#arma3_feedback_tracker

1 messages ยท Page 18 of 1

sinful kettle
#

A few pistol bullets cause the building to be destroyed

uncut briar
#

these are mp4`s exported from Steam Beta recording feature so they should be pretty standard.

regal nimbus
#

How do you go back and test 2.16 btw?

desert trench
#

legacy branch

uncut briar
solid marten
#

nope

azure crane
#

Discord ๐Ÿซ 

solid marten
#

there is no reason why it should have this access to my phone, i can post pics from my library, it doesnt need more permissions, dont want my photos on some chinese website

devout wave
#

playing videos on Discord in general just straight up doesn't work sometimes

sinful kettle
#

It's probably better to put it on FT

uncut briar
untold sky
#

to hotfix it I would need to know how/why its happening. And i don't

#

guaranteed messages, are guaranteed. They are not dumped

gaunt depot
untold sky
#

The SQF interface is all SQF interface.
The engine interface is not

gaunt depot
#

Apparently it happened to people with stable 2.16 joining perf 152149 server

untold sky
#

There was a bug before where scope was not correctly opened/closed in some cases, causing crashes. That's probably it.
It was never supposed to be accessible, but due to bug it was

#

If CTD then I need to get them.

modest sonnet
#

Yeah I've had to grab .rpts from community members. I'll see if they still have them else try and reproduce

untold sky
hexed burrow
gaunt depot
#

Also I noticed that it happens if you Alt+Tab during downloading

#

Just had it again now, happens very often

hexed burrow
#

what i saw was that the download wasnt stable, it was slow, then stocked for a bit, then went on faster, at the end there wasnt a file in mpmissions folder

gaunt depot
#
  1. Deleted everything inside MPMissionsCache
  2. Joined 100 players server
  3. Long download (30mb)
20241014213954_1.jpg
#

I noticed it takes VERY long time from starting the join to seeing the lobby now thonk

#

Tried high pop servers in close and far locations

#

Tried with no alt+tab, doesn't matter

untold sky
#

2.16.152149 - problematic
2.16.152100 - fine

2.16.152100 v18 on 07.08

  • JIP improvements
  • receive msg improvements
  • server receive time limit
  • anti-flood
  • bigger garbage
  • increased size of mission file transfer packets
    2.16.152136 v19 on 14.08
  • receive crash fix
    2.16.152172 v20 on 23.08

Question is if v19 was also fine..

modest sonnet
#

Do you just need mdmp or bidmp?

untold sky
#

only mdmp, and rpt is useful sometimes

#

I could understand if the mission download would always fail. But sometimes ?

hexed burrow
#

For the players i tested with, it failed all time at some servers. they tried another server, got a mission file, went back to the first server, download failed again

gaunt depot
untold sky
#

Oh I think I see it

untold sky
#

mission file is split into parts.
Previously each part was 462 bytes.
now its 1328 bytes

koth's 28205136 bytes are split into 21239 parts.
I thought maybe rounding error with how the number of parts is calculated, because that code there doesn't look correct to me.

But yeah again, if that were the issue it would always fail.
And I tried on my server, and its not failing.

#

It only writes the file to disk, when the transfer has completed.
But, it loads the mission, not when the transfer is complete. But when the server tells it to load.

So when the transfer hasn't finished, by the time its told to load it, it will fail.
But the transfer messages are guaranteed, they always arrive in order. So the load message will only appear after transfer is complete.

Clearly that is not the case sometimes. But I cannot reproduce that on my server..

#

Or alternatively, it cannot write to file because permission denied or whatever other issue with the file system.
But there was nothing related to that in prof v18-v19, so if the problem appeared there, its not that.

desert trench
#

Would a special debug build (server and/or client) be possible and help to see whats going on?

untold sky
#

I am doing that right now

#

But that still needs to reproduce the issue

#

Which I can't

desert trench
#

@gaunt depot on what servers can you reproduce it? (windows/linux, server and basic.cfg may play a role?)

untold sky
#

basic.cfg would be relevant if maxPacketSize was changed

gaunt depot
modest sonnet
#

Got an mdmp from someone, just grabbing the associated rpt

#

Alright, got them if it's still useful to you?

untold sky
#

why do you grab them before asking if useful?
Yes gib

#

Nop. I give up, can't reproduce.
I need to know, when it failed to load the mission, what is in MPMissionsCache.
No file? incomplete file? full file that's perfectly fine?

gaunt depot
# 
class sockets
{
    maxPacketSize=1444;  // default 1400
    initBandwidth=320000;
    MinBandwidth=16000;
    MaxBandwidth=4000000;
};
#

Affected server

modest sonnet
#

Yea okay, I'll continue to try reproduce it on my machine

untold sky
#

Why do people even mess with that. Just leave it at default..

untold sky
#

the crash is just AI pathfinding initializing, before the terrain was initialized

#

remove the maxPacketSize and see if that fixes it

modest sonnet
gaunt depot
hexed burrow
modest sonnet
#

Managed to cause a repro by alt-tabbing (fullscreen) while mission file downloading, Stable branch 2.18.1552239
High pop server (70/70), also Stable branch
No __cur_mp.pbo or <mission name>.pbo created in MPMissionsCache

And now I can't reproduce the reproduce ๐Ÿ™ƒ

gaunt depot
#

Its random, but happens often

#

Alt+tab doesn't seem to matter really

modest sonnet
#

Hmm, wonder if I can force it by stressing network

last jolt
# gaunt depot Happens on live servers, very likely started happening at exactly build 152149

2 weeks ago around the time when battleye servers went down there was an ACE.dll that was being blocked for people in certain regions. Once that was fixed there was another bug not allowing players to download the mission file. I had to send it to them manually to place into their mpcache folder, and from there after doing that they were able to join again with no issues as a temp fix. It was determined the players who had the issue all had pings of 100+ while players who stayed at nominal ping did not have this issue. The problem solved itself after 2-3 days. Since then, havent had the issue. Just a note cause the issue sounds very similar but I doubt its the same. It was like 10% of players effected at the time, unable to join easily

last jolt
#

Or increase your ping another way

solid marten
#

So I used Land_i_Stone_HouseSmall_V1_F , named it h and added this to init oneachframe {hintsilent str getallhitpointsdamage h} then set up MG and shot hitpoints which have armor 0.6 slightly more than yours. Could not damage with rifle but MG took it down pretty quickly. What is interesting is that it is the same on 2.16. Could this be that your ammo got a buff somehow?

#

Small caliber doesnt do anything on both builds, the damage to the hitpoints starts with 12.7mm

solid marten
#

minimalHit param will stop hits below this value damaging hit point. Also armor of the hitpoint is the coef of the global armor, so if global is 1000 and hitpoint is 0.3, the hitpoint armor is 300. Then passthrough will also modify the damage.

spice mason
#

Unable to receive the verification email from the feedback tracker. Can someone help me verify my account?

#

Username is Maximum_G

spice mason
#

Received the email finally.

uncut briar
spice mason
uncut briar
#

disable all additonal controllers in game settings if you have them.

spice mason
#

I need to use all of them. One for the flight stick, one for rudder peddals. one for the throttle, and a TrackIR for head tracking. I used the same four controllers before I left the game after the Jet DLC came out and didn't have this issue.

outer frigate
#

was the cfgmovesmalesdr config supposed to be merged in 2.18?

#

havent a clue what it means so dont know if its relevant

jagged nest
#

Any news on when a stable patch might push with the fixes?

alpine tulip
#

Isn't Profiling enough?

woven pollen
devout wave
solid marten
sinful kettle
jagged nest
#

I'm in no rush, just don't want to put my server at risk. I'd rather have players losing mags then duping AT Mines and wrecking havoc on bases on my server.

regal nimbus
#

A lot of server providers won't let you run the profiling branch anyway.

regal nimbus
# untold sky guaranteed messages, are guaranteed. They are not dumped

So if I see cases where some publicVariables or remoteExecs go missing (or arrive 20s+ out of order. Difficult to distinguish), that's some sort of bug? Last time I asked this you suggested that guaranteed packets might be dropped in cases with high packet loss, which matched what I saw.

silent bloom
regal nimbus
#

They only get knowsAbout 1.5. Haven't tested what that does with vehicle turrets though.

#

Against infantry it just wakes them up a bit.

silent bloom
#

Tried it with different locations and the turret aims at you straight away its like howwwww lol

surreal bough
#

PROFILING v3 headless client stopped working

regal nimbus
#

headless client or high command?

surreal bough
#

Now Iโ€™ll just change it to PROFILING v2 and check

#

HC PROFILING v2 same mission - same conditions - HC works as intended

image.png
#

v2 rpt HC

23:13:39 "HC/BIS_fnc_log: [script] initPlayerLocal.sqf"
23:13:39 "HC/BIS_fnc_log: [postInit] BIS_fnc_missionFlow (0 ms)"
23:13:39 "HC/BIS_fnc_log: [postInit] BIS_fnc_initParams (0 ms)"
23:13:39 "HC/BIS_fnc_log: [postInit] BIS_fnc_initRespawn (0 ms)"
23:13:39 "HC/BIS_fnc_log: [postInit] BIS_fnc_keyframeAnimation_init (0.976563 ms)"
23:13:39 "HC/BIS_fnc_log: [postInit] CUP_fnc_emissiveLights (0 ms)"
23:13:39 [CBA] (xeh) INFO: [67410,292.75,58.16] PostInit started. MISSIONINIT: missionName=EVO_FFA_RED_A3, missionVersion=54, worldName=brf_sumava, isMultiplayer=true, isServer=false, isDedicated=false, CBA_isHeadlessClient=true, hasInterface=false, didJIP=true

v3 rpt HC

23:07:18 Error in expression <ctCreate ["colorCorrections", 1776];
            BIS_fnc_camera_ppColor ppEffectEnable tr>
23:07:18   Error position: <BIS_fnc_camera_ppColor ppEffectEnable tr>
23:07:18   Error Undefined variable in expression: bis_fnc_camera_ppcolor
23:07:18 File \A3\Functions_F\Debug\fn_camera.sqf..., line 1070
23:07:18 Weather was forced to change
23:07:18 Error in expression <444)];

        with missionnamespace do {
            BIS_fnc_camera_ppColor ppEffectAdjust [1>
23:07:18   Error position: <BIS_fnc_camera_ppColor ppEffectAdjust [1>
23:07:18   Error Undefined variable in expression: bis_fnc_camera_ppcolor
23:07:18 File \A3\Functions_F\Debug\fn_camera.sqf..., line 949
23:07:18 [CBA] (xeh) INFO: [514877,717.379,0] PostInit started. MISSIONINIT: missionName=introExp, missionVersion=54, worldName=Stratis, isMultiplayer=false, isServer=true, isDedicated=false, CBA_isHeadlessClient=true, hasInterface=false, didJIP=false
#

I think this is where the HC freezes

desert trench
#

can you add -debug to make it log the execution context?

surreal bough
desert trench
#

try again then the next day. bis_fnc_camera is executed in various locations. important to know where its coming from

surreal bough
desert trench
#

@surreal bough try with -skipIntro - maybe also with -world=none

#

HC shouldnt play main menu cutscene i'd assume - maybe something changed/broke with 2.18 tho

#

potentially above sqf error breaks the connection process - or its something else

#

only Dedmen or KK could tell if something changed with -client -connect=127.0.0.1 in that regard

surreal bough
surreal bough
desert trench
#

so wasnt an issue with v2/v1/v0?

surreal bough
desert trench
#

ok. that said a few others reported HC not always connecting since 2.18 (not just v3) from what i understand

solid marten
#

Next update (prof or otherwise) I added limitation on who can join second position. Now only same Group members can take second seat. Side players can still connect and view video feeds. Empty UAV/UGV will be available to anyone of the same Side. This means if you fly quadrotor and land it and release controls it will become available to the whole Side. You have 3 choices: 1. pack it when land 2. lock it with script 3. wait until next release where UAV will know who created it from backpack.

surreal bough
desert trench
#

@surreal bough can you upload/dm me the rpt to check

#

also does it happen now 100% for this HC?

surreal bough
#

RPT is incomplete, there is nothing there

#

I rolled back the server to v2

desert trench
#

well skipIntro/world=none did its job - just connection to server didnt work either

#

can you share another rpt with v2 please

surreal bough
desert trench
#

apparently something broke with v3 - in v2 it loads the mpmission vs cutscene in v3

surreal bough
#

the culprits have been silent for a day... we're waiting

desert trench
#

BI did some anniversary thing today

solid marten
solid marten
uncut briar
#

They're not a copy paste, we used vanilla sources but improved and changed the models

#

@steady panther can have more details, also it's a DLC not a mod ;)

solid marten
#

Ah, sorry I am testing on 2.16 data, my fault

#

I found the problem for excessive bullet damage so this is getting fixed

steady panther
steady panther
solid marten
steady panther
uncut briar
solid marten
#

I do however get this error:

_display displayAddEventHandler [configName _x se>
15:37:13   Error position: <displayAddEventHandler [configName _x se>
15:37:13   Error Foreign error: Unknown enum value: "EMMMenuShow"
15:37:13 File vn\emm\main_menu\fnc\initMenu.sqf..., line 46
15:37:13 Error in expression <Name _x + "_intialised",true];
_display displayAddEventHandler [configName _x se>
15:37:13   Error position: <displayAddEventHandler [configName _x se>
15:37:13   Error Foreign error: Unknown enum value: "EMMMenuHide"
15:37:13 File vn\emm\main_menu\fnc\initMenu.sqf..., line 46```
uncut briar
#

fixed in internal build

solid marten
uncut briar
#

interesting

solid marten
#

yeah...

uncut briar
#

so all buildings without glass were getting excessive damage? :D

#

like it could not find the glass so damaged them even with a small bullet?

solid marten
#

havent tested but my guess this would be related and would probably depend on model

#

doesnt matter as it would be fixed to 2.16 state

untold sky
gray wharf
regal nimbus
#

If the message dropping is an independent mechanism then I'd really like an event for it at least.

#

So that we can tell mission-desync'd clients to reconnect.

solid marten
#

A pentagon, a heptagon and a decagon walk into a bar

Capture2.PNG
daring wagon
#

Please tell me that is a command to fill any area provided as array

gaunt depot
#

drawEllipse with negative radius maybe?

#

Having fill property in drawPolygon would be useful though, have to break it into triangles first right now

solid marten
#

draw your polygon from drawtriangles

#

I suppose I can add fill param to draw polygon but it has to be accompanied with manual center param

solid marten
gaunt depot
#

The dream command could be:
CONTROL drawFullPolygon [[COORDS], COLOR, (opt) FILL, (opt) CENTER, (opt) ANGLE, (opt) SCALE]

solid marten
#

you give actual point coordinates, dont think this makes sense

gaunt depot
#

Forgot about color and full, added

fresh hawk
solid marten
#

stop breaking 2.18!

ripe crystal
# gaunt depot 1. Deleted everything inside `MPMissionsCache` 2. Joined 100 players server 3. L...

Just wanted to put something here that I was able to confirm this issue as well. Also linking relevant conversation that I had with Sa-Matra:

#arma3_scripting message

Recently we had a map designer doing something really dumb and so I assumed that this was a bug caused by us. I still think the majority of it is. Players who were affected we had delete their PBO in MPMissionsCache and after they fetched a new PBO from server the problem was resolved.

I did however just try to help someone over screen share who said it didn't work for him and I was able to confirm that his entire MPMissionsCache was empty- the issue appeared to be in the way he was getting the mission from server. He also did appear to have write privileges of the folder.

Anyways moving to the interesting part that hopefully will help you guys a little more in diagnosing. I tried to have him use our Mission File Updater which source code can be found here:

https://github.com/olympus-entertainment/olympus-mission-updater/

And he was getting the attached image as an error. His game was closed so there really shouldn't have been any reason that it wouldn't have been able to write to that location. Perhaps maybe the folder didn't actually have write privileges or something?

Anyway just wanted to leave this here in the event it helps you guys resolve or troubleshoot the issue. Love all the new scripting commands you guys put out for us this update ๐Ÿ™‚

tevxn3i.png
ripe crystal
gaunt depot
#

@solid marten Did something change in regards to deleteVehicle on dead bodies?

#

Apparently if you delete a body and move yourself into remote vehicle you get ejected

#

Still not 100% sure its deleteVehicle but it points there so far

solid marten
#

repro?

silent bloom
#

Does anyone still maintain arma 3 vanilla AI FSM?

#

Maintain or fix rather

gaunt depot
#

And not always

# 
            deleteVehicle _body;

            if(local _body) then {
                _vehicle deleteVehicleCrew _body;
            };
```What I do on previous player's body
#

Issue goes away if I remove this

#

deleteVehicle then I do moveIn* to the same seat on the same frame

#

If I understand right, deleteVehicle/deleteVehicleCrew does a move out message which ends up moving out a unit rather than a body

#

Might not be 2.18 issue but an issue in general

#

Are move out messages specific to an entity or just seat? If just to seat then it explains it

solid marten
#

confusing, also you should probably do one or the other not both,

#

server might get confused too or you create a racing condition

gaunt depot
#

๐Ÿค”

solid marten
#

which one you remove and it starts working

gaunt depot
#

Gonna try different combinations to see what happens

solid marten
#

if you try to move in too soon I guess you would get ejected

#

as the dead body would still be in due to lag

#

make a function to delete body and move player in from server side, should be more reliable

#

especially when latency involved

gaunt depot
#

Good idea

gaunt depot
#

Gonna see if moving the commands to server side gonna help

solid marten
#

yeah probably will

#

deleteVehicle now can do what deleteVehicleCrew does I think

uncut briar
steady panther
#

On that note ๐Ÿ˜…
we had previously nicely working decals for burned ground and embers but now they dont stack same way again. Any idea if its something 2.18 related?

#

Before

unknown.png
#

any chance for some revelation on how decal draw order is formed? various different decals with different sizes stack up in seemingly random way

image.png
#

It looks pretty cools so it would be nice to figure out how it works and if it can be fixed

unknown.png unknown.png
limpid rune
#

I did report a bug with decals and hiddenselections but I don't think anything got changed/fixed with it

steady panther
#

when I made these I set them with different drawImporantce named property. the burned ground has 4 and the glowing embers have 10

#

at that time that seemed to fix their draw order

#

repacked with embers given drawImportance 100

#

but now other there are other differences in the draw order again

image.png
#

so the drawimportance does probably not matter

#

so understanding how this works would be beneficial to keep it working in the future

steady panther
#

is it simply their name? blobcatsweats

#

though even if it is, it was broken a bit there before I repacked it

uncut briar
#

What if you create them with higher Z/above terrain

azure crane
#

since markers are visible in the 3d view now, could the feature of opening map when clicking on marker category be disabled/toggleable?

steady panther
#

hmm didnt try reloading though ๐Ÿค”

#

no that didnt do anything :C

#

but now that I loaded it up the embers are again under everything. ๐Ÿค”

#

but I also did revert the drawimportance change

#

so got to try that again

#

hmm now it does not work again blobcatsweats

true blaze
solid marten
true blaze
#

Iunno never tried it before

true harbor
alpine tulip
#

You can just say "spam" and we plenty have them already

outer frigate
#

someone's never heard of SEO spam

steady panther
solid marten
true blaze
#

sadness

true harbor
true harbor
alpine tulip
#

We have a looot of spams. They never care where we are

true harbor
#

I wonder if the quickheat pro is actually a good heater

#

About the decals thought they will probably just say to use the class land_decal system unfortunately

outer frigate
devout wave
# true harbor In a feedback tracker? Really? That's such an obscure place to use for this

It's not specifically targeted at the BI FT. The bots just automatically trawl for any website where they can make an account and post. The FT has a high rate of "successful" spam posts because it doesn't have very good anti-bot protection during sign-up; other sites are better at stopping bots before they manage to post. You can see similar problems on older forums that haven't had their security updated in a while.

untold sky
# ripe crystal Just wanted to put something here that I was able to confirm this issue as well....

Players who were affected we had delete their PBO in MPMissionsCache and after they fetched a new PBO from server the problem was resolved.
It would have been very very very useful, to get that "bad" pbo to look a t it.

Perhaps maybe the folder didn't actually have write privileges or something?
It could be possible that the folder is inside microsofts OneDrive? That constantly causes issues.

untold sky
untold sky
ripe crystal
#

Windows OneDrive could also make sense. I have never had this issue but I also manually disabled that crap during Windows setup. Awful feature.

#

Fraali and I retrying to download our mission now- if we end up with one we'll send it.

ripe crystal
#

Fraali and I did not have luck but I'll see what I can do about getting it off of someone.

vague moss
#

Hey y'all, is there anyone who could try and reproduce the issue "Remote Control breaks AI aiming after release." https://feedback.bistudio.com/T179189 for me? Unfortunately, I only have this one machine to test and I'm not sure whether it's misconfiguration on my machine or an actual bug.

#

I'm not exactly sure if this should instead be in #arma3_troubleshooting, but it's regarding an existing/tracked issue. It's not clear to me whether this channel is for meta discussions about the FT or about issues tracked there, so if I'm in the wrong place, do tell!

regal nimbus
#

@vague moss Checked, confirmed and commented. This one should probably get some attention.

jagged nest
#

Has there been any mention of when the next hotfix may come to stable? I haven't seen any new comments on the mags disappearing feedback tracker for more than 7 days now. Just wanted to check in.

regal nimbus
jagged nest
#

Thank you

spice mason
desert trench
#

Ref to nonnetwork object Agent 0x0000025B6C51E0C0

#

is this with agent new? dont think i had this prior to 2.18/latest perf branch

#

(from local host session with just one player/only local host)

uncut briar
#

It seems to be coming from ambient animals, deleteVehicle on local agent

gaunt depot
#

2024, animals are still client-local pepe_acid

#

Technologies of network-aware fauna of A2 are forever lost

solid marten
gaunt depot
solid marten
#

they just didnt like you

dreamy bane
#

Arma animals DLC

outer frigate
unborn acorn
outer frigate
#

That would explain why I found the dog that is supposed to have a wander range of 1 meter about 10km away

gaunt depot
#

Would love to have mouse cursor to be resolution aware

#

Its so small with 4K resolution, very hard to find it sometimes

#

Maybe have it as a hidden option or something, I'd be fine with blurry upscaled cursor tbh

jagged nest
#

Mag reloading issues are not fixed with the Arma update today. Still happening 100% of the time.

#

Full magazines still disappear. Going to test partial mags now.

outer frigate
#

I never had full magazines disappear

jagged nest
#

Just did a test. It only happens if you try to reload your gun without firing. I just fired one shot and now all reloads are fine.

regal nimbus
#

I'm just gonna go back to my old point of "are you sure there aren't multiple bugs here?"

jagged nest
#

The issue where mags disappear if you drag the mag from your gun into your backpack still persists. 100% fail rate. Mag disappears every time.

lost ridge
#

everyone in the server has to also be on the hotfix

#

if there are still players on the original 2.18 version it can still cause this issue

jagged nest
#

I'm the one reporting the issue and I'm on the new hotfix and so is the server. Players can't even join my server unless they have the proper version.

surreal bough
#

interesting moment, I saw it by accident
absolutely empty server in both cases
stable version

image.png
#

absolutely empty server

image.png
#

v 5 prof

image.png
untold sky
# jagged nest Just did a test. It only happens if you try to reload your gun without firing. I...

Can reproduce on official server, escape from malden with me as only player.
But only one magazine disappears, the very first. Later reloads are fine.

Another player joined into the server next to me:
If I fire one shot, and then reload. The partial magazine with one bullet missing, also disappears.
Half emptying my magazine and reload. Same thing. The magazine that was inside the gun when spawning, is deleted.
Looking at inventory, I can see it appear in my uniform, but shortly after it disappears again.
Now even wtihouts hooting, every magazine disappears.
The other player that joined must not've installed the update.

I switched to a coop server to be alone again:
Fire shorts and reload, everything fine.
Restart mission.
Fire no shot, reload. Everything fine, no magazine lost.

So.. what gives?

jagged nest
#

I'm getting inconsistent results on my server. Mags are disappearing but sparingly. It's not happening every time. Also the last session, every time I dragged a mag from gun to backpack it disappeared. Now when I drag it's fine.

uncut briar
#

Ref to nonnetwork object <No group>:0 (Logic)

#

which points to:

       if (!isNull _camera) then
        {
            _camera cameraEffect ["terminate", "back"];
            deleteVehicle _camera;
            _bDestroyed = true;
        };

in BIS_fnc_EGSpectatorCamera

#

as we don't have deleteVehicleLocal command, could this be fixed to handle local objects gracefuly instead of spamming RPT?

solid marten
#

intredasting

gaunt depot
gray wharf
#

then add it ๐Ÿ˜› ๐Ÿคญ

#

so I should add it, or is it another issue? (for future addition requests please use #community_wiki ^^)

solid marten
gaunt depot
#

but I realized its not a big deal since you can call mission namespace functions anyway

modest sonnet
#

Are your servers still having issues with mission file downloads? Ours seem to have cleared up but I'm not sure if it's just players who would be affected using workarounds

solid marten
glad summit
#

not sure if this is the right place to ask/discuss, but has anyone else noticed projectile tracers "playing" from the opposite direction?

#

it only seems to happen when the tracer originates from outside of your view then passes through it

#

but passes through from the opposite direction the projectile actually came from

regal nimbus
#

maybe make a video

#

I doubt I'd have the frame rate to see that :P

gaunt depot
#

I still wish there was drop ENTITY to trigger particle source right away

#

Instead of sending huge array to drop ARRAY and have it processed each and every time

true blaze
#

Preprepared particles? Id love that too

graceful moat
#

We also have players reporting magazines still disappear on our server. Time for a hotfix of the hotfix.

desert trench
#

[BUG] "Ref to nonnetwork object" rpt entry should not be logged when not necessary
https://feedback.bistudio.com/T185807

made a summary ticket for it (plus linking previous reports).

that said wouldnt it make sense to only log to rpt with -debug active?

like there is valid cases for logging it but often a server admin or even mission maker (if from mods) cant do anything about it - given the frequency and amount (GB of rpt size in extreme cases), it might make sense to make it opt-in (or otherwise once per type per session/reduced via the (repeated 29x in the last 60sec) system)

unborn acorn
surreal bough
#

Can sticking to doors be fixed? in dayz I know it was fixed.
this is often observed among players, but judging by the screen, not only players))
my screenshot - I just saw this myself))

image.png
#

he opened the doors, I killed him and then closed the doors - the body with the door and left

#

but when I opened the door back, the body had already come unstuck

image.png
surreal bough
#

Why do they shoot through the wall? bis building

image.png image.png
desert trench
#

could be faulty model setup (view or geometry LOD i guess)

surreal bough
desert trench
#

if you create a ticket with the model name, maybe reyhard/someone else may still check it for 2.20

surreal bough
uncut briar
#

Is it intended that netId returns "0:0" for local objects but "1:0" for simple objects?

solid marten
#

in sp?

uncut briar
#

nope, multiplayer self hosted, from 3den

solid marten
#

odd @untold sky

solemn swallow
#

i don't quite understand how the view-pilot works but would it be possible to add a command and or a parameter that makes so a object is on the same "layer" render as the view pilot? the newest "Wardrobe - Outfit System" mod adds a really cool functionality that uses attachTo to attach objects to the player bone to customize vests and stuff like that, but he removed it from first person on purpose because the player model will always render on a higher "layer" than the objects, having something like that would be nice to help that mod be able to render the modifications we make on first person aswell

outer frigate
#

Have already asked for the bug to be fixed, got no response

sinful kettle
#

It's not a bug

outer frigate
#

Eh?

alpine tulip
#

I know it's a "feature" but lovely to have it somewhat altered

outer frigate
#

Given it doesn't even work for viv cargo...

limpid rune
outer frigate
#

I'm praying it's just something that was forgotten but yeah most likely

sinful kettle
#

Well it can be "fixed" for attachTo without any problems. You can just add an additional bool to its parameters
But you can't change this behavior everywhere

alpine tulip
#

That is I really imagine how it would work

limpid rune
#

...and similarly the 'bug' with dead units having their backpacks visible in cargo reintroduced as a toggle in the vehicle (showBackPackCargo etc)

outer frigate
#

Iirc Viv just uses attachto behind the scenes anyway

gaunt depot
sinful kettle
#

Well you can't just add an array to every object in the game. So no
But adding a single parent is cheaper (obj setDrawParent otherObj)

gaunt depot
#

Won't it be expensive to iterate through all entities while drawing internal view to find ones with setDrawParent refering to your vehicle?

sinful kettle
#

wat? Why iterate thru all? As you iterate over which entities to draw, you also check their parents, and render them in the same pass as their parents

gaunt depot
#

Well, up to you guys implementation-wise, I'd just prefer to let any objects be included like this, not just attachTo-ones

gaunt depot
limpid rune
#

those are decals right?

#

because yeah I can confirm that - i know decal simple objects break this way

sinful kettle
#

idk what's wrong with set

#

The code doesn't filter simple objs for set

#

Nvm I'm blind. You didn't say anything about set not working

gaunt depot
#

Yeah its just about getters, setters work fine

#

Still wish there was a way to set textures on non-classed simple objects (model path) somehow

#

Would enable so much stuff on script level

devout wave
#

A new function was added for flipping things, but that's not the same as giving everyone the ability to do it by default. It's just a function for use in scripts, not an action on all vehicles.

daring wagon
#

The only vehicles that have it by default are the small tracked UAV and the quad.

solid marten
#

and karts, vehicles that you would be able to flip over in real life

solid marten
gaunt depot
solid marten
#

i mean you can setobjecttexture on person with giving "clan" as index

#

but no way to return it?

gaunt depot
#

Can you?

solid marten
#

yeah

gaunt depot
#

I think you can only set it for the emblem, no squad.xml picture

solid marten
#

-> wiki

#

yeah the selection name is "clan"

gaunt depot
#

Oooh, interesting, I completely missed that

solid marten
#

me too

gaunt depot
#

I guess there is no getter

#

ENTITY getObjectTexture STRING for the selection?

solid marten
#

and it would make no sense to return it in array with other textures as it is not really an index

#

yeah will do that

outer frigate
#

Is the same for set able to be done or nah?

solid marten
#

I will not make a new command, just extend the existing even though it is plural

gaunt depot
#

Since you're looking at textures, there is still no way to remove custom texture or return original selection texture if its not defined in hidden selection textures in config

#

Same with materials

#

VBS had #reset texture, maybe bring that over to A3?

solid marten
#

make a ticket with repro plz

gaunt depot
solid marten
#

oh

gaunt depot
#

Gonna check for example vehicle which has no default texture in config and you can't get it

solid marten
#

ok but lets make it work with existing commands instead of making new

gaunt depot
#

Hmm

#

How about

ARRAY = ENTITY getObjectTextures ARRAY
```So it takes both indexes and selection names?
#

And also have it return original texture if it wasn't set?

#

player getObjectTextures ["clan"]; // someclanpicture.paa

#

All in an effort to reuse existing command name

solid marten
#

I will make ent getObjectTextures StringOrNumber

solid marten
#

On second thought player getObjectTextures ["clan"]; is less confusing

solid marten
#

@gaunt depot are you adding an example to that ticket what textures there exist but not in config?

gaunt depot
#

getObjectMaterials still returns empty materials though (no way to get default material), gonna add an example for that

solid marten
#

no no need for it?

gaunt depot
# solid marten no no need for it? 
v = createVehicle ["B_Heli_Light_01_dynamicLoadout_F", player modelToWorld [0,30,0], [], 0, ""];
diag_log ["mats", getObjectMaterials v];
v setObjectMaterial [0,  "\a3\Data_F\default.rvmat"];
diag_log ["mats", getObjectMaterials v];

=>

20:41:43 ["mats",["",""]]
20:41:43 ["mats",["a3\data_f\default.rvmat",""]]
#

no materials when there are obviously materials on that selection

#

My ticket was requesting a command to get these original materials, main purpose was to revert materials back to original after setObjectMaterial but maybe it can be useful for other purposes (checking material properties, etc.)

solid marten
#

what if you set it to "" something changes?

gaunt depot
#

Hmm, it does change the material back to original one!

#

I don't think it did years ago ๐Ÿค”

#

So it works properly, it reverts the material to original one, no #reset needed!

#

I swear it used to be an issue years ago and apparenlty got fixed at some point

#

I guess to revert the texture back with setObjectTexture you need to save original texture that is now properly returned by getObjectTextures command

#

Though no way to do it if its done on init or elsewhere

#

Hmm, I wonder if current behaviour of getObjectTextures could indicate that damaged textures (tex in Damage) are ignored ๐Ÿค”

#

Did a config search, apparently no cases where changeable texture gets replaced with Damage >> tex, only materials but these are "" by default so they work properly

#

Summary:

getObjectTextures

  • Command always returns real texture that selection uses, even if no are defined in vehicle config (hiddenSelectionsTextures[] = {};)
  • There is no way to reliable to revert the texture back to original one if you didn't save it on the moment of entity creation. Setting the texture to "" has its own function of setting empty texture so it needs special #reset texture from VBS to reliably reset it
  • No way to return "clan" selection texture, could use ARRAY = ENTITY getObjectTextures ARRAY version where right operand is an array of selections and indexes to return the texture from
  • Unlike normal textures, "clan" doesn't need #reset texture as "" resets it back to empty texture and normal behaviour

getObjectMaterials

  • Command returns "" if default material is used as very few vehicles have it defined in vehicle config.
  • To revert custom material back to default you just do setObjectMaterial [X, ""], no need for #reset
  • There is no way to get default model material as default returned by the command is "", unlike for the textures where it is actual texture
  • As there is no way to set "clan" material (why would you want that?), there is no need for ENTITY getObjectMaterials ARRAY getter
#

So, all in all if we omit the request to get default textures and materials regardless of current custom ones, here is what should be done:

  • Add ARRAY = ENTITY getObjectTextures ARRAY to allow return of "clan" selection texture
  • Add #reset from VBS for setObjectTextures to remove customly set texture
solid marten
#

wait so there are objects that return default textures from start and if you change it and then set it to "" they wont return to original? I'd like examples

gaunt depot
#

That's literally every object in the game

# 
v = createVehicle ["C_Hatchback_01_F", player modelToWorld [0,10,0], [], 0, ""];
diag_log getObjectTextures v; // ["a3\soft_f_gamma\hatchback_01\data\hatchback_01_ext_co.paa"]
v setObjectTexture [0, ""];
diag_log getObjectTextures v; // [""]
20241027211959_1.jpg
#

Setting to some material and then to "" properly reverts it to default material(s)

#

"" texture has a special function of empty texture, thus the need for #reset or something else

solid marten
#

so material ok, texture not ok? so we do need #reset?

gaunt depot
#

Textures: "" is not reset but special case for empty texture, #reset needed
Material: "" is reset, no special case needed

solid marten
#

might as well add for both for consistency

gaunt depot
#

I think VBS has #reset for both? No idea how it works there

solid marten
#

player getobjecttextures ["clan", 1, 0, "dssad", player]; //["","","a3\characters_f\blufor\data\clothing1_co.paa",<null>,<null>]

gaunt depot
#

to check if clan return really works

solid marten
#

["a3\missions_f_epa\data\img\orbat\b_111_texture_ca.paa","","a3\characters_f\blufor\data\clothing1_co.paa",<null>,<null>]

surreal bough
#

Iโ€™m so sick of being killed through the wall -please add FiredMan to the event - the target he shoots at and Iโ€™ll fix this problem!!!!!!

surreal bough
#

there are so many trivial problems related to just the game, and what are you discussing here?
AI still walks through walls, kills through walls, no sound of AI steps

regal nimbus
#

That isn't even on my top 10 of AI issues :P

#

If they're shooting through walls from building positions, shuffle them back a bit.

surreal bough
#

Yes, I have already said nothing that wayponts work crookedly and any domove with waituntil works much better

surreal bough
regal nimbus
#

There's more than one "kills through walls" case. I don't know which one you're complaining about.

surreal bough
#

if an infantryman or a vehicle knows that you are in the house, he shoots to kill - and then there is a model of the house that lets these bullets through

#

players almost literally know which buildings do not hold bullets and which do not

#

and they are discussing how to display trinkets correctly

#

my desire is to simply remove bullets if the shooting target has an obstacle

devout wave
#

KK works mostly on code things, not models. Him fixing script commands is not taking away development resources from fixing these buildings.

rose cloak
#

"has an obstacle" isn't really a given with some buildings as well ๐Ÿคฃ

image.png
#

(building from CUP, not vanilla)

surreal bough
devout wave
#

What, adding a target to FiredMan?
I have serious doubts that that's possible for that event. It's also used for players, who don't have a knowable target except for guided weapons, and even AI sometimes fire without a target.

near cedar
surreal bough
#

It remains to know that there is target designation and whether there is direct contact with the target

devout wave
surreal bough
#

Doesn't it bother you that AI missiles fly through trees?

devout wave
#

Where did you get the impression that the AI shooting through things doesn't bother me?

#

I just think it's unlikely that the FiredMan EH can be expanded like that.

surreal bough
devout wave
#

Btw, just a small tip, barging in and declaring that everyone else's problems are a waste of time doesn't tend to make them receptive to what you have to say.

surreal bough
#

so - the cry of the soul, nothing will be right

devout wave
#

Look, I know there's a translation barrier here, but you're coming off as rude and aggressive. I hope you don't mean to be, but it does have an effect on how your messages are received.
This is an issue that's been going on for years and years. I'm sure the devs are aware of it, given how many times it's been complained about, and if there was an easy way to fix it they probably would have. So it's safe to assume it's a pretty deep-rooted problem that's difficult to fix. Dev priorities aren't the issue; it's that one problem can be fixed in a reasonable time, and the other one is climbing a technical mountain.

surreal bough
#

off as rude and aggressive? - ohh - no

solid marten
#

If the building is the problem then the building should be fixed, unfortunately I dont know much about modelling. As for FiredMan, there is no target where it starts, it simply triggers on projectile fire event

surreal bough
#

I've been in the arma since 2005, I'm just glad that after 19 years I have a channel with the creators of the game)
accumulated

surreal bough
#

to find out if he has line of sight and manipulate the shot?

#

I understand that no one will remake the models

solid marten
#

there are some commands yes, assignedTarget getAttackTarget come to mind

solid marten
solid marten
surreal bough
#

these are still vanilla ones, which are used to create their islands

solid marten
#

how many? Do you have their types? I can check if we have tickets opened already

surreal bough
regal nimbus
#

It helps a lot to provide at least one specific case, so people can easily test it.

surreal bough
#

these towers get shot at the most

#

I know about the problem of object state on different clients, I made sure that it is indestructible on all clients

regal nimbus
#

There are three different building types in that screenshot that could be referred to as towers.

#

And "shot at"? You mean you're talking about AI shooting into them from outside?

surreal bough
#

everyone is shot through - you donโ€™t have to bother with clarification

surreal bough
regal nimbus
#

The concrete one and the Hbar one are pretty safe IME?

surreal bough
#

you can be on the stairs - they shoot at you through the walls and kill you

solid marten
#

are you using mod?

devout wave
#

Keep in mind that bullets do have simulated penetration in Arma, and those guard towers have fairly thin walls.

surreal bough
solid marten
#

is this vanilla car?

image0.jpg
surreal bough
solid marten
#

is this building a shoot through? If so what is its type

image0.jpg
solid marten
#

vanilla == stock game, no mods

devout wave
solid marten
#

can you confirm ai can shoot through its walls?

surreal bough
solid marten
#

they can change config

devout wave
solid marten
#

config can affect how bullets behave

solid marten
#

i think you can practically shoot through any wall if you force yourself into it

surreal bough
#

a feeling of a reason to pretend that there is no problem, ok - I didnโ€™t expect anything in principle

#

you gave me a shovel to dig with - thanks for that

devout wave
#

I just tested against a CSAT rifle squad and didn't observe (or feel) any bullets going through walls. There are a lot of windows in awkward places so it's pretty easy to get accidentally shot through one, though.
Wouldn't be super surprised if heavier weapons can legitimately penetrate them.

surreal bough
devout wave
#

No, and as far as I know that doesn't affect whether bullets can penetrate walls.

solid marten
surreal bough
#

no problem, if Iโ€™m in the mood, Iโ€™ll make a full-fledged video, or maybe you just donโ€™t play Arma?)

#

It's strange that I'm here alone on the side of the problem

desert trench
#

should be simple to verify via diag_drawMode if the model setup is faulty

solid marten
devout wave
surreal bough
desert trench
#

dev branch + diag.exe + diag_drawMode "MODE" in console

surreal bough
solid marten
desert trench
#

"ViewGeometry" should be viewblock for AI i think

#

"FireGeometry" if bullets should get blocked

#

"Wire" combine with above for bettter visibility

desert trench
#

BI QA has a nice automation process to split a model in all LODs and then you can overlay them ingame in combinations (part of their suite) - this way you can spot gaps or mismatchs more easily

#

like shown here - no roof in a certain LOD

#

if you dont want to/can test yourself then model name (cursorObject in console) + image of problematic area (or short video demonstration) is needed

solid marten
#

Tried these modes and only openings I can see where windows are, would be interesting to see the video

surreal bough
#

give me time, Iโ€™ll set everything up to shoot - everything will be fine

# 
player addEventHandler ["Suppressed", { 
 params ["_unit", "_distance", "_shooter", "_instigator", "_ammoObject", "_ammoClassName", "_ammoConfig"]; 
 if ([objNull, "VIEW"] checkVisibility [eyePos player, eyePos _shooter] != 1) then {
               deletevehicle  _ammoObject;
               //systemchat str _ammoObject
    };
}];

pathetic attempts to somehow fix the problem)

solid marten
#

@gaunt depot player setObjectTexture [0, ""]; player getObjectTextures []; //["a3\characters_f\blufor\data\clothing1_co.paa",""]

surreal bough
solid marten
#

So what is the problem here

surreal bough
#

even here I was able to record damage online now through the walls - is this not a problem?

agile trail
#

If the AI knew you were in there and the calibre of their weapons is sufficient to penetrate the building wall material, seems fine.

surreal bough
#

not funny

agile trail
#

Recommend you start doing some proper diagnostic testing using methods already explained to you, rather than silly remarks,

surreal bough
#

I do this on the go - there are many cases in the game when they just shoot you through a sieve

surreal bough
solid marten
#

I cannot repro. Even if I tell ai my whereabouts it refuses to shoot in that room let alone penetrate

uncut briar
#

I do exactly the same to AI, if I know they're behind a wall I can pen I just spray it ;)

solid marten
#

How do you put the building on the map?

surreal bough
agile trail
solid marten
#

I tricked AI to shoot at the wall, I can see the dust, but no penetration

surreal bough
agile trail
#

Vanilla game (diagnostic.exe)
Katiba 6.5mm
Land_ControlTower_01_F
Bullets hitting wall do not penetrate.

image.png
surreal bough
agile trail
#

Do you accept that running mods (ace, cup, rhs) can change the config of the weapons, and this can then make them able to penetrate even the thickest walls if misconfigured? Do you accept that without specifying which weapon was shooting at you, it could have been 5.56 or 0.5cal?

surreal bough
#

what kind of questions

uncut briar
#

To prove it's an engine/game issue you should have a vanilla repro

#

if it happens with mods but does not without them, it's mods fault ;)

solid marten
# surreal bough

Dunno if related: ```Killzone_Kid - c

Posted on Nov 27, 2013 - 23:09 (UTC)
Make sure you setDir BEFORE you set position. Setting direction after set position could lead to unpredictable behaviour. For example main part of the hospital building in Arma 3 can lose collision detection near both side entrances. AI will also get confused and will stop detecting obstacles if setDir is called after setPos.```

agile trail
#

Exactly what veteran says. We prove it's not the vanilla game, therefore it's YOUR choice of mods or your method of executing that is breaking it for you.

surreal bough
uncut briar
#

also RPT could be useful to see what exactly you're running. I do not trust people anymore, and "ace + rhs" is not convincing to me. AI shooting "through" walls could also be your mission or AI mods like lambs telling them to suppress.

surreal bough
#

I said above

uncut briar
#

cool, give RPT.

surreal bough
#

HC or server?

uncut briar
#

client and server would be best

surreal bough
#

client - my rpt?

uncut briar
#

and if you're running ACE, even CBA settings used in the mission could be relevant as ACE advanced ballistics can affect bullet performance

agile trail
#

rhs, cup, ace
rhs machine gunner with PKM 7.62mm
no bullet penetration

image.png
surreal bough
agile trail
#

Server can't keep up, too many incoming network messages. Remaining in queue: 16

true blaze
#

Is it possible that building is local only and AI dont see it? (Or missing from server, the object)?

uncut briar
#

no, then it would rip him in seconds, on his video it looks like only some bullets hit him and are definitely weaker due to going through a wall

surreal bough
#

I'll make you a video where they kill through the wall - I'm afraid I'll always get an answer play on the vanilla version

solid marten
uncut briar
#

the issue is no one cares about some video where you know jack shit what's going on

#

what's needed is reliable repro

#

a mission where the issue can be reliable reproduced, might be dependent on mods

#

so far Apollo for example tested it with diag exe clearly showing that there seems to be no issue in normal cirumstances.

agile trail
#

Not "some bullets go through some walls, please fix", but "this bullet from this weapon shoots through this wall with these mods loaded"

surreal bough
#

I understand what you are talking about - but I donโ€™t have the opportunity to do this on the vanilla version of the mission, itโ€™s a lot of work
I use mods only for variety or the client part, the essence of the gameplay and mechanics is vanilla

uncut briar
#

as far as I can see in the RPTs you've provided there's nothing out of ordinary, but your server is running bit more mods that you've listed and there's a "secret" mod that just has pbos named serverX.pbo doing god knows what.

surreal bough
uncut briar
#

so bunch of scripts from myriad of sources cobbled together. And you're 100% sure you don't have anything that could affect AI behaviour in there?

uncut briar
surreal bough
#

you're just talking about the problem)

uncut briar
#

If there's no scripts at fault it should be easy to make a reliable repro.

surreal bough
#

and stung there and stung there - in the end, nothing smart and simple, display it in vanilla

#

forget it - there is no problem, the scriptwriter and mods authors are crooked

devout wave
#

Well look, several of us have tried to reproduce it without mods and we can't make it happen. You're the only eyewitness, and you've only observed it while using mods. You see the logic, right?
It's possible it is a problem with vanilla, but right now, all the evidence suggests that mods are needed. BI can only change vanilla, so unless it can be narrowed down to something in vanilla, there's not a lot they can do.

surreal bough
solid marten
#

Not helpful tbh, many questions, like why there is no hit effect?

surreal bough
#

I didn't come up with this

#

I took it online now and itโ€™s always like that for everyone

solid marten
#

For everyone playing your server or for everyone anywhere, because I haven't seen many reports about it, well actually yours is the only one.

surreal bough
#

For everyone playing our server

#

I donโ€™t run to other servers with such questions

solemn swallow
solemn swallow
#

Leopard20 gave the idea of having something like this obj setDrawParent otherObj it would be nice to have something like this to make mods like the newest wardrobe mod be able to render both the vest and the objects that are attached to the player render normally

sinful kettle
surreal bough
sinful kettle
#

I've never seen this except in faulty geometries meowsweats
(+ animated parts near object bounds)

true blaze
#

Ive never had issues like that on ลกumava, playing with many mods ace, cup, rhs, ai mods etc

#

Just to be sure, server runs all the same mods clients do?

regal nimbus
#

I have had issues with AIs walking through the wall of that concrete bunker on the stairwell side.

#

But good luck replicating that.

sinful kettle
#

AI don't have really have actual collision (unless you apply velocity to them). They simply get "unstuck" if the game decides they are stuck

#

Only players have collision

surreal bough
#

what does this information mean?

Setting invalid pitch 0.0000 for HC
lime sedge
#

It's voice pitch. I think it can be ignored. Or you can set it in attributes.

sinful kettle
#

has anyone else noticed the game leaking memory for no reason at all?
I had a VR map running, and there were just a few helicopters, tanks and stuff in the mission and I was mostly flying around in spectator (testing a mod). All of a sudden I noticed that the game is using like 10 GBs of RAM (it was like 3 GBs at the start) meowsweats
I know I wasn't doing anything in the mod that would allocate so much memory (it's a dll and I use a custom allocator with profiling so I know exactly how much I allocate and where), so I'm pretty sure it's the game itself.
Just wondering if someone else has noticed this behavior
If not I guess I'll have to do a vanilla test myself later ๐Ÿ˜…

solid marten
#

I use a custom allocator
try the stock one

sinful kettle
#

No I mean for the dll ๐Ÿ˜…
I'm using tbb for Arma yeah

solid marten
unborn acorn
surreal bough
gaunt depot
#

Through an event handler perhaps?

rose cloak
#

well, existing setWeaponReloadingTime is phase-based as well

unborn acorn
gaunt depot
#

Editing the config on a fly is a big design no-no unfortunately

#

Otherwise we'd be able to add/remove/edit CfgVehicles/CfgWeapons/etc. right in the description.ext on a mission level

#

Which would be a better solution if you ask me

rose cloak
#

Arma3 campaign weapon randomizer speedruns when? blobcloseenjoy

gaunt depot
#

Want slower loading mag? Make a copy of vanilla mag and edit it

gaunt depot
devout wave
gaunt depot
#

2024, still no way to add custom marker colors in description.ext krtecek

unborn acorn
gaunt depot
#

or marker fills or icons for that matter

gaunt depot
#

That's why I suggested an event handler, so you handle all that stuff yourself

#

I'd be fine with your approach but knowing Arma I'm afraid it might be futile

# 
player addEventHandler ["HandleMagazineReloadTime", {
    params ["_vehicle", "_instigator", "_weapon", "_muzzle", "_mode", "_magazine", "_time"];

    if(_muzzle == "GL_3GL_F") then {
        _time = _time * 2;
    };
    _time;
}];
daring wagon
unborn acorn
gaunt depot
#

Offload the burden of storing weapon to time increase association to the scripter FeelsGoodEnoughMan

gaunt depot
#

Also explosive damage

surreal bough
#

I also want to run tests without using a HC client

surreal bough
sinful kettle
#

Your shots are penetrating the walls. You're not shooting thru them

#

This is most likely a model bug (bad surf properties)

sinful kettle
sinful kettle
sinful kettle
surreal bough
devout wave
#

land_house_1w03_f

#

The walls are wooden and not extremely thick, so it may be intentional that they can be penetrated

surreal bough
devout wave
devout wave
unborn acorn
solid marten
#

and why is this so important, would it be better to add more colours in config instead?

sinful kettle
sinful kettle
#

Tho the (inverse) bullet penetrability is probably too low for a bunker. It's only as good as the control tower on Stratis airport (tho it's thicker so it's better)

devout wave
# solid marten and why is this so important, would it be better to add more colours in config i...

previous discussion: #arma3_feedback_tracker message
Allowing custom colour definition would add a lot more flexibility than just adding more presets. (Obviously I'll take the latter if the former isn't possible.)
Being able to fully colour match things to UI elements, etc. would be one big benefit.
Personally I wouldn't describe it as "important", as such, more just really convenient. It's possible to live without, but it would open up more options to have it.

daring wagon
solid marten
#

no, the marker color is a class name

daring wagon
#

Adding more colours to config might add too much clutter. In Some missions though you want specific colours though

solid marten
#

and thats how it is broadcast

daring wagon
#

Also for debugging being able to add more marker colours would be neat.

desert trench
#

RGB would be array, vs existing string for class name, right?

solid marten
#

internally of course marker has rgb color but exposed to the user is the class name and all the networking is done with this name

#

so if the custom colors are added they will have to be in the same classname format

desert trench
#

#custom_RXXX_GXXX_BXXX_AXXX

gaunt depot
gaunt depot
solid marten
#

and then fix broken markers until Arma 4

surreal bough
#

this corner is the weakest point

true blaze
surreal bough
#

I havenโ€™t tested it in other modes - I donโ€™t see the point

#

Today I tested it in different ways: I turned off the HC, turned off all client mods - zero changes

#

Well, I donโ€™t think itโ€™s possible to achieve this with scripts) even if you want to

#

maybe this is a bug of the allowdamage command - I havenโ€™t checked this
yes - I make all buildings unkillable

#

This is the only real protection against everything)

image.png
#

the whole point is that we are playing back a counterattack - and being able to hide and defend is a key factor

sinful kettle
#

Maybe it is a game bug. The model looks fine

true blaze
#

Ill try to replicate on vanila with that bunker on ds, since you cant

surreal bough
#

last hope is to bury yourself in the ground)

image.png
surreal bough
#

everyone unanimously says that they donโ€™t have such problems

true blaze
#

IF its something with ur mods/mission/config, it aint arma problem. Then look for it in your mods/mission/config.

IF it happens in vanilla then its arma problem.

Ill do simple: spawn bunker, go in that corner, make AI shoot the bjesus out of it

surreal bough
#

HC don't forget to screw it on

#

although I tested without HC the same thing

regal nimbus
#

I still force it to 4GB because I prefer the original behaviour for non-huge modsets.

surreal bough
regal nimbus
#

On the other hand I don't know how you get to 10GB with a VR map and a few vehicles.

true blaze
true blaze
surreal bough
#

I remind you of the condition

_bulding allowdamage false
#

what life brings to players)

image.png
#

if you want to live, know how to spin)

solid marten
sinful kettle
surreal bough
#

not monitoring memory leaks is a problem

#

On my server and HC on the same machine I can overclock the processor to 90%, but the fact that 16 gigs of memory is not enough is a fact

#

and it is the HC that goes up in memory consumption

solid marten
#

@desert trench @daring wagon @gaunt depot is there a ticket

daring wagon
solid marten
#

yeah

surreal bough
#

The camera that sometimes jumps is also my problem? being treated, move to another seat

surreal bough
surreal bough
#

As far as I understand, itโ€™s only mine) apparently I have some kind of exclusive assembly of ACE, CUP and RHS

solid marten
uncut briar
#

could we get nearestObjects alt syntax that accepts array of objects instead of types?

#

would be a lot faster than using inAreaArray and sorting in an SQF loop.

solid marten
#

can you give example of this?

gaunt depot
#

And then setMarkerBrushLocal

#

This would make the markers very dynamic and sort of lightweight replacement for drawIcon, instead of sending huge array each frame and doing manual culling, let the engine handle all that.

gaunt depot
#

Anyone else would fine KilledGlobal/KilledAG which triggers Killed for entity regardless of locality? I know there is EntityKilled but its wasteful as it triggers for a lot of crap and MPKilled broadcasts the code when I need it only on certain clients, wasteful again.

sinful kettle
daring wagon
solid marten
gaunt depot
#

Sure there are workarounds but having targeted EH sounds more proper

solid marten
#

you still need to send it over network if you want to trigger it on remote clients

gaunt depot
#

Basically same Killed except triggers regardless of locality

solid marten
#

you assume that entity hit is simulated everywhere and not just where it is local or where shooter is local, but is it?

rose cloak
#

inb4 addRemoteEventHandler that fills remoteExecutedOwner blobcloseenjoy

#

although that should be pretty trivial to implement in SQF

gaunt depot
rose cloak
#

i'd assume it may depend on full MEH code being present where object is local

solid marten
gaunt depot
solid marten
#

it is killed in one place then broadcast the sync globally

#

i think via entity update message

#

the details of the kill are not necessarily broadcast along with it

gaunt depot
#

EntityKilled MEH has same arguments as Killed EH though

solid marten
#

but MEH brodcasts MEH data

gaunt depot
#

So server broadcasts a special message just for EntityKilled separate from entity kill/death message?

#

All in all, once again I found myself in need for a global Killed as I utilise global Deleted quite a bit, if there is a way to have entity-added KilledAG/KilledGlobal to trigger regardless of locality, I'd be a user of that.

solid marten
#

I will have to look the exact mechanism up to answer that which I don't want to do

gaunt depot
#

Also thought of suggesting to call it MPKilled but it could be confusing with MPEH MPKilled

solid marten
#

whats wrong with utilising mpkilled?

gaunt depot
#

Nothing world-ending, I'd just like to have a proper EH

solid marten
#

call empty function or if clientOwner == N call fnc

gaunt depot
#

So far EntityKilled could be a better replacement, just do

addMissionEventHandler ["EntityKilled", {
    call(_this select 0 getVariable "my_global_killed_eh_code");
}];
#

And add entity-specific code to each entity to call a needed function

#

As I said there are workarounds of course, but after seeing how good Deleted is, I wish there was global Killed

solid marten
#

global Killed is MPKilled

gaunt depot
#

Global in a sense of AG instead of AL

#

I guess it needed a better word

solid marten
#

but MPkilled can be added to anyone anywhere and it will trigger everywhere

gaunt depot
#

Not perfect, muh microseconds of needless EH calls and muh bytes sent of the network, think of the electricity wasted, think of the children and the future!

#

๐Ÿคช

solid marten
#

the difference is MEH entity killed broadcasts the resulting game data to everyone and it is assigned to mission, MPKilled broadcasts actual data to everyone and assigned to entity

gaunt depot
#

Special messages just for this reason?

gray wharf
solid marten
#

it returns correctly formatted manually set color. for example set #(1000000, 0.5, 1) will return #(1,0.5,1)

gray wharf
#

noice! thanks, that was actually the thing behind my question ๐Ÿ˜„

solid marten
#

nonexistent or bad format will set to #(1,1,1,1) , which is ColorWhite when you set nonexistent "ColorBlalalalala"

gray wharf
#

and can it accept HTML values ๐Ÿ˜

solid marten
#

Not sure if you can set icon color with classname but this works with all shapes

#

HTML is difficult to encode on the fly

#

I guess R3vo color picker would be easier with 0-1 range

gray wharf
#

yeah HTML was half a joke and definitely not a requirement ๐Ÿ˜„

solid marten
#

HTML format is actually not bad for parsing but but user experience it is probably easier with numbers

solid marten
#

I need to check how markerAlpha works with this, be prepared that color will not have alpha

#

Ok marker Alpha is additional modifier on top of existing color aplha, worth mentioning

#

OK wiki example works just like it says

#

I will add about alpha

uncut briar
#

or do you mean an example of the current inAreaArray equivalent?

sinful kettle
uncut briar
#

objects array

#

any you pass, so it filters them by specified radius around center

#

and filters SORTED by distance

#

equal to something like this:

_objects = [obj1, obj2, ..., obj350] inAreaArray [player, 200, 200]
_objects = _objects apply {[_x distance player, _x]};
_objects sort true;
_objects apply {_x#1};
gaunt depot
#

GIB ARRAY = ARRAY sort CODE

uncut briar
#

would be nice, but still slower than alt syntax for nearestObjects for this case

sinful kettle
gaunt depot
#

I wonder if _objects sort {player distance _x} could be as fast with SimpleVM as native distance sorter

#

Whatever the case, ARRAY sode CODE is a must

sinful kettle
gaunt depot
# 
private _closest = _objects sort {_x distance player} select 0;
private _farthest = _objects sort {(_x distance player) * -1} select 0;
#

or select -1

sinful kettle
#

Because code should be: {params ["_v1", "_v2"]; (player distance _v1) < (player distance _v2); //return if _v1 is somehow smaller than _v2}
Which is shower than computing the values once, storing them in a separate array and sorting that array (which is what the apply method does)

gaunt depot
#

Just let the scripter quantify the properties

#

Sure item by item comparison callback is nice, but I just wanted a better version of current sort

# 
allPlayers sort {name _x}
```And by strings too
sinful kettle
gaunt depot
#

Yeah, I guess asc/desc flag for non-numeric value maskes sense

#

I'd say reverse ARRAY but it doesn't return ๐Ÿ˜ก

# 
Searching 980 files for "_sort sort"
26 matches across 16 files
#

forEachReversed is great though:

Searching 980 files for "forEachReversed"
34 matches across 18 files
# 
Searching 980 files for "getEntityInfo"
18 matches across 3 files
```not much yet
true blaze
#

The thing that penetrated was 30mm AP, on thinnest walls

solid marten
uncut briar
#

yes

#

could be an additional arg to inAreaArray too I guess.

solid marten
#

that indeed be a more suitable command override

#

calling a code as Samatra suggested is not going to be as fast as predefined routine, Lemmie think about what can be done, if we do it it has to worth it

#

Something like ARRAY sort ["SORTTYPE", optional args] where args would depend on the type of sort chosen, could be as fast as qsort can be

#

what types of sort would be helpful, go nuts.

for example
SORT_BY_NAME_ASC
SORT_BY_DISTANCE_2D_ASC, argPos
....

gaunt depot
#

90% of my sorts are distance, mostly 3d

dreamy bane
#

maybe sort by isPlayer, iskindof (human vs vehicle etc)

true blaze
#

Arma 3 Filtration update

gaunt depot
#

Since sort by distance is the most used one, I'd make it a dedicated command.

# 
ARRAY = POS sortByDistance ARRAY;
```No asc/desc flag. You can use `select 0` to closest, `select -1` and `forEach`/`forEachReversed` later. Or just `reverse` if you need it the other way around *that* much
uncut briar
gaunt depot
#

Instead, I'd implement my ARRAY = ARRAY sort CODE idea and @sinful kettle 's sort by direct comparison of arguments

#

ARRAY = ARRAY sort CODE:

private _closest = _objects sort {_x distance player} select 0;
private _farthest = _objects sort {_x distance player} select -1;
#

ARRAY = ARRAY sortXY CODE:

private _sorted_by_distance = _objects sortXY {player distance _x > player distance _y};
private _least_fuel = _vehs sortXY {fuel _x < fuel _y} select 0;
#

Two items are supplied as magic _x and _y

#

Suggesting _x and _y by analogy with other commands

#

Since the whole point of the command is to use two arguments, supplying _this and doing params is wasteful, might as well put both arguments into vars right away

#

Also GIB reverseRet/applyReverse to copy, reverse and return the array

#

arrayReverse

#

ARRAY = ARRAY reverse BOOL where BOOL is to copy or not the array and also return it

# 
private _a = [1,2,3];
diag_log (_a reverse true); // [3,2,1]
diag_log _a; // [1,2,3]
diag_log (_a reverse false); // [3,2,1]
diag_log _a; // [3,2,1]
#

tbh I'd say just bite the bullet and change reverse to return the array instead of Nothing

#

Its not like there are any EHs that act on returned array

#

in case somebody had reverse as last command in some HandleDamage somewher

regal nimbus
#

I just want setZ really

#

that returns the changed array, obviously :P

gaunt depot
#

I'd love ARRAY = ARRAY setZ NUMBER too

#

I have a lot of use cases for setN though, with hashmaps in particular, hate doing HASHMAP merge [HASHMAP, true]

solid marten
gaunt depot
#

Well, it will be convenience AND performance, no need to build an array of arrays only to break it back after sorting each time

#

Not as fast as native sorting of course

solid marten
#

what array of arrays

gaunt depot
#

Currently I mean

regal nimbus
#

I actually felt that native sorting (with sort) was oddly slow.

#

I guess because it has to keep confirming the data types for each comparison.

gaunt depot
solid marten
regal nimbus
#

Might wanna take a look at that function though. Should have a fast path for all-numbers at least.

solid marten
uncut briar
#

i would be fine with additional arg in inAreaArray hide

solid marten
regal nimbus
#

Yeah, but that's an O(N) pass while the sort itself is N log N.

#

I could be wrong though. I was just puzzled at a couple of points that sort was the slowest part of some code.

solid marten
#

lets assume that it is as optimised already as it can be, so sorting numbers would be the fastest you can squeeze out of it

dreamy bane
#

anything faster than BIS_fnc_sortBy is a bonus

regal nimbus
#

nothing is really. That's not a bad function.

#

(it uses sort for the actual sorting)

#

just apply/sort/apply IIRC

solid marten
gaunt depot
#

Agreed, lack of 2d/3d is a problem

#

asc/desc less so

solid marten
#

only?

regal nimbus
#

you want ARRAY sortByDistance ARRAY, looking a lot like inAreaArray.

gaunt depot
#

As for other sorts, I have a hard time coming up what kind of sorts you might needed at engine levels

regal nimbus
#

Or you could add a parameter to inAreaArray :P

solid marten
gaunt depot
#

Honestly modifying inAreaArray* sounds like a best solution here. Old command, very flexible, etc.

#

inAreaArray can't test spheres though

solid marten
#

sort is old command too, even more flexible and versatile

regal nimbus
#

I suspect sort is the slowest part of apply/sort/apply for any significant array now, so yeah, I'm not sure if adding commands is helpful.

#

Assuming that simpleVM works on stable now.

solid marten
#

no apply could be the slowest

#

it calls code

#

for each iteration

uncut briar
#

unless it does not due to simplevm

#

which afaik does it "in engine"

regal nimbus
#

simpleVM apply is pretty quick IME

solid marten
#

is simpleVM something dedmen added?

regal nimbus
#

Yeah.

solid marten
#

then dunno

regal nimbus
#

I'll do some tests later.

solid marten
#

my bet it wont beat native predefines sort functor

uncut briar
#

might be close enough tho

regal nimbus
#

For what array size? :P

solid marten
#

enough to measure

regal nimbus
#

sort's obviously going to be the slowest part at some point

#

I'm guessing closer to 100 elements than 1000

solid marten
#

there is multithread switch for big arrays not sure if enabled

#

1000 is small array

uncut briar
regal nimbus
#

Anyone remember how to force it off? Like ;; at the start of the code or something?

dreamy bane
solid marten
#

we'll see I can certainly prototype it with something simple first

solid marten
#

fun fact: if you stretch pentagon horizontally you end up with cybertruck

cybertrack.PNG
near cedar
#

KK you could get paid millions of dollars for that in Silicon Valley

regal nimbus
#

n log n for sort does fit pretty well.

#

Conclusion: You won't gain that much just from moving the apply to the engine.

regal nimbus
#

btw, if you just want the closest then this is pretty good since simpleVM:

_distances = _array apply { _x distance2d player };
_index = _distances find selectMin _distances;
near cedar
#

nearestObject?

gaunt depot
#

Also GIB selectMinIndex and selectMaxIndex

solid marten
solid marten
#

@gaunt depot could you remind me why did we even add xxxIndexes version? I'm testing something and it is always little bit slower than returning objects or even positions

gaunt depot
solid marten
#

?

gaunt depot
#

When you have custom data with positions and you want it area checked

solid marten
#

yeah, if you give it array of positions it returns positions faster than indexes

gaunt depot
#

Yes, but you then have to turn positions back into indexes to get the data that's alongside these positions

solid marten
#

ok, i pretend i understand what you are talking about. the difference in speed in negligible, probably due to the fact that when you return an element it uses the fastest iteration loop

gaunt depot
# 
private _data = [
     createHashMapFromArray [
         ["pos", [123,321,111]]
        ,["name", "some data"]
        ,["something", "else"]
    ]
    ,createHashMapFromArray [
         ["pos", [321,123,333]]
        ,["name", "some other data"]
        ,["something", "else again"]
    ]
];

private _indexes = _data apply {_x get "pos"} inAreaArrayIndexes [player, 10, 10, 0, false];
{
    systemChat format ["Player is near %1", _data select _x get "name"];
} forEach _indexes;
#

My practical use are smoke sources that can get deleted earlier than I want it (Can't change timeToLive) but I update their positions while they live, then I can check against that position and retrieve other data about smoke source by index

#

Another real use case:

    client_mapDraw_drawTexts = []; // Hashmaps
    client_mapDraw_drawTextsPositions = []; // Positions

        ... then later ...

    {
        private _hashmap = client_mapDraw_drawTexts select _x;
        ...
        _map drawIcon (_hashmap get "text_array");
        _map drawIcon (_hashmap get "subtext_array");
    } forEach (client_mapDraw_drawTextsPositions inAreaArrayIndexes _location);
#

its more complicated in that forEach with zoom checks, conditions, etc. but you get the idea

solid marten
#

ok, how about having hashmaps as well accepted into array where position will be hashmap key called "position"

#

you cannot return index but can return matched hashmap

#

ah no you can return index too

gaunt depot
#

Sounds very specific, I think its fine as is

solid marten
#

getPos hashmap

gaunt depot
#

Would be pretty much a _hashmap get "pos"

solid marten
#

so if you have [hashmap, hashmap, hashmap] inareaarray "whatever" instead of extracting positions from each hashmap and placeng into array then backmatching indexes to find which hashmap, you will have the actual hashmap returned

solid marten
gaunt depot
#

Yeah I get your idea, skipping that apply {_x get "pos"} in my example

#

Issue is what if there are several positions per hashmap or array of positions

#

Very specific usage

solid marten
#

you just make special key #position that will have special usage in hashmap

#

its an object with great potential

#

aaaaaanyway, thats just thinking aloud, I better go back to optimising inareaarray

regal nimbus
regal nimbus
# solid marten can you show the whole code comparison plz?

Building the test data:

testArray = [];
_playerPos = getPosATL player;
for "_i" from 1 to 10000 do {
  _pos = _playerPos vectorAdd [random 1000, 0, 0];
  testArray pushBack _pos;
};

Test for 100 elements:

_array = testArray select [0, 100] apply { [_x distance2d player, _x] };
_array sort true;    // comment out to test apply only

The range select is cheap enough that it doesn't make a significant different to the result.

solid marten
#

so you are sorting arrays in array, it's gonna be slow

regal nimbus
#

well yes. That's like every practical use of sort :P

#

If you could make a much faster one that returns indexes then that would be useful.

#

but it feels like you could do that internally for the [number, array] special case.

#

The qsort operation itself just needs to deal with [number, index]

solid marten
#

So why are we here? What I was proposing is slightly different. I am sorry but I kinda missed where it turned into this comparison, because it is irrelevant to the approach I was thinking of

regal nimbus
#

Well, my point is that sort itself is pretty slow, and probably much slower than it should be for the common cases.

solid marten
#

ah, I remember now, you said sorting just numbers is slow but then show this array in array sort. yes this is slow, do you have an example for just numbers where it becomes slower than apply after 100?

regal nimbus
#

Well, it will for some array size, but I don't see the utility.

#

I think every usage I have is [number, something]

solid marten
#

If I were to improve on it I will need some sort of sqf equivalent which would be the fastest way of achieving said sort with currently available tools so I could compare

solid marten
#

reference in the same array i mean

regal nimbus
#

Yes.

solid marten
#

I think I can improve that

regal nimbus
#

The trouble I think is that sort is dragging the whole thing around internally.

solid marten
#

it is a poorman's version of having array sort code, which would probably be faster

regal nimbus
#

hmm, actually sort is still pretty slow when fed with plain numbers.

#

maybe I should fire up a C compiler and check expectations :P

near cedar
azure crane
#

I keep seeing people who cannot join servers and are stuck in the mission loading screen until they make a brand new profile and try to join with it. Was there any consensus what is the cause?

uncut briar
#

Haven't seen that one in ages, but that's quite old bug (3+ years). I think no consensus.

azure crane
#

unfortunate

#

made a request on CBA's github to add some form of export-import function for the custom keybinds similarily how it works for addon options, because it is very annoing to set everything up by hand everytime this happens. Hopefully doable

#

I can get the player's profile files but not sure if this can help troubleshooting the issue

regal nimbus
#

Maybe if you can replicate by using the profile data yourself...

#

but I suspect that's not how it's gonna work.

#

If they switch back to the old profile, does the problem return?

true blaze
#

Yes, we had such problems (never myself) but if i used their profile, i was able to replicate it

dreamy bane
#

wow, all that spam in tracker

gray wharf
#

bring it

solid marten
#

@gaunt depot do you use inareaarray a lot?

surreal bough
gaunt depot
solid marten
#

ASL is really world position

gaunt depot
#

Sounds great!

sinful kettle
surreal bough
#

the feeling that AI sees the player, but at the same time they also walk along the floors and stairs

wary mason
#

For old inactive feedback issues is it best to recommnent on them or make a new post

sinful kettle
#

Comment

wary mason
#

okie dokie

sinful kettle
regal nimbus
#

I think every single one of those is 2d though.

wary mason
regal nimbus
#

partly because the airspace control function is bad.

solid marten
regal nimbus
#

Yeah, I never had to worry about that. I remember there was an inconsistency with another command.

solid marten
#

@gaunt depot so what did you do about InAreaArrayIndexes inconsistency?

#

So inAreaArrayIndexes expects center in AGL and positions in ASL, not what wiki says, which is even more confusing. There is no way of explaining this on biki so I'm just gonna go ahead and fix it and hope it won't detonate

gaunt depot
#

But yeah, I support just doing the right thing and fixing it

solid marten
#

I will test and then will try add to prof. are your vehicle positions in ASL?

gaunt depot
# 
both_vehicles_vehicleSmokesPosition set [_forEachIndex, ASLtoAGL getPosWorld _x];
gaunt depot
#

I'll change it with productVersion check when its fixed

hardy cipher
#

Hi, I posted a tweak request for the launcher. It would be nice for the launcher to prompt users to download CDLC compatibility data if they try to import a modlist with CDLC that they don't own. Currently, importing a modlist that has CDLC you don't own will simply give an error message stating the fact, but instead, if it asked the user if they wanted to download the compatibility data and replace it in the modlist file, it would be very helpful, especially for newer players.
Thank you.
https://feedback.bistudio.com/T186037

solid marten
gaunt depot
# 
both_vehicles_vehicleSmokesPosition set [_forEachIndex, ASLtoAGL getPosWorld _x];
```was to show that I check `_shot` against array of AGL positions
solid marten
#

but do you really want to fill array with positions instead of objects?

gaunt depot
#

I have to check against positions because _x could go null while I still need its last position

solid marten
#

ah

gaunt depot
# 
    private _index = both_vehicles_vehicleSmokesPosition inAreaArrayIndexes [_shot, 15, 15, 0, false, 15] select 0;
```older version of this line before I figured `inAreaArrayIndexes` uses entity ASL instead of AGL and this code doesn't work
solid marten
#

and in nonindex _shot worked on its own right?

#

it expects agl for center asl for positions you made it flip and it also works, whoa

#

i will make both expect agl by default (currently default for inArea and inAreaArray) and pos world option will be optional

#

world and asl are similar, asl does placing point transformation into world coord while world does bounding center transformation. think for player or vehicles it doesnt matter

#

internally everything is kept in world format

gaunt depot
#

Speaking of positions, wish there was getPosAGL to skip ASLtoAGL/ATLtoAGL, plus we can add large red banner in getPos wiki article to stop using it, so people don't end up with AGSL by accident messing up their scripts

#

Not a big deal for me personally though

gaunt depot
#

@solid marten Why boundingCenter for marker positions?

#

Is this how engine-driven map icons are drawn? Using object bounding centers?

solid marten
#

3d icons?

gaunt depot
#

or you just fixed the wiki and it was always like that?

solid marten
#

getposworld is pos of bounding center

#

there is also placing point

#

getpos getposatl getposasl all return placing point

gaunt depot
#

boundingCenter command return differs from getPosWorld a lot