#dev_rc_branch
1 messages Β· Page 7 of 1
So I can then uninstall all browsers and just use arma as a Webbrowser. Nice...probably also uses less ram than Chrome.
can it download stuff like viruses?
Maybe because you don't have !verify ( #offtopic_bot_cmds ) maybe that blocks you from threads.
Mh.. I would put that on the same level as whitelisting a malicious website.
If the mission maker whitelists something, that they cannot keep safe, its basically the same as whitelisting something unsafe?
And if they know they cannot keep it up forever, they could include the .html directly in their mission pbo, so it never gets lost.
Can not download nor upload any files. (Well javascript could, but it cannot access your disk, so it cannot store or execute any malicious files)
Can you also not post in thread?
π’
why sad?
#dev_rc_branch message because asking to post in thread if able
Another question would be the same thing, about "Extension UI"
https://community.bistudio.com/wiki/CT_EXTENSION
Extension gets mouse/keyboard inputs, and renders into a texture that's displayed as a control in-game.
Technically its about 90% the same code as for WebBrowser (which just sends mouse/keyboard to steam, and gets a texture back)
You could ask the same security questions for it. And do some popup/approve solution too.
But IMO that does not make sense, because extensions can already do anything they want.
Even if we don't do extension UI. A extension already via DirectX access can render anything overtop of the game.
And even without the new DirectX access, extensions could also find the DirectX things by themselves and access it anyway without our support.
And even without that, the extension could act like a steam/discord overlay and render things into your game that way.
And even without all of that, extensions could open a second windows window, that transparently overlays over your game and renders things over top of your game.
So TLDR; for extensions, all of this and MUCH more is already possible anyway, so trying to place artificial restrictions (that would be not that hard to circumvent) onto it doesn't make sense to me.
And Extensions are still subject to BE whitelisting.
We are making this easier to access now, but the old way to render an overlay in a game, is just one google and a couple stackoverflow posts (Or searching through my github) away.
Extension UI would be cool
Extension and WebBrowser controls are in there.
But both are heavily WIP, but they work as currently documented on wiki
so i'm on latest dev-branch and i just noticed that hitting enemies in virtual arsenal with HE grenades makes them not explode. if i hit anything else, the grenade explodes as expected π€
unloaded all mods just to make sure-- same. hitting target with HE grenade = no explosion. hitting ground = explosion
Hand nades or UGL nades? 
UGL
didn't test regular grenades
i guess grenades are hard to test, since they bounce
can confirm (with "arifle_CTAR_GL_blk_F") 
@unreal arrow 
working as intended
I will re-test on legacy

Revision: 152173
Yo, @full sonnet , was that you that I discussed a bug with the damage eventhandlers a few months to a year ago about? I was wondering if the changes to the helicopters you demonstrated where they survive crashes were in this dev branch release.
If you need a refresh, I had a bug where I tried to disable impact and scripted damage to helicopters via a damage eventhandler and it was seemingly engine sided they were blowing up or something. Not sure if this rings a bell.
Don't remember
Here's the thread on the forums where I experienced the problem, not trying to delve into it too much aside from reminding you perhaps of the chat. Since it's off topic for this board, only the question about the aircraft damage stuff was specific to the dev channel however.
https://forums.bohemia.net/forums/topic/236333-handledamage-eventhandler-aircraft-not-working-as-expected/
Hello, Im trying to make it so all vehicles survive impact damage (from wrecks), but can get killed from everything else. The script works fine on land vehicles, it even de-activates on its own, but helicopters dont seem to abide by the same rules, even after blocking the damage types that they g...
You mean surviving after landing upside down? If so I was playing a helicopter scenario a few days ago and this did happen to me and my heli didn't explode
do you mean this? https://feedback.bistudio.com/T181306
from what I gather he is trying to have vehicles negate all damage besides guns/missiles etc
so any collision damage etc should be ignored
but air vehicles specifically want to blow up
repro steps:
- spawn heli (tested with pawnee)
- add event handler that ignores all damage : (vehicle player) addEventHandler ["HandleDamage", {0}];
- gentle crash > rotor gets destroyed for ~1 second and magically appears again full health and the rest of the parts stay fine
- hard crash > kaboom
same thing with jets, only difference with jets is that landing gear are fucky and always break, even with allowDamage false
Not really, I was trying to negate ONLY impacts and wrecks damage like @normal current said.
But it seems there's maybe something scripted at an engine level causing it regardless of the eventhandler. There's a lad in my forum post who actually recreated the steps and documented it beter than I could.
Trying to turn off impact damage from wrecks and crashes on vehicles leaving only explosions from weaponry and projectiles was my goal there, but something scripted at an engine level is making the eventhandler moot.
Yeah, helis still have crash death working that ignores the HandleDamage
I wonder if this still happens https://youtu.be/Pz8UGlqLcvI
lol
it doesn't it just dies ignoring the EH
Didn't know we had Project Orion in Arma
@unreal arrow had that happen actually during my testing. I don't know what I did but lmao.
Added more feedback
CSLA RM-51 doesn't seem to work with ArtilleryShellFired, the EH seems not to fire at all
I don't know if it's the only one
RM-70 doesn't work
The other artillery units from vanilla, CSLA, GM, WS/RF seem to work with it
Is that a pylon-based launcher?
I've seen that used in mods so that the rockets come out of the right holes.
Ah. It's possible, i saw i could change the magazine of the RM-70 from the editor attributes, so i'm assuming it is?
Did it let you change each rocket individually?
no π€
but aircraft pylons also just change the magazine -- at least vanilla ones
okay only the RM-70 has this, RM-51 does not
RHS' grad/bm-21 uses pylons for its rockets too, so can try it if you have it installed
ill test with SPE as we use pylons for the calliope just to confirm if the issue is pylons
nah works fine
so CSLA is doing something different
Thanks for testing!
They've have (had?) an issue with their RM-51 as well, doesn't work with doFireArtillery or the scripted fire mission waypoint last time I checked. I'm not sure if they fixed that, but it would appear something is off with their rocket artillery.
do they have artillery computer available?
Should have
artilleryScanner = true;
in config, does it have it?
this is the only condition that looks like might fail
yeah, i tested using the artillery computer π
I'm adding another mission option "IgnoreUpsideDownDamage" which will ignore this silly damage accumulator that is quite unpredictable as well
It seems handledamage works fine to negate the damage apart from this random variable in equation
Why not make it go through HD instead?
because it is risky
and unnecessary
you only want to cancel it or not, it is so minor you dont want to modify it, but it would fire every simulation which is not helpful
Every simulation? You mean every frame you're upside down?
I see no problem with it
Have it as mission option
if option is on, it goes through HD
yeah no, I played with it there is no benefit of having access to it
modifying value does not do much, except when you zero it. also in most cases other forces kill the heli before this damage even has a chance to accumulate. so if you cancel other damages this one will eventually reach critical value and explode the heli, which brings element of randomness you want to get rid of in this case
also it has no usual handle damage info, so not suitable for that format, as a separate EH yes, but then again no point as it insignificant
I'd get rid of it completely in the engine looks like someone was trying to overkill helicopters
btw while we are at helicopter damage... in advanced flight model, if the helicopter touches ground with anything that isn't dampers, it viciously explodes
https://www.youtube.com/watch?v=bhkDFaI1e6Q
dont mind my piloting
this is also an issue with the helicopter in reaction forces-- if you have the gears up, the dampers rotate from vertical to horizontal, so if you then touch down, they don't touch ground anymore... and the helicopter explodes.
why does this happen
the advanced flight model has weird issues with like (modded?) vehicles and keyboard/mouse controls. with helicopters i just end up constantly spinning without control
I guess we were talking about different issues then
One thing that is still beyond HD for helis is collision damage
It seems like there are two systems at once, both classic heli crash logic AND epe logic at the same time
I think EPE goes through HandleDamage and can be filtered but classic isn't
Or the other way around, not sure
This is evident by source and instigator, source is null in one kind of damage, and heli itself in another
Repro:
- Spawn heli
heli addEventHandler ["HandleDamage", {0}]- Ram the heli until it blows up
Do we have a method of being able to tell if an object or projectile exists on the current machine only?
I'd like to avoid sending references to objects/projectiles don't exist globally over the network.
If we don't have methods currently, could we get them returned in getEntityInfo/getShotInfo? I can make a ticket if wanted.
Do you know if that works for projectiles too?
Uff
Maybe not. But I'd expect shot info to have the local/virtual/fake flag
Yeah the isReal flag
It's false for remote projectiles.
Don't know if there would be a local projectile that's also not real
It looks like netID returns 0:0 for local projectiles
But we might switch over to getShotInfo isReal when 2.18 comes out
Thanks!
make crew invincible, then it doesn't blow up
I humbly submit that that's awful and it shouldn't work like that
we have to leave some arma in arma c'mon
wdym
oh that reminds me, why is arma simulation not realistic enough out of the box?
would it be possible to turn non-essential vanilla things into mods that may be turned into community maintained content?
Example?
hmm, i suppose things like campaigns, showcases, attachments, modular inventory system, idea would be ship a known good version and then have the user update it with the community patchwork
or like map features or idk 2d editor
idk, something like extending engine limitations by making built-in stuff more modular and pluggable
shipping with a out of box experience that can be easily upgraded with the community maintained continuation/evolution of the features in question
am i too weird/confusing?
That sounds like a lot of really big changes to how the game and/or engine works. Not to mention the licensing/liability issues of letting community code in.
There is/was an initiative to take community contributions to the Functions Library, but anything more in depth than that is likely to require more QA and dev resources than are available.
solution: have the community volunteer for that task entirely
Arma 3 is not open-source and I think it's quite unlikely that BI will make it so
BI will never open source one of their active projects
Let alone one of their biggest cash cows
That's both a dumb business move and not how BI does things
okay
well and the day it becomes inactive as far as bi workforce is concerned?
I guess you'll be busy playing Arma 5, by then
And if they have not open sourced OFP, Arma 1 and Arma 2 yet, I don't see them open sourcing their third (fourth) installment
Depends on which "Flashpoint" you're asking about π
The original Flashpoint (not Dragon Rising) was written by BI and Codemasters was merely the publisher
Then they kept the trademark and that's why it was renamed to Arma: Cold War Assault
And Codemasters hired another studio to create Dragon Rising which only used the same name but was a completely different game
In 2011, Bohemia Interactive rereleased the game as Arma: Cold War Assault, establishing it as part of their Arma series and distributing it for free to owners of the original Operation Flashpoint.
https://en.wikipedia.org/wiki/Operation_Flashpoint:_Cold_War_Crisis
Was a joke about Arma being Arma. This is really weird damage, caused by a dead crew inside invincible helicopter, If I find the cause I will fix it of course.
Hello there! Not sure if devbranch followers noticed but after some research and discussion about this issue - #dev_rc_branch message - it was decided to unify everything to use "short rod" geometry for characters.
Hitpoints configuration in base character was adjusted in base character config so if you are using short rod geometry in your character, then there is nothing to do from your side. If you are still using older geometry model, then I would strongly recommend to use geometry from samples.
Hopefully soon (tm) there should be also o2script to help with mass replacement of geometries across p3d similar as it was done with Fire Geometry in past
would be nice to get a log when encountering an outdated model, as there is lots of content that might not ever be updated
Another good question is whether all the CDLC models already use the 'short rod' geometry or if some still need updating? 
at least ws and rf should use the short boys already
Speaking of animation geometry, I really wish offset rotational axis speed would get addressed somehow, been an issue since the beginning of A3.
When rotating the character in animation that has offset rotation (like running forward), you end up with very large velocity, which gets propagated over network and messes up movement prediction on remote players.
You see this with players appearing to run at unusual speeds indoors, getting phased through walls, etc.
Mondkalb mentioned GM uses short rod
Shown: Remote player ran up the stairs, quickly rotating left, huge max velocity registered
SPE too, went through and checked them when it came up along with the armorStructural
Is this channel suitable for game engine update suggestions?
Generic suggestions could be better to put in #arma3_feedback_tracker
No dev branch today, sorry π«
Optimization update π€£
FPS++++!!
Optimizing Poseidon are we?
finally, 60fps
YArma
the new battlebits update looks weird
Too much ethereal oil
"Beauty is in the eye of the beholder" - More FPS
so we can finally play ARMA 3 in Operation Flashpoint graphics as intended
what are you cooking!!!!!!!
his GPU apparently
It's fine, the texture hardware is having a nice rest
the only thing this can fry is old intel iGPU π
not so easily!!!!
i like it, this would be a great gamemode
Just set texture quality to lowest, and you'll basically have that
Was there any traction to forbid "Very Low" textures from server side? I hate that sweaty players all use that for vegetation visibility advantage.
You can detect it and kick the player
Can't you just get it with 'getVideoOptions' and either kick/show full screen warning/cat image?
Until changed
Blur their screen saying you got soap in your eyes
Sure there are ways to hinder players but I'd prefer if there was a way to enforce it instead
Same as say shadows which are super important during night for example
No reason to have them if you want to get advantage in both visual and FPS
Guilty as charged.
You can enforce it yourself, just don't let him play.
I wonder if you could open the UI with video options, set it to the lowest acceptable and click the ok, all from sqf.
Interesting idea π€
It's possible. Done in Enhanced Artwork Supporter
It'll break soon
Well soon is relative, 2.20 is a while away
It's sooner than never
what was the next one again? 2.18? or 2.19?
You can do this, I have implemented it in a sort of scuffed way on an Altis Life server. Force open UI on mission load, force button actions, if it's lower than what you require, set it, then delete other options. ScriptedEventHandler for display 5 iirc to delete unwanted options or set minimum view distance required when you open the display. Kinda cancer to do, but works.
100% would prefer a command or server option to do this though
I think it's sensible that there isn't a good way to force graphics settings. Someone might be playing on low settings for completely legitimate reasons, and it wouldn't be very nice if people could force them straight to Extreme everything and 200% res scale and torch their GPU
check graphic settings and current fps π
if settings are low and fps are high -> straight to jail
Its a way to create level playing field in PvP
I understand why it's wanted, my point is that it would be very easy to abuse or misuse in a way that could potentially damage people's hardware, or block them from playing the game.
just force everyone to low settings π
Not to mention all the other non-graphics settings
Tbh playing PvP on low settings is borderline cheating so I wouldn't mind blocking some people from playing a given game mode on a given server, as the owner
You already can do that safely, by kicking them, or blacking out their screen, or disabling their simulation or whatever. The issue with forcing graphics settings is that it's not a clean safe block; it has consequences outside the current mission, especially but not only if the forced settings are persistent. Graphics overload could crash the whole PC, damage the GPU...and if the settings were to persist after the current mission, someone might even be fully stuck, unable to even load the main menu to change them back.
- If you can already prevent someone from playing by doing workarounds, then there is no harm in adding a simple command that does the same, but without workarounds (assuming the "I wouldn't be able to play with such settings" line of thought)
- Then just do it non-persistent, but mission only
- If setting higher graphics settings would damage your GPU, then... you should ask your manufacturer for a refund, seriously
wasted script command, big harm
If it's being used, then it's not wasted. No harm π
It really was disheartening to have to choose whether I wanted to stay true to myself and play the Arma that I wanted to play, while knowing that I'm probably giving the other "hardcore" gamers an advantage or switching to Arma-minecraft mode and being on par with them, visibility wise π¦
Kick = loss of traffic
Which becomes more and more scarce
description.ext so its mission specific
Wait, so which display settings give you an advantage, aside from better FPS?
lowest textures, lowest objects
its arguable of course, but there is slight advantage due to loss of texture detail
Helps for picking out targets at long range against background?
I didn't think it went low enough to be useful.
grass is locked to terrain grid size, isn't it?
hmm, can kinda see it with "very low" textures.
yes, clients use server setting at minimum
but
if the client has a way to use setTerrainGrid, the grass can disappear
youve got more to worry about if client can arbitrarily execute code
Looks like very low settings make much more difference at close range than at distance, so I'm not sure it'd help much. Makes all the nearby stuff blurry, but long range detail is similar.
Yeah I tried making an example and the difference is not that big over distance
Also some tree loses its leaves completely
bad models :/
most of the tree lods are horrific
too late in dev to have them changed though afaik
One LOD is missing leaves or something
All trees and bushes seem to do that actually
Happens randomly
thats not engine issues thats p3d issues
one of the tree lods is indeed missing leaves
iirc lythium uses it extensively
oh ok my bad lmao
well hey, if you like thin trees it's right up your alley at least :D
There's some engine oddity with which tree LOD it picks, but yeah, bad LODs.
yes
same often happens when using binoculars or any zooming
lol that's not so bad fwiw, lod equivalent of mipmapping ftw
reminds me of comparing this to how source engine implements lod switching
No vegetation at some angles
Also there is fraction of a second during lod blending where you can see everything behind the vegetation
doubt this kind of stuff can be fixed with video settings lower limit though
Or maybe it can, I couldn't get bad results in standard at all
If there is this kind of LOD glitching, it happens way too far away to notice
Very low "Objects" is the culprit for this mess
Works much better on standard objects setting (next after lowest)
So if anything to limit, its the objects setting
These trees are still bad though
You need both standard textures and objects for this issue to go away
So yeah, both needs to be limited to stop gameplay advantage due to LOD issues
I wish there was a way to disable LOD and force the game to use the best looking one always (for distances below 1k), on a plane or heli it makes almost no sence since objects are so far away that full LOD doesn't benefit the scene)
I have modified the objects setting beyond ULTRA via tweaking the config files in the My Documents folder and it resulted in a lot less stutters when zooming in/out and less FPS spikes for me (when turning around fast)
And FPS was maybe lowered by 15% at most (for me)
Don't know if the huge number I used (on the config file) was enough to force all objects into max/best LODs all of the time but I could no longer notice any difference between normal view and zoomed view (hold right click) in terms of quality (texture poping was no longer a thing)
Perhaps with the upcoming shadows improvements (no longer render for all objects even the not visible ones?) and the raycast improvements (objects behind other objects making them not visible are no longer taken into consideration for some physics calculations?) the impact would be even more neglegible. It is a matter of waiting for 2.20 π
Dx12 will fix all thatπ
Honestly I am impressed by all the optimizations lately, perhaps the benefits of a Dx12 implementaition have already been surpassed by now or will soon be. Arma 3 has never felt better than now. Still no one will complain if the devs have the audacity for even bolder endeavours π«‘
no thanks
vulkan >
uhhh, is arma 3 on mac implementing Metal?
or doing a dx conversion?
Are they even developing the mac version anymore?
How many people is playing Arma on a Mac right now?
...5?
yes, they're just slow for whatever reason, it been outsourced to a different studio
"For whatever reason"? Perhaps due to the Mac market being so small that BI is basically gifting those 5 players all the costs to outsource the continuing develpment on that platform
Most of my friends with macs just install windows via bootcamp anyway, it is the only way to game efectively there
yeah, how about apple silicon users huh?
where was this mentioned? first time i'm reading about this
I made them up, just my personal wish list π
..
lol
There was an issue with shadows being included in calculations even if shadows are disabled. So doesn't really matter for the avg. User
also shadows distance 0, still worse FPS, unless you really disable them 
there is also a problem with enabling/disabling shadows ingame
need to change settings and restart arma for them to apply
previously it worked on the fly while ingame
and if you set shadows to like 3000 m and you have a mountain in front of you, buy you're still looking in a direction that has many many kms of objects that have shadows, despite them all being behind the mountain, FPS still tanks heavily
all of that started after making accessible the extreme settings in the patch
RIP
so basically there's no object/terrain occlusion culling for shadows????
Its kinda complicated to do occlusion culling for shadows since a shadow from an object thats completely obscured should still be visible
ex. Large tower obscured by a building casting a shadow next to/past the building on the ground
there is some solutions but idk if Arma used them
Update 2.18 RC is now available: https://x.com/ArmaPlatform/status/1833861763948810423 
yesss can't wait for 2.20
what does RC really mean?
Release Candidate
This is an official update?
this is a Release Candidate, available for testing as (for example) a separate branch in Steam. If testing doesn't show any issues - it'd get pushed to full release state (as in: main branch)
Added: Mission option to ignore upside-down self damage for helicopters
Didn't dedmen completely remove this previously? Or is this another, separate system for auto-damaging flipped helos?
I removed the instant kill.
Now they just get normal damage upside down
Or in some cases they act... Arma-y
hmmmm
?
look; you want them FPS or not??
so it doesnt seem to happen at all times
but often enough that i have noticed within 5 mins in 3den
so either my pc is dying or something happened in todays dev-branch update π₯²
it can always be both
It looks like it could be related to Dedmen's model loading fix
But there was nothing about it in the changelog
restarted game just to make sure, but still.
i mean, it's easy to test. if someone has dev-branch active, just boot up the game, go into 3den and flying around a few seconds. should be visible then
what does "mission option" mean? description.ext only?
I hope we can tweak it with mods.
Is it only loading slow? Or not at all?
if i dont move it doesnt seem to change at all. if i fly around, it starts to load stuff
Mh but it's models, not textures.
Maybe my counting was off by one
Wo gehobelt wird fallen spΓ€ne
aaf with new adaptive camo system
interestingly, not every model suffers from it
my guess is house lod textures are working because they are small? compared to vehicles at least, which even in the last lod will still load the regular textures and not some special mlod texture
hmm. I found a couple of walls and buildings where the top LOD hadn't loaded in, but after I moved around for a few seconds it did. Not sure if it was the movement or the time.
I probably messed up the preloading. I will find out in time
ok, I flicked to Georgetown and nothing is loading in :P
Now that dev branch is so broken. I recommend trying out RC branch, it has the real big goodies
it's the langoliers eating the world
Clearly you must implement DX12 direct storage now π
so no need anymore for window full screen, instead of full screen, to avoid the game freezing when trying to alt-tab to come back to the game?
It fixed one case of alt tab issue. There are probably more
Made a quick test with my old modded Exile server and a "Vanilla Exile server"
No issues so far π
Also made a Yaab test, 127 FPS.
With latest perf build it was 123 FPS.
Most of the performance things aren't in yet
playerTargetLock
just need a set version now
have a custom radar system, for an apache, that leaches the input from getsensortargets, but only shows the top 16 important targets, so we have to have custom lockon systems and i cant control the sens radar, always slowly moving towards my goal
that is awesome to see
- 4 FPS?
when you launch a dynamic mission with AI, the fps can jump +- 50 fps on the server
the client fps starts to depend on the server fps if the server fps is less than 20
turn off the grass - it will give you a boost of 10 fps
i must be experiencing perceptual blindness, i never noticed a grass option
setTerrainGrid command
terrain quality
I think with setTerrainGrid you don't have to sacrifice terrain quality to get rid of grass 
Or maybe they both do the same thing and I misremember π
both same thing
removing grass = lower hills/mountains = rocks floating in the air and enemies visible on the other side of the hill/mountain, through it, since it's lower, when lower quality
That can still happen at high terrain quality settings too, at longer ranges. It can be pretty annoying in tank combat.
I don't see how that solves the problem
it doesn't 
Starting getting errors like this on the dev branch today when closing dialogs, happens randomly but often for long periods when it starts.
vram usage is fine?
https://steamcommunity.com/sharedfiles/filedetails/?id=2987839876
Removes grass/clutter so that players can use Full terrain detail while having no grass
Sorry, let me rephrase what I was saying before.
"At high terrain detail settings, the terrain LOD still changes at long range, causing objects to appear to float, or be visible through hills."
Removing grass won't fix that. There will still be discrepancies in the terrain at long range, but now everything else will be uglier too.
you can disable grass by modification of clutter class in the given terrain config - thus you can use very high terrain detail with no grass and other clutter
that said its a poor way to achieve the goal. it should be separated in the engine finally
Okay, so it's a solution to what Groove was talking about (terrain quality and grass being attached), but it was pinged at what I was talking about (there is no setting where terrain never deforms) which it doesn't solve at all. So that's where I've got confused, I guess.
Welcome back Arma 2 no DLC texture
or Arma 2 Free
183855
back to the future
lol
Hey question, if someone were to make a feedback tracker ticket, what are the chances of UAVS getting a 1000m height setting between 500m and 2000m?
There already is one 
It is Depending on the UAV. If there is an option, could it be a manual option? But 2.18 is on RC, so that will need a while.
Is there? in current branch or dev/rc?
Ticket
I asked for it some months ago or so
Hello Is there no longer a development branch for servers? This page https://community.bistudio.com/wiki/Arma_3:_Dedicated_Server mentions there is one, however I cant find it on steam, and the file it suggests editing doesnt exist
no need to post the same message everywhere. Doing so just here would be enough
I wouldnt consider two place everywhere, but noted
and judging by this https://community.bistudio.com/wiki/Arma_3:_Steam_Branches#Arma_3_Server you have the rc branch available when it's online
Not exactly very useful considering its mostly offline
Thanks though, It's odd the link I sent mentions the ability to select the branch through a .cmd file, but said file doesnt exist
It is online rn
Where do I see that? The site just says "Stay Updated"
Ah lower down in the right it says the access code, thanks!
Do you know how to get the servers to be selectable when using the games dev mode? Its greyed out
Dev or diagnostics?
Dev
That should work iirc, diag has it disabled
Which exe cooresponds to the dev build? My server isnt showing up in LAN even though its set to the RC build
Hm the regular 32 bit one has access to 2.18 commands so I assume its dev, and yet the server is missing
eventually got it working with the RC branch, thanks everyone for your help
Dev branch client should also have the server exe's in it
its a exaggeration and it was attempted to say to just not cross-post in the server.
there isn't really a dev branch dedicated server because its not necessary it seems, but the server binaries are included in the client as you've gathered.
I couldnt get the server to come up when using the dev branch client and the server was on rc, worked with both ac though
well, yeah, maybe worth filing a ticket since it sounds like a bug?
oh yeah, i forgot, i was going to say that as well
Would it be possible to add a command to offset playerβs position marker on map/gps?
For example playerβs actual pos is [123,456] but the map shows his position at [789,987]
Cool use case would be for electronic warfare / jamming scenarios
Better suggestion: let us hide more map marker things
the GPS still shows your current grid even if you hide markers
The grid is just a control so you can hide it too (afaik)
lol
That doesnt achieve the effect wanted
Everything should be the same, just the position of player offseted to somewhere else on both maps and gps
And AI too if they are affected by the command
just q, anyone has issues with the 2.18 RC servers we put up ?
have issues with the whole game :p
You on Dev or RC?
lol
Could we increase some visibility of drawLine3D at night? Its still almost invisible at night even with the new width param.
And maybe correct color under NVG and TIs to make it cooperate better with drawIcon3D?
https://imgur.com/a/oXIxr3q
onEachFrame {
_width=20;
_beg = player modelToWorld [-50,100,5];
_endR = player modelToWorld [50,100,5];
_endG = player modelToWorld [50,100,10];
_endB = player modelToWorld [50,100,15];
drawLine3D [_beg, _endR, [1,0,0,1],_width];
drawLine3D [_beg, _endG, [0,1,0,1],_width];
drawLine3D [_beg, _endB, [0,0,1,1],_width];
drawIcon3D ["",[1,0,0,1],_endR,0,0,0,"Red"];
drawIcon3D ["",[0,1,0,1],_endG,0,0,0,"Green"];
drawIcon3D ["",[0,0,1,1],_endB,0,0,0,"Blue"];
};
code used on the screenshot
FT ticket
Roger that
I've played with four other guys on my own server with 2.18 RC last Friday and Saturday for many hours. We had no problems with vanilla and addons/mods (mostly CUP)
thanks
@full sonnet I found an issue in RC.
Placing a trigger by double clicking in Eden (The small asset menu) the trigger gets placed at 0,0,0. Other entities work fine.
If you then select the trigger in the entity list and press "F" a script error pops up.
>
18:06:57 Error Type Not a Number, expected Number
18:06:57 File A3\3DEN\Functions\fn_3DENCamera.sqf..., line 150
18:07:11 Error in expression < then {_pos set [2,(_pos param [2,0]) + getterrainheightasl _pos];};
};
} foreac>
18:07:11 Error position: <getterrainheightasl _pos];};
};
} foreac>
18:07:11 Error Type Not a Number, expected Number
18:07:11 File A3\3DEN\Functions\fn_3DENCamera.sqf..., line 144
18:07:11 Error in expression <ected "");
if (count _pos > 0) then {
move3DENCamera [_pos,true];
};
};
};
>
18:07:11 Error position: <move3DENCamera [_pos,true];
};
};
};
>
18:07:11 Error Type Not a Number, expected Number
18:07:11 File A3\3DEN\Functions\fn_3DENCamera.sqf..., line 150
And the game crashes at some point 
See DM for dump file
we also found a bug with 3den and rc branch
when starting our ws extraction scenario in 3den in sp mode, error popups about variables show up
this only happens in eden, and only in sp. if i start the mission from the main menu, there is no error
also this only happens on the rc branch, not on stable
really weird
hm no issue with any other CDLC
SPE was working before i switched the branch π€
down is still in progress. so cant tell yet what files this is
(switched from perf branch)
would it be difficult to have a function for telling the sensors to lock a valid target
like
_bool = vehicle player setSensorTarget Object;
seems the branch switch or something else corrupted my SPE install - verify game download 15 GB of it again.. can launch RC without errors now
I was get some cfgSurfaces... errors when tried pack addon via addonBuilder on Dev
Success when set RC branch
Build failed. Result code=-1073741819
Binarize task failed.
20:02:16: '/' is not a value
20:02:16: No entry 'C:\Program Files (x86)\Steam\steamapps\common\Arma 3 Tools\Binarize\bin\config.cpp/CfgSurfaces/BuildingRubble_exp.dust'
20:02:16: '/' is not a value
20:02:16: N
more here
https://pastebin.com/3d7MrRda
Pastebin.com is the number one paste tool since 2002. Pastebin is a website where you can store text online for a set period of time.
Same on RC, just try pack any vanilla or mod .pbo
It tries to render trigger at nan,nan,nan position.
The "crash" you sent, is a freeze. With a trigger at nan,nan,nan position. It freezes when you open the map.
It tries to render the elipses, with -2147483648 line segments. The number of segments can never be negative. But whoever wrote this apparently thought it should support negative numbers anyway, but then also didn't handle negative numbers at all causing a freeze.
Its a refactoring messup, and applies to all trigger creations. I split a function in two, but forgot to call the second part, from within the first part.
Thank you π«
Glad I could help.
Mh yeah.
lxws_requiredintel is undefined.
but lxws_helipad, lxws_task1-9, lxws_inttable
I can see many variables be present. But I can't really diagnose why that variable is not set, I need to know where it supposed to be set at, that script is probably not running. If I know which maybe i can find why
thing is, for a test i just defined the variable again right before that error pops up. and then on mission start it'll complain about the next variable π
via addonBuilder on Dev
Tools dev branch? You mean?
Afaik we didn't update the config there, so that shouldn't be a new problem?
the variable is initially set in initserver.sqf file
Well I don't know where they're supposed to come from
maybe it loads too fast?
Ah
or too slow
Yes.
initServer.sqf is called AFTER the variable errors happen
and it's called in initplayerlocal.sqf if i remember right
so i guess load order has issues
https://community.bistudio.com/wiki/Initialisation_Order is on your side there, BUT.
initServer.sqf is scheduled. If its started, while the scheduler is already full. It will not execute the first line.
its controlled by initFunctions.sqf
if (!is3DEN) then
{
if (isServer) then // server execution (hosted or dedicated)
{
if (fileExists INIT_SERVER_SCRIPT) then
{
[] execVM INIT_SERVER_SCRIPT;
INIT_SERVER_SCRIPT call bis_fnc_logFormat; // log only if executed
};
};
//--- Run mission scripts
if (!isDedicated) then
{
if (fileExists INIT_PLAYER_LOCAL_SCRIPT) then // client execution (incl. hosting client)
{
[player, didJIP] execVM INIT_PLAYER_LOCAL_SCRIPT;
INIT_PLAYER_LOCAL_SCRIPT call bis_fnc_logFormat; // log only if executed
};
[] spawn {systemChat "A";};
[] spawn {systemChat "B";};
prints A, B
[] execVM "A.sqf";
[] execVM "B.sqf";
prints A, B
I didn't mess up the execution ordering 
But what definitely happens here, is we spawn both at once.
If scheduler is full, initServer.sqf might suspend after first instruction.
And then because it "ran", it is at the end of the queue.
Next frame, initPlayerLocal starts at the start of scheduler timeframe, and gets the full 3ms to itself.
So.. race condition?
You cannot know that initServer.sqf has finished before initPlayerLocal.sqf runs
but why it was never an issue before and is with rc branch now?
i've slapped a waitUntil { !(isNil "lxWS_requiredIntel") }; in the beginning of initplayerlocal.sqf π
do we have an eta for the official release? will have to try to schedule a ws patch then for soon after
Not before October, but it basically depends on all the (internal / external) testing going on now. We're tracking a handful of issues so far, probably doing a RC update next week or so.
ok, thanks. sounds good
in the past we occasionally had issues with linux servers. makes me wonder now if issues with the script scheduling caused those.
If I'm not mistaken Tools Dev has several changes that were not pushed to stable with 2.16, they would be pushed with 2.18.
Among them changes to the audio encoders and binarize.
https://forums.bohemia.net/forums/topic/169362-tools-development-branch-changelog/?do=findComment&comment=3525025
hmmm, would it be possible to have a stable-compatible experimental branch of tools?
Tools dev branch is generally stable compatible
At least. Currently. We don't really do compat breaking changes on tools
probably a long shot but wanted to ask before i put a ticket in, is it possible to correct the link between client fps and firerate of weapons, im sure its probably not, but wanted to ask
isnt this issue as old as arma?
There once was a fix for it, a decade ago
I've read this like Skyrim's bard sang "Ragnar the Red". Need a full song.
it's the beginning of a pirate song
"there once was a fix, like a decade ago, hey ho hey ho"
Dev got an arrow to his knee before he could commit the change though.
https://www.youtube.com/shorts/n_9Ov2vQi_8 got that in mind
(Original song - downwitharron on Instagram) The endless loop of your normal coffee drinker. Bean juice makes me go fast but NOT FAST ENOUGH!!! #animation #meme #art #comedy #coffee
What a lie. There is no "too much coffee"
well i hope it can get fixed
i made i fix with a script to alter the reload phase,, but it messes with the audio so not a good solution
i think to recall "things" were fixed with the to SQF export from Eden - if so, does anyone have FT or changelog at hand?
- doesnt covert bool correctly in setVariable
_this setVariable ['SPE_ARTY_AI_DIR',1,true]; (_this setVariable ['Enable_Dir',0,true]);
=>
_this setVariable ['SPE_ARTY_AI_DIR',true,true]; (_this setVariable ['Enable_Dir',false,true]);
- double semicolon at times
_this setname "Reinhold Poehler";;
_this setspeaker "spe_male02ger";;
_this setpitch 0.979453;;
- faulty case for some sqf commands
- obsolete space at end of each line
Do you know the name of the script that does the exporting? And which pbo it's in?
It's probably a script right?
It is. It's in the eden pbo
1200 lines of joy. Have fun
Oh shit
I found even more issues
Need to check what's behind 0.5 == 0.5 when I am back home.
Set variable with %s placeholder in attribute doesn't work at all
Just gets parsed to 's' instead of the variable name.
This is also bad as layer names are not unique
Should have the layer index appended.
. . . screen photos?
Not possible.
I am at work. My tablet has no discord.
Hmmm exitWith seems to also be an issue as it exits the entire script or at least a large part of it.
@full sonnet Do you want us to create a ticket with a collection of potential issues/improvements?
Last change to that SQF file was december 2021.
So I'd say any issues within it weren't fixed "recently"
You can create a ticket for KK 
https://feedback.bistudio.com/T184849
@scarlet root Please add your issues as well.
(SQF ones only)
(I lost will to live and it's sqf caused, does this issue qualify?)
that's mentioned in the EULA
I've seen lots of mods on the workshop running the https://community.bistudio.com/wiki/setObjectScale command on a loop so it works on units/vehicles and not just on simple objects/attached objects. Would it be posible to expand the command so a loop is no longer needed? We all know that Arma doesn't need any more loops running on the background π
I've also seen it being extensively used in missions on a loop too, for example to make a xenomorph be as big as a queen etc etc... or johnnyboy's giants script etc
no
otherwise it would have been made ^^ by yours truly @full sonnet
no
it isn't supposed to work at all, loop or just single run
Actually don't know why it even works, that's a mistake
The wiki on that is correct.
Only for attached and simple objects, and inside eden editor.
Any other case is not supported and may stop working any time
|| break it π ||
it works like since god knows when
you can do funny stuff with it, but it's buggy as hell
this is as small as i was able to make my character when testing it. smaller than that and it bugs out completely and you become invisible. still, even this is already causing issues when switching terrain surface textures
Ok, so a request turned into a bug being found π¦
And now it is my fault that lots of mods, missions and scripts that use the command will be broken π’
should write your name into the changelog
I am already "beloved" by the community so nothing much to loose there, I am also already taking the blame for Dredmen not fixing the rocks due to me being so intense about it
He also dropped Dx 12 because I asked, maybe I should ask for extra optimization so devs also drop that and I drag all of you with me
i literally just made that mod to prove a point to dedmen
its not meant to work even slightly
(the point being that it is somewhat doable)
Oh well, it was fun while it lasted... maybe it should be made to work officially but that probably aint happening just because it was me who asked
me and dedmen spent a solid period of last year just going:
kjw: can i do X
dedmen: no, not possible
kjw: i did X
dedmen: 
likely wont get "fixed" for BWC (and dedmen would feel sad about breaking a kjw mod ofc)
So many issues being overlooked, just add this to the list of "won't fix" too and be done with it π
Nothing more to see here gents, move on now
PS: Can you mod DirX-12 just to prove him it is posible π
capital ships?
that said, many things are possible, but that doesn't mean it's a good thing to do
True, lots of stuff on the workshop took a lot of work but just look silly in the end... Like far future stuff having great models but then A3 vanilla looking maps taking you completely out of the immersion. There is also the weapons with gamey GUI based sights or iron sights that take up 80% of the screen, both not fit for a realistic game engine like arma's or the melee aliens that just teleport around and look glitchy rather than immersive or scary
ill dm
No I already knew about it being bugged
I doubt that. But you are a doubting person so I guess that's fine
Dedmen: not possible
Also dedmen: which commands do you need to make it work?
(Dedmen loves us, but secretly
)
You said it wasn't π
Show proof 
Never
gonna test and if it doesn't work, i blame dedmen
lod loading works again. good job, nobody needs blame
do not post links without description #rules π€
@full sonnet I think the model loading "fix" has broken something. The nearRoads command doesn't work during pre-init anymore and returns an empty array 
When I put a breakpoint before it and wait a second before continuing, it does work so it's because of the async loading stuff
Preload was split into steps and every frame now only does one step.
So probably after the 5th frame it'll work
Yes I'll fix it. It should preload everything on mission start
out of curiosity, why was it split into steps?
what's weird is that nearestTerrainObjects does work but the road objects don't have the "network" type yet so I guess road network is built in another step?
99% of the time, nothing is preloaded.
It just spends time going through everything, to check if they might maybe want to preload something. Which takes alot of time basically doing nothing.
what about PreloadTextures and PreloadConfig definitions?
completely unrelated
Oh no some script broke and is now producing "Not a Number" values, what can we do.
SQF Debugger saves the day. Just set a breakpoint, and add a condition to only trigger on NaN.
And tadaa.
It feels so under utilized.. The VS Code extension wasn't updated/improved for 4 years (and yet it still works fine). And there is still so much to improve on it.
I want to invest some time on making it better, but is it really worth it if seemingly no-one is using it? π
is it really worth it if seemingly no-one is using it? π
Buy some Google ads for common errors that people are likely to googleπ¬
Didn't even know that existed
π’
Yes
This stuff tends not to work out of the box, and unless you're the author you have little chance of fixing it. So everyone gives up and goes back to printf debugging.
You know "NaN" isn't a documented result for typeName, right :P
shush, you!
done π
what can we do.
look into the RPT context and see where it's nan in the stacktrace?
IDK, every time I've tried to use the debugger for SQF I wound it faster to just check the rpt and usually I can immediately find the issue in my scripts.
I tried to use it twice and it wasn't overly helpful compared to just checking the RPT and inspecting the code
I used it once and it was super helpful 
I use the SQF linter extension and that by itself prevents a bunch of errors π
Best thing to do with debuggers is step through your apparently working code through and check that it's actually doing what you think it's doing.
I do kinda miss that.
Doesn't matter much unless your logic is pretty complex though.
And variable states!
I can see where the nan is. But its hard to track down where it came from
In this case, it came from an array.
And it was set into the array inside another function, where it was pulled from another array, that came from another function, where the value was calculated by combining 6 numbers.
One of the numbers was calculated by a loop that keeps adding a value to it. On one iteration of the loop, a sqrt call had a negative input value making it return nan, that was then added onto the value, making it nan too, which was then combined into the _loss, making it nan too, which was then returned by the function, which was then put into an array, pulled out of the array again, put into another array, which then later was picked up elsewhere, iterated over, and then when trying to get the value out again the NaN was found.
Callstack doesn't get me far on that. It tells me there was a NaN in the array.
I would then have to edit the code with filepatching, or repack the pbo and add tons of log tracing into every place to check to try to find where it came from.
or... I can just do one click in the debugger to set a breakpoint instead.
And debugging it, the function gets called every second or so. And 90% of runs its absolutely fine.
But, setting a breakpoint, and setting a condition to only trigger when NaN is present, lets me look at all variables and track it backwards to see where it came from.
It was still quite limited though, its time to add some features.
Like editing the values of variable, while paused
i do now
That's sick. Definitely going to give it a try again.
I tried it years ago, did not get how to use and fallback to diag_log debug.
Biggest issue was to set up prefix.
I made a tutorial but meh.
Next version will show every script function, no matter what mod it is and you can just debug it directly without needing local files.
That will be way more useful, especially if it can trigger on "exceptions"
I've finally learned how to do VS Code plugins. But now have to wait on the maintainer who wasn't active for 4 years, merging pull requests.
Or I bite the bullet and make a fork and publish that but I don't like having two things that do the same thing but one just being old
is that tutorial still around
welp i suck at something, il try again when i awaken
@ KK, just wanted to follow up on ArtilleryShellFired, and if there's any possibility of getting a fire mission counter/ID with it for better tracking what the AI is doing?
failed attempt at pinging 
No, intentionally avoiding it π it isn't like, critical or anything
it seems a3 tools on dev-branch have issues with mikerotools
specifically p3d files.
trying to pack my addons with it corrupted the files
after reverting from dev-branch tools back to stable, everything was working normal again
Can you repro issue with addon builder and binarize enabled? Mikero mostly just runs binarize for the p3d files
we did confirm the issue with running binarize directly via console
and it wasn't just certain p3d files, it was all of them that we tried. binarize would create some .deb file, then throw an error message and now the p3d is corrupted
What error message?
How do you judge "corrupted"?
Can you send me p3d and binarize command line to repro?
The only reference to .deb that I can find in our code, is diag_SQFCDebugDump command, that writes a deb file.
Can't find anything in binarize code that would do that
busy right now but i can send you something in the evening
but wait, i have screenshots of the error
ok i lied, dont have any of that error. will have to wait till evening
p3d file was smaller after that than before. like, a lot.
I shall wait for the repro then.
I tested the process on a couple models and didn't have issues, but maybe your model is just alot more complex to show stuff I didn't find
BTW was this fixed? Or do you need a ticket?
Nice. Thanks π
Do we get a dev branch update this week?
No Dev, but you do get a 2.18 RC Branch update right now. Just a few fixes and translations.
https://forums.bohemia.net/forums/topic/183855-release-candidate-branch-announcements/?do=findComment&comment=3533181
Special link context for @lucid totem, so he doesn't #rules shame me again π¨ββοΈ
Engine support for headgear models switching based on NVG being on or off (not used in vanilla data)
so no pilot helmet visor flip? π₯²
only nvg
i mean, vanilla pilot helmet does not use the feature
since it does have both models, visor up and down in the game files
but (not used in vanilla data)
:poutcat:
@unreal arrow sorry for ping but I'm assuming 2.18 is getting closer to release... For the ArtilleryShellFired mission EH, is there's any possibility of getting a fire mission counter/ID with it for better tracking what the AI is doing? It would be super helpful!
I briefly looked at it and there does not seem to be an engine param for it to return, so this means it needs some kind of implementation, so not for 2.18
Aww okay β thank you so much for checking!
π₯Ί
yeah, i have some issues right now
trying to make that repro but... now it works :x
and i dont understand why. we tried for hours that day until we finally realized that swapping between dev and non-dev tools fixes/breaks it
and it wasnt just on my pc, the same issue happened on two different machines
have you checked logs (pboproject, binarize, etc) for anything special to stand out?
looks all the same as usual
what i remember from the error we had was that the console printed some error about memshape
need to wait a bit until we can check it on the other machine again. but for now i am unable to reproduce the issue ..
the only "memShape" that exists, is in disabled code, and that would also never be printed
yeah, that error was printed by pboproject. i can't remember anymore what happened when we were running binarize via console
except that it broke off the compile and damaged the p3d
so... i guess we gotta wait until i can confirm it again π₯²
hmm, would the dev/beta a3t be suggested/recommended for general use with stable/perf/prof authoring?
dev build can only be play with other dev build
I also had the memshape issue the other day and had to import my p3ds into blender and reexport them to make it go away
What is the memshape issue. Please the actual message that you get
pretty sure it's the pboproject issue. updating to latest rev should work fine
also a reminder to always backup your files in case of corruption
With the new p3d encryption being a thing, is that something modders can use? Or is it just for internal use within BI?
"internal"
Check the wiki page linked on tools changelog
Thankyou! It says its opt in encryption for mods, and then links to a page saying its not documented for security reasons but will be shared with select creators via this discord. Is there a way to get on that list of select creators?
If you work on some big mod, where encryption would be relevant
Gotcha, I've got a mod with 40+ p3d's coming out in the next few weeks, but I take it you mean more so cdlc sized content. Nevertheless, thankyou for the info, I appreciate it!
You can DM me about it
hm. sometimes vehicles and items on the ground will "jump into the air" and fall down again. was this issue ever investigated?
i notice it quite often on the ws terrain, but also rarely on altis and more often on livonia
like this? https://www.youtube.com/watch?v=QPslL0SpD6Q
not sure if it's the same issue, but here an example with the car: https://www.youtube.com/watch?v=uabaxN77f40
happens with all vehicles and also weapons on the ground etc
cant see anything in .rpt log that could point to it
It could happen when its woken up from sleeping.
But it should never go to sleep while its active (engine running and driver inside)
this issue is quite old. noticed it the first time around when ws was released
I have only ever seen it when starting up the engine.
Never when just standing around like that
so while writing about this, i made a funny observation
it looks like this behavior stops if i alt+tab once and go back into the game
Noticed that only on Sefrou-Ramal world after some update of the game after WS releasing.
it happens a lot more on our terrain than others, but i have seen it happen everywhere by now
https://www.youtube.com/watch?v=K-EV9A2nkJ4 another example
Sanity check, is this in SP?
Yeah was just wondered if it was sort of desync
Confirm -- it happens in MP too.
and i remember QA giving us tickets with crapton of logs with something like "object x is under terrain, moving up" or something along the lines
but right now if i am looking at the logs, i see nothing
maybe server logs are different, dunno
That message is only for ThingX.
Not vehicles
ah, so that's probably with the weapons jumping then.
let's see if i can repro that right now
yeah weapon holders are ThingX
You can just play some time in Last Stand mission.
ok, cant repro right now. dropped a rifle next to the car and now not even the car wants to jump π
Does it happen right from the start or after some time?
After you get in the vehicle and get out does it stop?
nope, doesnt stop
sometimes it stops but not always
so my theory is that it might have something to do with the decal objects. however, i have seen this issue happen even in areas without any decals at all
in the areas with decals i can just reproduce it more easily
so if you start extraction in sp, get in a car and drive to the location from my video, then wait, you should be able to see the effect. hopefully. maybe.
at least i can reproduce it there very reliable.
Let me try
Doesn't happen for me. I tried a few times 
Maybe it's related to FPS? What is your FPS?
I only see Alchemist and Extraction
locked to 60
last stand is a mp coop scenario. but you would have to play it for a while and clutter up the ground with stuff. so not really fast to test, imo
but yeah, that's what i mean.. it's not always happening. often i am perfectly fine. but then suddenly... lots of issues out of nowhere.
I have locked to 58 FPS, and, as I wrote, for me it happens always on Sefrou-Ramal world. Didn't notice that for other worlds. Maybe I just wasn't paying attention.
And I don't remember it starting right after WS CDLC release, it started happening sometime after v1.1 release.
I couldn't repro it 
I've seen the issue a few times before but I thought maybe this was more reproducible
It's also commonly seen in MP when you are observing a vehicle from outside.
When you get in as driver/pilot, the jumping stops.
In this instance, fairly certain it's a localisation/ownership issue.
But I've never seen this issue in, for example, Takistan world from CUP.
We have had this issue as well in our Spearhead 1944 cdlc. It also happens in Singleplayer if I am not mistaking
so happy to see i'm not the only one π
Refering to the video lexx posted
Lexx, are you the driver of that car while observing it?
We also have had tons of glitches with weapons being constantly lifted and dropped back to terrain.
This has not always been present in the game.
If you observe it jumping while standing outside it, then get in, does it continue to jump?
yes, can be seen in the 2nd video
i drove the car to that location, got out, then back inside
Hmm, that looks a bit like bad car physx
It also happens with tracked vehicles for us.
Have you compared multiple vehicles, such as tanks from SPE, RHS, SOG, Vanilla to see if they all do it under the same conditions in the same spot? Definitely possible to make this happen with incorrectly configured physx on the vehicle.
I have not. Perhaps I find some time today.
Could this be an issue with worlds, not vehicles, weapon holders, etc?
and items, as written. it's not just related to vehicles.
I know
Because it also happens for Offroads in Sefrou-Ramal world.
considering the issue happens less on altis and more on our terrains, i'm pretty sure terrain settings play a role in the issue
maybe grid size, etc
Terrain cell size?
What cell size you use?
I mean..that could be related. @scarlet root do you know our cell size for Mortain and Normandy? Mortain was much smaller I think.
getTerrainInfo is a thing in case you guys didn't know π
Yeah. Not at home right now.
2 and 3.
I noticed this with some weapons that I hadn't fully fleshed the LODs out for. Eventually I'd come back and add them in and the issue would resolve itself. Unsure as to which LOD in specific, though.
it seems there is value display change with e values. from what i understand it should not change the value itself
TBD if any sqf command or tool is affected (probably not i'd assume)
Compiler upgrade changed float handling, including float to string formatting
Kinda reminds me of a common issue where a disabled tracked vehicle will continually flip between two facings. Not sure if it's the same thing though.
Figured that one was MP only though, don't remember seeing it in localhost.
(my bad - had old vs new flipped)
sorry, very late response but a3t is arma 3 tools not the game. using dev versions of those are compatible with stable game apparently
Does new 3d markers in 3den builds by engine somehow or is it available to investigate how it works in some external script ? Wondering is it possible to draw(show) markers these way outside 3den
It's in engine. Similar can be done with drawIcon3D command.
for icons - yes, but not for the areas. Like if you need to draw ellipse marker for players
Like the mission markers in GTA
The area markers are just a model, that is scaled up.
Problem is setObjectScale only has uniform scale, so you cant do the same scaling with it
i was hoping for this too π€¦ββοΈ
I didn't see a ticket about non working end on Zeus, am I blind or should I create it?
Ah shit
Fun
~~Has anyone else noticed that AI is broken on dev branch?!
~~
https://feedback.bistudio.com/T185257
Is dev any different currently from stable?
this is a test build that Dedmen sent me
I see
Yes
But not sure if it's broken on the actual dev. Let me try that one
I have to verify my game and all my mods too to get it back 
No not broken on the public dev
Ah I see the build Dedmen sent me probably contains his async AI stuff and it was probably incomplete
Access Denied: Restricted Maniphest Task π₯²
Yeah I hid it π
async ai stuff?
it's public again if you want to see it
The async AI visibility things
#perf_prof_branch message
for a moment i thought the infamous ai fix dlc is real!
so about the mentioned "objects are jumping"-bug from a few days ago. by now i actually really believe that it is fixed (in sp) if i just alt+tab out of the game and come back
Now all we need is a command that can be run on the init.sqf that simulates an alt+tab π
ok, now alt+tab did not fix the issue. tl;dr -- no idea
any TI change in last update ? Haven't seen in changelog but looks like something has changed
no
i thought this issue was fixed like a billion years ago π
what even is that ?
The RCO (HAMR) rifle optic. There's a gap between the glass and the inner face of the casing.
on 2.19.152232 I get a million lines in 5 minutes of
d:\Bis\Source\Dev\Futura\Es\Memory\memAlloc.hpp(588) : Assertion failed 'CheckSameThread(_owningThread)'
Yup, known and already fixed last week
is it pilot lod?
Hotfix 2.18 was released: https://twitter.com/ArmaPlatform/status/1849072803443601605
Love u all no homo
Is there a full build version number somewhere? Changelog just says 2.18.
152302
Thank you.
Fixed: Partially filled weapon magazines could disappear after reloading in multiplayer
π₯³
t/.co/Ho... 
Link is broken.
You briefly scared me I wrecked the link, but that's just Discord preview shenanigans?
Links inside discord embeds seem to be broken. At least for embedded twitter posts. But the actual link when viewed on twitter does work.
I blame Elon.
Yeah. Sorry.
That is why you have to rename the link from x.com to vxtwitter.com for video or fxtwitter.com for pictures/posts.
Or just post the link here directly π
or just dont post twitter links π
Or just don't use twitter
Or the internet
bruh
uv is misaligned. needs to be moved down, then it should be fixed
and the MH6 mirror is not a mirror because the helicopter's reflection is on the left
btw π
uv needs to be mirrored then
this bug is canon
it's on Miller
period
Bug is feature confirmed :p
Can someone remind me. Should direct connect in server browser try to connect or just show server in list? I have an issue with it. Nothing happened
Seems like bug. When I use button with script command it connects. When I use filter on Internet tab server shows up but takes time. Direct doesn't work.
The latter
Calling anyone with UI on texture experience; I'm banging my head against a wall here.
I'm coming back to an UI-to-Tex scope mod to finish it up and get it live.
I have everything working absolutely fine with no issues except that after packing it up into mod format I get: bin\config.bin/RscInGameUI.reapir_optic_ui
After clearing this error in game, again everything works like a charm.
Obviously I sought to fix this one remaining issue and nested my UI to Tex optics class into: class RscInGameUI but this is when all hell breaks loose.
All of a sudden, all my ui to tex elements load onto the 2d game interface and not into the physical 3d scope sufaces.
And now instead of getting the bin error, I get a new one saying i have poorly formatted render to texture name? my UI on Tex: #(rgb,1024,1024,1)ui("reapir_optic_ui","reapirui","ca")
I even tried loosing the 3rd array element for ca at the end to reflect the biki but no difference.
Does anyone know what ui optics class elements I can use for UI on tex that wont place all the UI elements into the 2D screen?
I've tried everything but can't get past this final issue before releasing π¦
I think the weirdest thing is how everything worked pefectly (except for the missing config bin error) until I nest my optic config class into RscInGameUi.
I'm more than happy to be proven doing something stupid here and fully expect that to be the case but I'm at a loss
config.cpp paste bin: https://pastebin.com/KWDF3P0T
We use UIonTex for our MA37 ammo counter. I can show you some of the config for how we have that set up if it would help?
Not really dev_rc_branch related anymore, however, so we should probably discuss further in #arma3_config
Clearly the main issue is that the display is in RscInGameUI, however
Your display class should just go directly in the root config
The issue is very likely your weaponInfoType = "reapir_optic_ui"; on line 77
just make it weaponInfoType = "RscWeaponZeroing";
weaponInfoType is used for drawing 2D elements on the screen when zoomed into the 2D ADS mode iirc, and is a totally different functionality than what you're using it for
No dev again this week, because the performance stuff failed QA π«
Thank you very much for the response, and agreed; I'll continue discussion in the other thread π€
That only means you got more time to refine things π, keep up the good work πͺ
@full sonnet About doing explosion intersections MT:
Do you MT each explosion or do you collect all explosions and do their intersections together? (I don't think the former is really worth it)
There usually is only one explosion, not multiple
And the stuff needs to be done when the explosion happens, I cannot wait to see if maybe some other explosion might happen soon
Some of changes were released in 2.18 hotfix. Or at least mentioned as released
That is correct, the previous dev-branch was before the hotfix.
hey, would it be possible to look into issues with usb microphones not being picked up properly by the game at all?
basically improving the robustness against OS changes, is wasapi/wdm/mme/dx used here or its own conversion? thoughts about asio support? XD the last one is a meme
Added: New marker shapes: TRIANGLE, PENTAGON, HEXAGON, HEPTAGON, OCTAGON, NONAGON, DECAGON
Can we expect extending Cfg3DEN β Attributes β ShapeMarker with new shapes in 2.20? This is Eden Editor marker 'style' attribute
If KK builds us a new attribute control. π
changing already existed not allowed?
localization request is made, but not a priority atm
Okey progress on Webbrowser.
I found a way to load a html file without needing to ask the user for approval.
So you can embed .html files inside your mission/mod pbo. And display them in a webbrowser.
Including javascript and css (must be embedded inside the html file, no separate external files. Though you could load extra files via SQF and javascript, as far as SQF is able to)
Javascript, watching the situation from afar
Inb4 React For Arma π
And JS UI modding, in general
The annoying thing is just having to stuff it into a single file.
But there are webpack solutions already out there that do that right? π
Images only possible via data url's.
But theoretically you can ask SQF to give you a data url for some other file, that you can then send back via javascript and embed somewhere.
I will probably prepare some code for that, I need to test it anyway
You still use Steam right?
yea
Wouldn't it be better to include CEF or sth in a3 itself?
Yes. But won't do that.
I will add a way to turn a image out of a pbo into a data url.
So you can import any paa file, into the javascript.
That way @trim crane can rebuild his killfeed again, this time as a single html page instead of hundreds of ui2tex π
Rust WASM Arma UI 
If someone wants to try that, They shall do that before 2.20 releases and they find out some small thing is missing π
But I don't know the security profile of wasm, if it can launch external web requests even if the javascript its called from isn't allowed to do so, that'd be a problem. But I would assume wasm devs considered that.
nyx tank has red tracers (blu) instead of yellow (ind). plz somone fix plox
this is unacceptable
So half year available

//_image = controlNull ctrlWebBrowserAction ["GetTextureAsDataURL", "A3\ui_f\data\map\markers\flags\Netherlands_ca.paa", 1024];
//_image = controlNull ctrlWebBrowserAction ["GetTextureAsDataURL", "A3\ui_f\data\GUI\Cfg\LoadingScreens\GC_wallpaper_10_co.paa", 1024];
_image = controlNull ctrlWebBrowserAction ["GetTextureAsDataURL", "A3\ui_f\data\IGUI\RscTitles\SplashArma3\arma3_lite_splash_ca.paa", 1024];
_ctrl ctrlWebBrowserAction ["OpenDataAsURL", format['
<body style="
background: repeating-conic-gradient(#FF1493 0%% 25%%, transparent 0%% 50%%) 50%% / 64px 64px;
overflow:hidden
">
<img src="%1" style="width: 2040px;"></img>
</body>
', _image]];
assembling html in SQF
Arma Alpha Lite confirmed
doesn't this make ripping textures from the game very easy? 
Wasn't it always? Like realistically
Sure it's always been possible. But with this you won't even need a tool to extract the full, high-quality texture, hence why I said very easy 
ofc textures are not very important but still...
meh, who cares is my opinion on that π
if we could do complex interface stuff in html now, with button events (running a script when clicking a button on the html page) etc. i accept everything
this makes interface creation a million times easier. this whole rsc config stuff is just a huge pain in the butt
Arma life internet let's go(even if i don't play a3life)
Play Arma add Fronted Dev line as experience in CV.
You mean.. like unpacking the pbo using the official tools and double clicking on the file to open it in the official tools and using the official option to save it as png?
Or displaying the texture full size and taking a screenshot?
Probably should block ebo's, I'll think about that.
But why would you go to all the effort to get it into javascript and get it back out to then open it in browser and save it, if you could just take a screenshot of it.
It doesn't get you the full quality, its webp compressed. It'll be very close, about as close as a screenshot.
You can easily script to display the texture you want to rip 1:1 on the screen
Fair enough π
Amazing dev branch update notes! Best game devs ever!
I have a quick question about one of them:
Added: Asynchronous AI and sound simulation
Does this add any new game audio controls?
No
will break OS compatibility and that's unwanted, and just add unnecessary dependencies
This is a performance optimization measure
excuse me, do you know what does asynchronous mean?
it means things can happen independent of each other
It's not unnecessary dependency. For reliability you have to have a specific version of a library
So let's say in a few years steam suddenly decides to remove its browser, or make a change that breaks stuff
Then this feature is broken. And if by that time the game is no longer maintained, every mod/mission that used it will be broken too
and if the game is no longer maintained there might be an exploit in CEF so malicious mods can get desktop access 
though you're far more likely to just use a regular native extension if you want do naughty things
No this is why I asked π€
Asynchrony is the state of not being in synchronization.
Asynchrony or asynchronous may refer to:
We also have extension UI.
So anyone can implement CEF by themselves in an extension, and switch to Extension UI to accomplish the same thing.
If Intercept stays working (which it will), it can do that and handle all the script commands and eventhandlers too.
If need be, it'll be fixable with a mod.
JS can request files from engine, without involving SQF.
With that you can also implement loading Javascript libraries.
function LoadLibrary(filePath)
{
eval(await A3API.RequestPreprocessedFile(filePath));
}
LoadLibrary("z\\myMod\\addons\\lib\\FancyChartsLibrary.js");
Can even run preprocessor, on Javascript files.
inb4 building missions in JS now instead of SQF
Webbrowser is only available when you have a UI, so you couldn't use it on server side
is this secure yet?
||Netherlands jumpscare||
It just uses loadFile/preprocessFile for the text files.
I didn't change the texture loading stuff yet. Don't know if I will, probably not.
hmmmmmm
will dev console be available to everyone?
Yes. Probably only when running game with -debug (Mainly to protect normal players, from having it open mid game and tabbing them out)
RscTitles UI runs in fullscreen, and overlays a webbrowser as a HUD element (checkerboard pattern to visualize it)
Script eventhandler gets path to the weapon texture, and the names of killer/killed and runs javascript
https://github.com/dedmen/ArmaWebBrowserTests/blob/master/ArmaMission/init.sqf#L18
https://github.com/dedmen/ArmaWebBrowserTests/blob/master/src/KillTracker.ts#L13A
Typescript receives that and asks the Arma API to give it the contents of the texture, texture fetch happens async (loading texture file could take a while) and when the contents have arrived (probably a frame later) it adds a new row that CSS animations pops in https://github.com/dedmen/ArmaWebBrowserTests/blob/master/src/KillTracker.css#L2
blerp
Killing one kills multiple?
Waiting for such a killfeed now:
https://x.com/zack_overflow/status/1853494692169896361?s=33
π
why not -console ?
Because I'm not adding another start parameter for this.
That's probably the best fidelity I can come up with
it's missing MLG sounds and vuvuzelas on multikills
Where's the Halo guy voice going KILTACULAR
what are the chances, in 0..100% range, of seeing setPlayable in a working state?
Next update adds omnimovement
Been asking for it for a long time
I'd make a proper addLobbySlot command instead or something
With an ability to specify role name, squad, etc.
ticket explaining usecase and expected+current behavior?
@verbal venture
addLobbySlot explains it
Don't u have a ticket
Paste the old one
Adding a slot into lobby is something that I need a bit. Project I play on deletes all AIs and respawn is turned off, so the empty slots from lobby disappear, meaning that if you die (slot gone) and your game crashes, you are not able to get back in and spectate.
I want to create a spectator slot on mission start and if it get's occupied, create a new one. Allowing people to always at least spectate
https://feedback.bistudio.com/T66277 but it does not have a usecase, just mentions "no worky"
Will it really be possible that a player (slot) can be created by a script command?) Hallelujah
I don't see why not, currently.
But it requires further investigation that will only be done with ticket
ok
Do you really want a new one, or should I just add the usecase as a comment on the one I linked?
some time ago ive tryed to modify some stuff on slot selection but didnt found any leads in the wiki, its posible?
reuse the old one, if its in comments then its fine
yea! the @groupname thing is from CBA now you mention it
Added.
tank you!
setPlayable and unsetPlayable are both not possible
#arma3_scripting message
because slots are locked as soon as the session starts, or something like that
Do we get dev update today?
Nope
Webbrowser can load local files in mods.
Would you expect
webBrowser ctrlSetURL "z\mymod\addons\index.html"
or rather something like
webBrowser ctrlSetURL "file://z/mymod/addons/index.html"
The first is easier, but it being a "URL" maybe it should stay a URL and obey the syntax
Second imo
I prefer it too, but will people be confused/dislike, mainly that its forward slashes wheras everything else in Arma filesystem is backslash.
Though it works the same way on windows too
it's seturl not setpath so I'd go for 2nd too
π That it'll be then.
People will have an alternative if they want to
_ctrl ctrlSetURL "file://z/mod/file.html" is equivalent to
_ctrl ctrlWebBrowserAction ["LoadFile", "z\mod\file.html"] is also equivalent to
_ctrl ctrlWebBrowserAction ["OpenDataAsURL", loadFile "z\mod\file.html"] is also equivalent to
_ctrl ctrlSetURL ("data:text/html;base64," + controlNull ctrlWebBrowserAction ["ToBase64", loadFile "z\mod\file.html"])
Duplicate API to do the same thing is probably not good, maybe I'll get rid of that
Adding the protocol only seems to make sense if other protocols are supported
data: protocol is supported.
And if browser is set to allow external URLs then http/https is too
I found one more example above π Though thats a bit wild
https://community.bistudio.com/wiki/CT_WEBBROWSER#Local_Content is available today on #perf_prof_branch
But with two small tweaks. file:// doesn't work, and javascript api has a typo on prof.
So if anyone wants to play around with it, its ready to use.
I also have a sample project with webpack+typescript+custom font and css https://github.com/dedmen/ArmaWebBrowserTests/tree/4499d2b0354815eede928e2fc425b4c3ef421d26 (Don't use newer commits, they'll require newer game version than available)
Development on this is probably going to go on for a bit, I'll need feedback and people to tell me things that I've missed.
Wait is this the doom in arma
Will it work with getMissionPath?
Yes, but you'd have to convert it into file:// with forward slashes, if you want to use the SetURL variant.
The file finding/loading works same as loadFile.
So loadFile "index.html" or "file://index.html" would both refer to file inside mission
That was external URL
does the browser communicate with the game via alerts?
π
yikes, I guess that's the only thing you can hook into? :D
yeah I got alert and confirm, and thats it
Engine reserves alerts starting with "__A3"
I guess not a big issue if it will be packaged into a lib
The A3API is already injected as a "library" into every page, but that communicates with engine directly, not with SQF.
For things to SQF, they need to be done manually with the JSDialog eventhandler.
I'll add an example for that to the wiki example
I have no idea what I could do with this, but looking at the code, seems fine
Added example of how to communicate with sqf.
_ctrl ctrlAddEventHandler ["JSDialog", {
params ["_control", "_isConfirmDialog", "_message"];
// Insert message as text into the page, by assembling Javascript code to insert a new element and set its text.
_control ctrlWebBrowserAction ["ExecJS", format ["const eSpan = document.createElement('span'); eSpan.textContent = 'Script says %1!'; document.body.appendChild(eSpan);", _message]];
true; // We need to tell it that we handled the "dialog", by returning true or false.
}];
``` ```js
// Send a message to script
A3API.SendAlert("Javascript is sending an alert"); // This alert will trigger a JSDialog (Alert Dialog) eventhandler
End result:
<span>Script says Javascript is sending an alert!</span>
If you wanted to communicate with SQF, you'd probably use JSON.
A3API.SendAlert(JSON.stringify({document: "test"}));``` ```sqf
_ctrl ctrlAddEventHandler ["JSDialog", {
params ["_control", "_isConfirmDialog", "_message"];
fromJSON _message; // hashmap [["document", "test"]]
fromJSON _message get "document"; // "test"
true; // We need to tell it that we handled the "dialog", by returning true or false.
}];
inb4 cougars in my area - I know they're fake because they are not endemic to Altis
unless you have Reaction Forces!
Dedmen, while you're working on perf improvements, can you also plz check why loading a game from save takes a really long time?
For instance, run any showcase mission (such as Infantry), save, then load
I can't find any good explanation why such a simple mission should take so long to load. I think something is broken
For some reason, when loading from save, it always takes 15 seconds to fill up the mission loading bar (the one after loading the terrain). But when I run a scenario for the first time it's almost instant
Happens on both perf and dev branches btw (probably stable too)
what increases savegame time (storing and loading) a lot too is object states - ie many hideObjects or damage states.
possibly due to storage structure seemingly rather expanded.
repro is: SPE, SP, PZKW, start mission, save-load game
Yeah but the Infantry mission is almost as simple as a few units running from waypoint to waypoint π
This even happens if you just save as soon as you load into a mission, so there shouldn't even be any state changes
Also that 15 second delay was consistent in every mission I tried. So something is off
Hence the "too" - it might be a general design topic of the data structure, next to other type of issues with rather the engine handling.
Curious if the system is still basically the same as it was made for OFP, or redone at times, or just duct tape revisions whenever necessary.
That said i think to remember in OFP save-load was very fast (edit: at least with somewhat modern hardware its ~1s for save/load)
Let me test A2 real quick π
Ok I had A3 2.15 on my laptop actually
Looks like it's a new issue
On my laptop (which is much weaker than my PC) it took 1s to load
Ok no it's not a new issue. It just takes longer on my PC. Weird 
I wonder if it has something to do with having too many cores
Update: using -cpuCount=4 didn't help so I guess nope
did a quick test with A1 (god was that terrible in so many ways) and A2 (only getting 20-30 fps with high settings - and ppl complain about A3 optimization π )
A1: saving 2-3s, loading ~1s
A2: saving 3-5s, loading ~2-3s
sameish wiht A2:OA (but even worse fps.. 18 vs 21 fps in benchmark 2 πΈ )
A3 infantry showcase takes me ~2s for save and ~3-5 for loading
Yeah something's wrong with my PC. Debugging
Have you tried turning it off and on again?π
It's been bugging me for a few weeks now π
So the issue is related to WaitForMultipleObjects call which is a kernel call
For some reason it's spinning too long 
loading a savegame does feel awful slow
If someone else with Intel 13th gen CPU (or maybe 14th too) and Win10 could test this I'd appreciate it
I guess it's win 10's fault
Do you meet any of the above conditions?! π
I mean what is your CPU and Win? π
Is it also 13th/14th gen and Win 10?
win11, 12th gen intel core i9-12900k
i didnt stop time, but feels like loading infantry showcase right after start takes like 10 s
ok i just stopped the time. 16,81s
saved right at the beginning and reloaded
Yeah it's really long.
Here's a vid if someone else wants to see/try for themselves
(notice how fast the restart is, and how slow loading from save is)
13700, Win 10
yup, pretty much looks the same for me
10-ish seconds on Win10 with 5800x3d, 16-ish in the second run
6-7-ish to resume from "Showcases" menu
Load time: 13.2s
Restart: a few secs
AMD Ryzen 7 5800X, Win 11
A mini-update for 2.18 was released (future EF CDLC support): https://twitter.com/ArmaPlatform/status/1858857358589346128 π£ββοΈ
You could send me a vTune recording of the load and I can look at it?
Tried infantry showcase.
At the start in the forest, save game (about 2 seconds), load that save (about 3 seconds)
ok so the issue is related to terrain quality and view distance.
Ultra quality: issue happens at around 3300+ view distance
Very high: tried up to 8000 (which is as large as Stratis itself) and it took 3-4 seconds so still much faster than ultra
