#2233797
Nothing definite
Vehicle countermeasures tightly connected to current effectiveCommander. Can it be separated or this is somehow related to AI?
RHS vehicles have two types of countermeasures but in multiplayer they always screwed.
I'm not sure if this has been asked before (I'm sorry if it has):
Would it be possible to be able to pass arguments to https://community.bistudio.com/wiki/Arma_3:_Scripted_Event_Handlers (to BIS_fnc_addScriptedEventHandler and maybe BIS_fnc_removeScriptedEventHandler)and to get the event's ID (and possibly name) within the executed code (just like https://community.bistudio.com/wiki/addMissionEventHandler )?
Currently I believe the only option to pass both is to pass it via a namespace variable, which is viable, but it would be nice to be able to pass/get them to/from the code directly.arguments and eventID
Can make a ticket if it's going to be considered.
use the magic variable _thisScriptedEventHandler
Thanks, didn't see that - however I don't think you can pass arguments, can you?
It was missing from the wiki. I added it just now 
nope, I don't think it's possible to pass args
actually it was mentioned here:
https://community.bistudio.com/wiki/BIS_fnc_callScriptedEventHandler
Wasn't sure if that applied to BIS_fnc_addScriptedEventHandler - but it makes sense that it would. Thanks for clarifying!
also you can pass args according to that page...
Before it's gone and forgotten:
#arma3_scripting message
Therefore my request still stands - can make a ticket if devs want to implement it
I just noticed we have commands called countSide, countFriendly, countEnemy and countType
would it be possible to add "select" variants? e.g. arrayOfObjs selectSide side and arrayOfObjs selectType arrayOfTypes (the array of objs being on the left, unlike those commands where the array is on the right and look weird and unnatural, like the count command) for faster selection than ARRAY select CODE?
Spam comment
https://feedback.bistudio.com/T163245
nukeded
Would it be possible to introduce a new value to the mission.sqm which is always unique to the mission and does not change unless manually changed inside the file? I would like to store some data that is only valid for the specific mission, but briefingName, missionName, missionNamesource can all change.
Every mission contains randomSeed=8307455;
actually they have multiple randomSeed's in them 
Yeah I looked at randomSeed but was not sure what it does. Wiki says it's for randomcommand
but its unique per mission, so if you could pull it out of sqm that might do it?
If it's unique then it's all good.
As we're talking about the randomness... the RNG generator in arma seems weird.
Like if you put 30 objects with 75% presence quite often you will get either very few or almost all of them present.
I've not done any scientific tests but also with selectRandom it sometimes feels like you're getting same results too many times in a row.
I guess its code might be quite old and old RNGs were not that good.
|| Might be also just some sort of observation bias. ||
Dilbert explains it quite well 🙃
https://www.bogotobogo.com/cplusplus/images/RandomNumbers/Dilbert_random.gif
Yeah.
random means exactly that, random, unless there is some sort of pattern or order it is random
I raise you: XKCD
https://xkcd.com/221/
I raise you: XKCD
Wasn't this comic drawn after the Sony security fiasco? (or another console - I don't remember anymore)
I know, but RNGs can be of different quality.
Actually, it looks like it isn't but I wasn't the first one to think that: https://www.explainxkcd.com/wiki/index.php/221:_Random_Number (look for "jailbreak")
heh!
I think I seem to remember that the group who did the jailbreak used that drawing in their CCC slides about the hack 🙂
Do you ever wonder about the people who make tickets about items being despawned and set the category to "Advanced Flight Model"
Is it that they think AFM causes bodies to fly away...?
A is probably at the top? 😄
https://community.bistudio.com/wiki/lineIntersectsObjs
Syntax: [begPos, endPos,withObj,ignoreObj, sortByDistance, flags]
Parameters:
...
ignoreObj1: Object - object to ignore
ignoreObj2: Object - object to ignore
...
this is related to #community_wiki
anyway, I'll fix it
Hm whats the games 'official' spelling of Chemlight?
Because the item/grenade is called Chemlight(<Color>)
But the 3den module is called Chem light
this is with english language selected
Spam comment https://feedback.bistudio.com/T163437#2282377
dealt with, thx
can you move DayZ feedback?
https://feedback.bistudio.com/T163466
Craft a campfire
Craft a sharpened stick
Put meat on stick
Roast meat (without clothes and with clothes)
🤣
all fixed, thx 👍
@untold sky did you make some changes to static weapons get in action? I seem to remember there was a change but I'm not sure
the game doesn't show the get in action for "B_GMG_01_A_F"
try from the front
do you have any other scroll actions? There was some fuckery with reversed get in dir mempoints and action not showing due to that if you have no other actions on the list.
yes, they were rotated the wrong way around
I think that caused that specific object to be broken 
We should've updated all affected objects
while walking around, have you tried constantly scrolling your wheel?
