#dev_rc_branch
1 messages ยท Page 6 of 1
What a voodoo damage system
yes
Thanks God I made my own
Markers are bunged up in dev-branch.
Setting marker properties could set them on a random other marker (only if you are using setMarkerDrawPriority)
"Tweaked: View geometry of a few more structures (House_1W03_F, House_2W03_F, House_2W04_F, House_2W05_F, IndustrialShed_01_F and u_House_Big_01_V1_F) "
This one is also slightly broken.
Land_House_2W02_F
See message here #arma3_feedback_tracker message
Loosely connected. There are a couple of Enoch ruins (the standalone grey ones) that are missing damage = no named property, so if they take enough damage they just poof out of existence. (Not even any effects).
Expected results. Latest Feature 
Fixed: Double echo sound when firing Katiba or SPAR-16 indoors
Please don't forget fix the silenced Mk200 as well, see ticket comment https://feedback.bistudio.com/T178139#2589058
If you're changing view/fire geometries then that House_Big_01 also needs to be fixed so that grenades can be thrown through the window. https://feedback.bistudio.com/T84758 9 year old FT ๐ญ
The model shown in the video is u_house_big_01_v1_f.p3d, and the same issue also exists with u_shop_01_v1_f.p3d. Both of these models are missing openings for the windows in their geometry LOD.
I think @unreal arrow is working on different solution, rather than making hole in geoemtry
Okay
But can we have the hole ๐
Or do you mean different solution with same effect?
found a age old texture loading bug while looking into ui2texture issues.
When a texture is too large, and doesn't have enough mipmaps, it is possible that it simply doesn't show up.
Because the game will calculate that it wants a mipmap, that is lower resolution than the lowest available.
Which it considers "the texture is so small, that the player cannot see it anyways"
But.. This texture isn't small, I can clearly see it. Or rather I could if it loaded it.
This only happens with procedural textures with limited number of mips. Because packed paa textures always generate enough mips.
But it could also happen with MipMapTool modified paa's that remove the lower mipmaps.
At the distance in the screenshot, and per resolution set in the procedural texture and at lowest texture quality, the game tries to load these mips (index starting at 1)
64: 3 (results in 16x16 texture)
128: 4 (results in 16x16 texture)
256: 5 (results in 16x16 texture)
512: 6 (results in 16x16 texture)
1024: 7 (results in 16x16 texture)
So it really really wants to load a 16x16 texture.
Now if you make a 1024x1024 procedural texture (perlinNoise, or ui2tex, or r2t) with less than 7 mipmaps. Thus a 16x16 version doesn't exist.
The game will simply not load anything.
In my example there is a 1024x1024 texture with 5 mips. As you can see you can't see anything.
You certainly wouldn't expect this size of texture on the screen to be "too small to see, thus don't need to load it"
This was fun to track down...
This caused me alot of brainmelt back when working on ui2tex that only allowed one mip. They would just not load.
that might explain a bug I saw with procedural textures not rendering at a distance (like text, license plates etc) even when theyre part of the lod
I had some very ugly workarounds for text/ui procedurals, that mostly worked and sometimes was absolutely broken.
But I finally figured out the real issue. It'll be fixed in next #perf_prof_branch, so maybe you can see if broken stuff works after that
if its about num plate technology, then this has some hardcoded drawing limit which I don't remember atm
no ๐
As for different solution, lets if KK wants to do some sneak peek
i'm really curious what this different solution would be
Beaming the grenades of course.
inb4 clip through houses
it was solved in ofp already, the grenade teleported to the bottom center of the house whenever you throw one at it
No, my solution wont fix that, there is some artefact in the model that makes half of the window pass through and half not
Tried it now, depending on at what part of the opening you throw grenade, it may pass through or may bounce.
As for different solution, lets if KK wants to do some sneak peek
Next dev

I'm more curious how lowres the tex is
funny, just today i was wondering if i can script that without dying of pain
Possible yes, not dying, no
I can reuse the code from triggers for area markers.
But I wonder if that wouldn't be annoying if they'd always show up like that
3den toggle button?
uff adding buttons is big uff
That'd b e so much more effort
maybe I can just make it be low height and only outer edges
The main problem with area markers is that you never know where they start and end. I wrote a script to visualize them with triggers temporarily
I think I'll just take the trigger box. But without top/bottom on the box.
And then make it limited height and let it fade out.. somehow.. maybe.
Yeah it do be annoying to see them :/ mh
This though should be okey? I think that would be un-annoying enough to just leave it
Can also make it fade out a bit harder when unselected

And so unexpectedly, this also works. (if you can remember the keybind)
But moving in 3D doesn't work yet 
hah fixing the orientation makes it look like its painted onto the ground
that's kinda cool
it's flat painted on the 2d map after all, might as well be flat painted on the 3d surface!
it is 3D still. You can even set the height now
I never noticed how hard it is to see triggers on non-VR maps.
The 3D model for trigger and markers can already be controlled by set3DENModelsVisible script
Yeah. The colours are sadly semi hardcoded in config. Would have been nice to have a command for that
never had a problem with seeing trigger areas
but i always thought it would be nice to be able to color triggers
Could this be added to all modules with areas? Having this for minefield module, would be great
devQuality? so developers get higher quality Arma3 ๐ง
There are many modules that were never updated after Eden Editor was added. I think I have a ticket for that somewhere.
You know my tickets better than I do.
don't know if there is a way to find out if a module has an area
Module areas are already drawn in 3D.
I'm good friends with the search box :P
hmm looks like on dev-branch it is not possible to disable autonomous mode on drones
allow lights cant be toggled either
i just reverted to stable to check and there everything works as expected
Do modules even have areas themselves or do you sync an area (trigger) with them? 
Sorry I meant to reply to this comment
There are modules that have built-in areas, e.g. covermap, civ presence
They are custom attributes though iirc.
They would need size2 or size3 and canSetArea property to properly function.
No dev this week
Karel found the answer to the marker zone being annoying issue.
Simply only show the zone, when the marker is selected.
sometimes it can be so simple
I think I'll keep the icon rotating.
On map screen, the icons already rotated (so unit icon pointed forward where the unit was looking)
in 3D view the icons always simply pointed upwards on screen.
The icons rotating is kinda nice for the vehicles itself.
But then looks ugh for the passenger icons
yeah, that's not so great
Ah I was just doing it wrong.
I can just turn off passengers.
I'll try to make it good
yeah there we go. Just need to fix that floaty group indicator icon being wonkey ๐
Sweet
If am wondering. If markers now are only drawn when selected/hover, would it make sense to make their height unlimited?
oh shoot.
Scaling is bugged in 3D 
Never rotate a marker in 3D around z or x
yes. But I don't like it :harold:
Also, what's Unterholz ? ๐
Anyway, back to the topic
Another thing is, you can now change the height of the marker, but the visualisation area doesn't start at ATL 0 then
It rotates a bit wrong, but besides that doesn't seem too bad?
Basically a floating area
Once it's rotated, try to scale it ๐
I don't want to deal with translation issues ๐
uwaaahhh. I see.
I don't see that.
I just see that scaling on X, works the wrong way around, dragging it bigger makes it smaller
ah i got it
So basically same logic needs to apply to it as with triggers
Rotating around x and z does nothing
Mh.
Yeah could force it to ground level. But if its on a mountain side you would pull it upwards.
But yeah we can make it infinite height now anyways when its only selected
Good news, snapping to surface works and fixes wrong rotation ๐
Also, should the z translation axis do anything? Markers only have size2 attribute, they should always be ground level.
get3DENSelected "marker" # 0 get3DENAttribute "position" // [[3432,2940,87]]
Now one can do things like this
Not sure if that was possible before
Scripted markers actually have a 3D position, and its also stored as a 3D position
Yeah, just noticed, default marker height in Eden Editor seems 5
thats just the surface of VR terrain
Doesn't return 5 for any other entity
It was not possible to do in Eden directly, without hand-editing the mission.sqm.
Markers were always stored as 3D in mission.sqm, and also created as 3D at game start
Alright, cool then. I think I've tested everything I wanted
I didn't have to change anything, the only thing is that you can now edit the height in the 3D view.
But that was just because UI was missing before, only the UI is new
Now I wonder what happens if I replace size2 attribute with size3 so one can also set height in attributes UI
Switching between local and global coordinate system works too?
yes, the rotation being applied you mean
So
- Zone only when selected โ
- No Rotation around XY โ
- Full height zone โ
- Icons rotations โ
I think that was all?
I think enabling the Z position coordinate in the Transformation attributes section would be sensible, if it's not a lot of trouble to do that (now that the same thing can be done in the 3D view)
Mh.
It would be a config change which is hard. And it really only is useful for power users.
95% of users won't deal with marker height
I'll Mod that if it doesn't break anythinf
Should be all. Much appreciated
It already is 3D internally, so shouldn't matter yeah. I think its better done on your side
One thing maybe now is, placing markers forces you onto map screen.
That isn't really needed anymore now that you can place in 3D.
But, in general map markers are for the map, so it still makes sense to place them on the map
I'd be OK with removing that tbh, mostly it's just annoying when I accidentally click on Markers instead of Modules
In a perfect world we'd have a setting for that. But that requires config
Yeah
oh wow. Was this always bugged?
Clicking on the "Markers" tab, forces map.
But you can then close the map, you are still on the marker tab. You can now click on a marker to place it but you don't see anything, it just ignores the placing if map isn't open
Always has been
And if you open the map again, the marker placement is still queued on the cursor :D
Trying to find the marker placement code.
The marker preview on the map, doesn't call the marker preview code, nor the marker icon rendering code.
The marker placement has custom icon drawing code.
Which is why the size/shadow doesn't actually match up between preview and placed marker
y tho
Ah unless you're placing a Rectangle or Ellipse, where it actually does use the drawMarker functionality.
There is specifically a "allow marker placing" check that only allows markers when map is open.
Fun fact, when you place a unit, then allowMarkerPlacing is enabled, because.. yeah.
I'll turn that off. So you can place markers with map closed, but the map will still auto open when you click on the markers tab
That makes the faked custom icon drawing even more jarring

lol.
Ah unless you're placing a Rectangle or Ellipse, where it actually does use the drawMarker functionality.
Yeah, and because Rectangle and Ellipse don't have any icon. It does all the effort to set up a marker and draw it to... draw nothing.
The scaling there is F'ed up for Area markers too, but you can't see it because they are invisible ๐
Add one thing, break 3 other.
But I think I got it now
Tweaked: Eden object icons now rotate with the object
Added: Eden markers can now be placed in 3D (non-map) view
Fixed: Eden marker placement preview icon didn't exactly match the marker (slight size difference and missing shadow)
Fixed: Eden 3D marker could be rotated in pitch/roll axis (should only be yaw)
Tweaked: Show 3D marker area only when marker is selected
Tweaked: 3D marker area is now "infinite" height (previously was limited because it would be too annoying if it always showed, which it now doesn't)
๐ค
They'll all be in next prof, don't know when
@full sonnet Any easy way to expose 3DEN's line drawing into scripting?
I know these are 2d unlike say DrawLine3D which are actual geometry going through objects, but this will still be very useful even if it will be drawn over everything.
Maybe even call it drawLine2D
no
I think that thing is a 3D model that is just scaled
with simple object + set object material + setObjectScale in 3D that could be doable
Meant these
but I opted against 3D scale, but 3D for local only objects might be viable
Yeah, I think thats a model thats scaled
Atleast when I initialize it, I give it models
Maybe not, can't look now maybe friday
They look like the same kind of line as is used for the translation widget handles, and I'm pretty sure those are UI and not models (appear in front of terrain, not affected by lighting, no perspective, don't seem to be affected by post process effects)
3D Position works very nicely
I have one last wish. Would it be possible to show markes both in 2D and 3D when they are hover on/selected, even if their alpha is set to 0?
Currently it's super annoying to work with markers that are set to be invisible.
The code must be already there, but I guess it's something like (1.5 * currAlpha). A marker with 0.2 % alpha will be better visible when hovered over/selected, but if its alpha is 0 it will still be invisible.
This would completely break Antistasi :P
How?
Because we have an awful lot of logic-related markers that are alpha 0 and need to be untouchable.
ah, fine
That hover logic for markers is just borked, instead of increase alpha by a % of the alpha set in editor it should just change it to a fixed value, preferable 100%
Sadly I cannot control that via script without changing the original attribute value. Which is ...not ideal.
Haven't looked into the mission but I wonder what's the main purpose? Positions, inArea checks instead of saving into variables?
In most cases it's not even a good method these days, but it's how it was written.
What would you pick these days? Trigger entities as they support Z?
I did consider switching to locations to some stuff, because nearestLocations is a lot better than what you get for markers.
but there are workarounds.
r3vo is right, invisible markers are a huge pain to work with
its a config entry for "iconAlpha"
Normal is 50%, selected/hovered is 100%.
It works the same for all icons.
So there is no 1.5*, the most is 100%.
So invisible markers still be invisible.
I can make a
if (isSelected) then makerAlpha = max(0.5, markerAlpha)
so if selected its bumped up to 50%, no matter what its set to
why not always 100%?

configfile >> "Cfg3DEN" >> "Marker" >> "Draw" >> "2D" >> "alphaHover" //1
configfile >> "Cfg3DEN" >> "Marker" >> "Draw" >> "2D" >> "alphaNormal" //0.5
What do these do?
inb4 "literally nothing"
Looks like it. But from the naming of the properties they are supposed to do exactly what I suggested
What I said
Normal is 50%, selected/hovered is 100%.
It is multiplied with the marker alpha.
So if the makers alpha is zero, then nothing you can do
But it doesn't work like that for markers currently
But it does
It is multiplied with the marker alpha.
Yes, that's what I described earlier with 1.5 * currAlpha (guess it was a bad description)
So if selected or hovered it should use (configfile >> "Cfg3DEN" >> "Marker" >> "Draw" >> "2D" >> "alphaHover") as absolute value
Why not 100% though?
I still want to see that its supposed to be in some way transparent, even if selected ๐ค
Hmmm, not sure, with alpha 0.7 the difference is probably not noticeable
Imho It should just be 100% as with all other icons.
configfile >> "Cfg3DEN" >> "Marker" >> "Draw" >> "2D" >> "alphaSelected" //1
Would also be consistent with that
๐ฅบ
Is there a chance that in engine there are more conditions than documented? https://community.bistudio.com/wiki/Eden_Editor:_Configuring_Attributes#Condition I remember objectHasInventoryCargo not being documented but works too.
possible? yes, actual? need to check
Would be appreciated ๐
ObjectBrain
ObjectControllable
ObjectAgent
ObjectVehicle
ObjectSimulated
ObjectDestructable
ObjectHasInventoryCargo
LogicModule
CommentCamera
Those are all the variables in these expressions.
it checks if the objects id == camera object id.
The id is -999
that should be the eden camera
I see. So attribute only for the camera.
get3DENSelected "comment" //[-999]
I see. So camera is somehow a comment. Interesting.
out of curiosity... would it be possible to bind the "release uav controls" action to the "get out of vehicle" action?
i really hate having to use the scoll wheel action when i can just press v
and let's be honest, those actions are kinda similar anyways
The following event handlers (https://community.bistudio.com/wiki/Arma_3:_Eden_Editor_Event_Handlers) were added to the biki some time ago but don't seem to do anything. Do these actually exist?
-
OnRotateGridDecrease
-
OnRotateGridIncrease
-
OnMoveGridDecrease
-
OnMoveGridIncrease
a bit off topic, but funny
If you place the drone without crew in eden and then drag and drop some AI into it, when you are controlling it and press v (double?), the AI unit you placed into the UAV will pop up 
I realized my question made it a bit overcomplicated. would be enough to just add "release uav control" to the hotkey menu under commons. for now I made myself a tiny mod that does the same, but would still be nice to have this as a native function
share it pls thank you
cool feature
They do.
They are tied to eden operations that...
Are never used
(Atleast by engine code)
MoveGridIncrease
MoveGridDecrease
RotateGridIncrease
RotateGridDecrease
Seems it changes the size of the grid, if you enable grid alignment
sus
Guess we have some undocumented actions
Was able to document all of them but 5 https://community.bistudio.com/wiki/Eden_Editor:_Actions
These XXXGridIncrease/Decrease actions are so buggy 
AddCustomConnection is not handled (Only by UI onMenuSelected handler)
ChangeSeat is not handled (Only by UI onMenuSelected handler)
AddCustomStart doesn't exist?
PublishCustomComposition
A element in the compositions tree must be selected. Then it opens either the publish dialog, the workshop search (category selected), steam workshop page (someone elses steam composition selected)
Unknown well.. .. Well.. Yeah. Nothing, placeholder
Now I wonder, how did I end u p with addCustomStart instead of addRandomStartwhen all I did was copy the list from you to excel ๐
Alright, all done. Cheers
Added: Headgear model can now be switched based on NVG being on or off
huh?
I can see that being used by modders for futuristic headgear with integrated nvgs or other high tech stuff
For example pilot helmet. Visor up if NVG is down
but what if people just want it to stay up regardless of NVG is down or not
Ah I think for that one I couldn't make the FT ticket public
oh sweet
Then they will not configure a alternate model in their config
whats the config params for that? same as nvg?
oh so it's for mods then
@tall pumice has been tasked to help with documenting that on wiki ๐
class CfgWeapons
{
class ItemCore;
class H_HelmetB : ItemCore
{
class ItemInfo;
};
class H_PilotHelmetHeli_B : H_HelmetB
{
class ItemInfo: ItemInfo
{
modelHMDOff = "A3\Characters_F\Common\headgear_helmet_heli"; // Use this when NVG is off
uniformModel = "A3\Characters_F\Common\headgear_helmet_heli_visor_up"; // Use this when NVG is on
};
};
};
Inttresting?
does it work the same as how nvg works at the moment? been having a crap ton of problems with that & hiddenselections recently
yes
ah
wait, this could be used for that Heavy combat helmet (altyn) from RF without the NVG effect
RF requested that feature
nah, you will always need an nvg, it won't work without it

i'm just happy that i can swap the visor on our pilot helmet!
i really hate the headgear clutter in arsenal
wait theres visor up/down variants?
so how will you workaround it? like invisible NVG if Heavy combat helmet is used?
simply dont
yes
no, we can't work around it
are they not in the release build or am i just really blind
it's there even in promo pics and trailer
7-3-2024EXE rev. 151510 (game) EXE rev. 151510 (Launcher) Size: ~71 MB ENGINE Fixed: Remains collector was not removing invisible dead characters from vehicle wrecks Added: Weapon holder support to the remains collector (WIP) Added: weaponholderManagerMode, weaponholderRemovalMinTime, weaponholde...
yoris beat you to it ๐ค
I was the one who beat yoris to it ๐
damn, i should have requested the headgear swap feature also for facewear
@full sonnet plz ๐
no more sunglasses when nvg
i will send you a carton of ribena
Tweaked: Grenade impacts on characters now deal corresponding minor damage based on the kinetic energy and angle of impact
Does this apply only to hand grenades or also to GPs? And what are the upper limits on damage? (Insert Generation Kill 40mm smoke scene here)
ohhh is this the window fix
Is there a way yet to have nvg slot devices that have an on and off state without adding actual night vision or thermal yet? I remember requesting that when Apex came out ๐ https://forums.bohemia.net/forums/topic/143930-general-discussion-dev-branch/?do=findComment&comment=3040339
Honestly, I rather take a performance hit than to play with the current aggressive LOD switching. Its also pretty annoying that at one point, the LODs start flickering between the two LODs, back and forth... There has to be some work invested into this, as it is now, its really bad The truth is t...
Shame. Would allow that visor up/down action by having an integrated non-night vision NVG device in the helmet
works great btw
btw any plans to have the vanilla heli helmet swap by default? feels like it wouldn't do much harm and is a cool detail. and wouldn't need extra mods then :>
Last status I got was that its being considered
ok, plz consider harder ๐
- honey, hi! where are you?
- I'm at the bus stop, waiting for the bus
- ok, hurry up then.
- ok, I'll wait for the bus faster ๐
This is wife logic
Hence, this is correct logic and no discussion is to be had
Lest you wish to sleep on the couch
@scarlet root Fixed: Eden Editor multiplayer scenarios would apply the player name to the "player" unit, even if a different "playable" unit was chosen
This is probably the issue we had with the nametags
Offtopic from this thread - #arma3_terrain message
Two things regarding drawing markers in 3D in Eden Editor that I have just noticed.
- Marker texts are not localized, instead @STR.... string is shown
- Markers are shown even if their parent layer is set to "Enable Transformation" false
are marker texts localized anywhere else?
i guess this is the type of thing that should be asked here.
could it be possible the remove the 4096x texture limit and allow higher resolutions? i've read that original OFP allowed up to 16k textures.
why would you need such a high res texture
if there really is a limit, then please no. all those tacticool specblackops mods will then have vests with 8k texture pockets on them
On the 2d map
Maybe in the entity browser too, but not sure
Arma can barely handle 4K textures properly
mostly for skyboxes.
what you mean? on all my tests, i didn't had any problem with textures of this size
why
can run into out of memory issues related to VRAM
haven't had it in a while though, but not sure if it's because of my newer pc or cumulative changes to arma / perf branch
well, it's not like changing the textures size limit would instantly turn all of then in 16k textures.
The original textures would continue there (i'm pretty sure that the game will only show 2k textures or above, if you have then set to ultra as well).
but on my older PC it was easy to run into it with GM + ultra textures
there is a big diff b/w high and v.high ultra
not once have i ever had problems with 4k textures flooding my vram
nor have we ever had issues with 4k textures reported in usp
well, even on a 4090 or a quadro?
you'll find lots of reports on this server and on the feedback tracker
At least some by me but also Groove_C
people who don't know how to click the "very high" button on textures instead of "ultra" then
Well... before you go on insulting the intelligence of people unnecessarily:
- very high and ultra seemed to have similar results
- high and very high have marked difference in VRAM
- if ultra or very high is leading to <20 fps and eventually crashing, then that means the game isn't handling this very well, which is what I said in the first place. Reducing quality settings is a workaround to prevent the problem being triggered, not proof that the problem does not exist
But of course, if one person doesn't have a problem then that means the problem does not exist 
it looks more like a specific problem you had because of your older PC. How much VRAM it had?
KJW โ 05/04/2024 22:20
can you help me prove a point plz 4k textures arent going to kill lower end gpus arent they
Dedmen (โฉ๏ฝ-ยด)โโโ๏พ.*๏ฝฅ๏ฝก๏พ /(O_O)/\ โ 05/04/2024 22:40
shouldn't
just lower texture quality if anyone has issues
if your vram is too low for the settings you're on................. lower the settings
If you insist on missing the point then that's up to you
Vanilla game has no problem with ultra textures because it doesn't use 4K textures
GM does have issues
using mixed assets or running missions with different asset sets in the same session means a mix of good looking assets and less good looking assets (because vanilla on high and GM on high are different quality)
or constant toggling settings
Other games will more gracefully handle VRAM, dynamically reduce quality, or less gracefully -- textures go missing, or textures are flushed leading to stuttering
Arma seems to not do this very well or reliably, which is what i was saying. Now if you have 8K textures, the problem might get even worse.
6GB. Now i have 12GB. Groove_C reported the crash even with 20+ GB iirc
24 GB vRAM
oofs, i guess its very bad
Can you give me a string to test that with?
I am at work right now but try @STR_Enable.
Even Livonia in the basegame can create this issue. Just load more and more detailed textures in the VRAM by closely looking at buildings/trees. Takes like 3-4 minutes. Once the vram is fully loaded textures will start flickering and the game will crash. Havenโt tried it since I upgraded to 24 instead of 20GB of vram, but groove had the issue too I have no doubts I can recreate it.
This seems odd given that the whole of vegetation_f_enoch.pbo is under 400MB.
You just load more higher LOD textures in the vram. I could easily recreate and Iโm trying to make a video of it now
Yeah but the entire texture set for vegetation is a couple orders of magnitude smaller than the VRAM.
Iโll try to make a video of it now
If you have a consistent and fast replication method, I'm sure Dedmen can figure it out anyway :P
I have yet to provide him with that video for a lot more then a year
Well something has improved. Itโs taking a lot longer now
But itโs slowly filling. Couldnโt get it above 17.1GB with only the map, but adding vehicles now has me at 19.1GB
Almost there at 23.3gb
well, 64GB of pagefile, 14.4GB of RAM, 24GB of VRAM and tanoa gets a stuttering hell when flying over the center of the island
but no crash
but this PC also almost costs 5k
but vram remained maxed until I now closed the game
Yeah that sounds a lot like what I used to see with 6GB VRAM
In my case I wouldn't look closely at trees but would try and fly over most of the map at low altitude
But mostly focused on buildings and GM vehicles
sucks that it leaks vram
I don't know if it leaks or just doesn't manage it well enough
As a counterpoint, if you stick it in Standard display settings then 2GB VRAM works fine :P
So I'm really not sure what's going on here.
strange, maybe its too much quality? is it compressed or not?
My observations were:
- Vanilla assets (2K), even with very high and ultra there was no problem, it stays around 3-4GB out of 6GB available (but i don't test indefinitely). Stepping down to high reduces the memory footprint but also noticeably impacts texture quality.
- GM assets (4K), very high and ultra very quickly full up to 5.5GB and stay there, but in relatively short time spans it could cause severe stuttering or a crash. Reasonably easy to force this condition by changing maps, placing different assets in editor, zooming in and out or moving over terrain, going in to virtual arsenal to edit vehicle customisation or unit loadouts, and then repeating on a new map. The first sign of trouble is usually textures taking multiple seconds to load.
- high textures are 3-4 GB iirc (maybe it's equivalent to 2K ultra?) and didn't seem to cause issues.
- these repro steps may seem artificial but it's kind of what you do when mission making (which has organically triggered crashes, sometimes even on an empty Gabreta)
- system config for these tests: i5-5775C, 32GB DDR3-2133, GTX 1660 Ti 6GB, game installed on PCIe 3.0 NVMe SSD.
so it doesnt unload the textures when switching maps?
Possibly yeah. I know there's been discussion here or in #perf_prof_branch that flushing isn't working properly or something, but that's beyond my knowledge
interesting question: what happens if you switch from one map with specific assets to another with assets that arent shared with the previous map?
I usually start from an empty map and spawn random assets. I've had it if i copy-paste assets, but more recent years that stopped causing a crash, in which case i found i needed different assets
Crash frequency did decrease in recent years, where it mostly led to stuttering and very low fps (<5)
Going to try with this memory though, since i've had good results on my new system https://github.com/GoldJohnKing/mimalloc/releases
oh hmm, which malloc would be recommended? (particularly for windows 7)
Vram stays basically unchanged. I switched to Altis and Tanoa and he game couldnโt recover anyone. This is on a 4090, 7950X3D and T700 PCIE 5.0 NVME, X670E chipset and 64GB CL32 DDR5.
Still nothing would be maxed out, except for the pagefile.
sucks, sounds a bit like a memory leak to me
yeah but, dont altis and tanoa have shared assets? i was more implying like custom terrains that have nothing in common with each other and use their own assets or like switching from vanilla to cup for example
The game unloaded like 2-3GB of VRAM but it would fill that up in no time
I was recording but OBS crashed and the file didnโt have the proper settings
oh bummer
that kind of scenario is why obs recommends mkv
you can always remux it quickly to mp4 from the file menu
I know ๐
how much vram you have? 24gb?
and it gets filled pretty quickly? what graphics settings you got?
Yes 24 GB. It takes a considerable time now. I made live commenting yesterday.
All maxed, 5-40km view distance. ๐
ah okay
I can try with GM later
gm? golden master or game master aka zeus? ๐
What I tested was vanilla arma
gotcha
Global mobilization DLC
oh okay
Because SK mentioned that I creates a lot more load. Also I didnโt run any special allocator.
hmm, that cdlc has a lot of assets or higher quality than vanilla?
oh okay, sorry for being a bit forgetful and stuff
@full sonnet
Any impressions on why a terrain baked object would not be able to utilise destructionEffects when the model has the class = road property?
Did you try using the damage property in model instead?
Yup.
I tried a great many things. ๐
Setting damage = building in the model would allow the object to sink into the floor on death. But it refuses to play any destruction effects outlined in the config of the object. So no particle effects, no swapping to a ruin object if specified.
Loosely related, I have also found out that if a models center position is in water destruction effects behave... strangely. Their finish flag gets set to true almost immediately. Causing premature ending of delayedDestruction.
Just a question regarding addForce command, how will a person react to it if the setUnconsious parameter is set to false?
Surprised?๐
Markers are shown even if their parent layer is set to "Enable Transformation" false
But I think that would be correct.
Units and vehicles and comments inside layer are also shown when transformation is disabled.
localization is fixed
Afaik the limit is 8k, nor 4k?
Roads don't usually have destruction effects do they? so I wouldn't expect that to work ๐
I can't find anything but it's alot of stuff to check
bonk
No. I guess they don't. Though the only object you typically would find with class = road is a bridge, which in vanilla cannot be damaged. I was more wondering if the class was intentionally being excluded from features such as destruction effects, I noticed it can't support a hitpoint lod either as another example.
If there was no intentional exclusion, then would it be feasible to enable it on that class was my next thought.
Don't want to spend the time to search further.
I couldn't find anything about named property "class" being road being related to destruction.
Road shapes can also get merged together, maybe it could loose the effects in that process
As you can see in this video, the reaction to this slap was a nap, but I'm wondering what will happen if i setUnconscious to false
find out
It should just be that they're ragdolled but not unconscious, same as if you get lightly tapped by an object in the current version.
its the same as when you get hit by a vehicle
you fall over, lay on the ground a bit.
But you are not unconscious
from my tests, it seems the limit is 4k.
other mods that deal with textures also go on as further as 4k only.
oh yeah, for some reason shooting people from a distance causes a unrealistic ragdolling to happen
new or old?
wdym? this is with ace in profiling/stable. sorry for posting about that here but the conversation reminded me of that
thats known with ace
Anyone know if there is a line I can throw in a config that allows the NVG model to stay on the character when they are inside vehicles?
there are multiple lines for different vehicle seats
however, for ffv seats and some other exceptions, nothing will help
Added: Vehicle marker lights now animate with animated memory points does this mean the light points can now move with model.cfg animation (use case such as proper position/nav lights on the Black Wasp folding wingtips, which are currently just an emissive texture) or something else?
I really hope so
as someone who is not a programmer, Added: DLL extension access to DirectX what kind of funny things can we do with that?
These and the ones below the message: #arma3_tools message
render custom stuff inside the game without injection.
damnit, why am i not a programmer
any cool ideas? ;D
i have some, but no way will i ever be able to do it
Streamed youtube videos watched on the laptop object ingame ๐
or other tube.
I doubt this would realistically be doable or if it would even look remotely good but: video call, face extraction, apply speaker's face to soldier's face texture ๐
But the actual result will most probably look so scary and mutilated that it will prevent you from sleeping at night
OFP vibes
London underground?๐ค
random images of ryan reynolds
Yes
I don't think that was confirmed working yet. Intention was for animated collision light position
Just tested, exhaust is not supported (so far)
i wonder, is this safe from possible security bugs/vulnerabilities?
Watching movies and shows! said the arma 3 cinema gang
You'll be using an extension for that so there's absolutely no security
Any ETA for the next update?
Yes
that's just asking for a zero-day rce that bypasses battleye
extensions run on BE have to be WL by BE
otherwise its the exact same level of security as no BE (none)
its about as secure as running ace
Can confirm this
Also hitpoints smoke effects dont
btw. new RC ready https://forums.bistudio.com/topic/183855-release-candidate-branch-announcements/?p=2949173#entry2949173
Release Candidate Branch Announcements - posted in ARMA 3 - GENERAL: In this thread well post announcements about Release Candidates (RC) of future main branch updates. That includes the opening of the RC-Branch ahead of an update, possible change logs between RC updates and details on (multiplayer) stress tests. This thread itself will remain closed to allow for a clear overview of announcements, please feel free to discuss this branch here.
Mh that bug that re-appears sometimes where specifically on linux server every unit is stuck in an animation...
When animations are loaded, they are cached by a hash.
The hash is
_hash = CalculateStringHash(animationRtmName + Format("_%x", pointerToSkeleton));
Making a hash by appending a pointer as a string to a name is sus.
But formatting the pointer as a 32bit hex integer.. while a pointer is 64bits long 
Also the animations are all only compared by hash, so if there is a hash collision...
Also the hash we store is 32bits long, but everywhere where we use the hash, we convert it into 64bits.
Also the animation also contains a reversed flag for.. loading them reversed. But that's not incorporated into the hash, so reversed and unreversed versions of the same animations, have the same hash. Even though the data inside them is vastly different.
fun fun fuuun.
Was testing the new magazineUnloaded EventHandler and noticed that it cannot be used as a muzzle event like fired and reloaded can. I mentioned it on the feedback tracker. What're the odds of this being something we could do?
https://feedback.bistudio.com/T174935
uff I need to copy it and make a second eventhandler just for the muzzle events
Annoying.. but if you need it..
It would be greatly appreciated! ๐
If you're feeling extra generous, a version of SlotItemChanged and WeaponChanged tied to the weapon would probably be extremely useful. Both of those would be great for the optimization, or at least simplification, of RHS' add an attachment and get different anims system. It would also allow you to specify a certain weapon to check for muzzle switching instead of having to add it globally with an if statement to specify to the weapon class.
Not sure if the new WeaponReloading one would be necessary, since you should be able to find out if the weapon is reloading based on whether reload has fired and reloaded hasn't... but I'm not positive that works the way I think it does.
Here you are:
SlotItemChanged - https://feedback.bistudio.com/T174935
WeaponChanged - https://feedback.bistudio.com/T178384
MagazineUnloaded - https://feedback.bistudio.com/T177742
Thank you again for your many efforts! They are very much appreciated!
Tweaked: Turning off white-listing for Remote Execution (this is a temporary measure until solutions for problems with the current approach are found)
you expecting people will run that? :P
are we definitely getting platform updates after 2.18?
btw since last dev branch i can alt+f4 out of a mission inside 3den and the game won't crash anymore.
progress people, progress
A progress not to break the progress
What are the chances that Chernorus from Dayz will ever appear in Arma 3?
install cup
talk about the release from the developer
install green sea terrains
we'll be lucky to get to 2.30 let alone get another map
There are a lot of problems there
green sea terrains only pvp
There is no implementation of roads for AI
I don't see that as a problem 
The flag is in your hands
2.20 is already on the list
2.18 is basically confirmed, it's just a matter of seeking a release window somewhere towards the end of summer.
2.20 is planned, but purely as just being the next platform update.
At this stage in A3's life, it's healthy to account for the scenario where suddenly further official support is slowed down more or even stopped for whatever reasons (f.e. business). However, there is no actual indication of that happening in the short-term! So our expectation is things will flow as they have recently at least for 2024/2025.
Confirm they are all muzzle events right?
I see Reloaded is a muzzle event.
SlotItemChanged/WeaponChanged on a muzzle is weird. Because these rather apply to the weapon itself. rather than a muzzle on the weapon.
Like if you wanted it to trigger on the weapon (all muzzles), you'd have to copy the EH to each of them which is weird
But afaik we don't have weapon wide config eventhandlers so it must be this way right?
WeaponChanged you would want to fire on both old and new weapon?
So SlotItemChanged will just fire on currently selected/active muzzle.
Also SlotItemChanged will only fire for attachments changed on the weapon.
For entity it would also trigger when the weapon is added to the entity (for example binocular slot). It won't fire on the weapon for that.
Muzzle events:
"MagazineUnloaded" (probably with the on prefix? whatever the other muzzle events do)
"WeaponChanged" (Fires on both old and new weapon)
"SlotItemChanged" (Only weapon attachments, not weapon added/removed from soldier)
Are in, please take care of #community_wiki people and test it on next dev branch, I don't have time for either ๐ฅบ
Like getting setObjectScale and setPlayable to work in MP?
That would be nice for 2.20
||And open sourcing the entire thing so the community can take over once Dedmen is forced to let go||
I think if it was practical to make setObjectScale fully MP-compatible (it does sort of work in MP now), they would've done that when they implemented it. That's not one of the ancient commands from A2 or OFP, it's a new one. It's already at the limits of what the engine can do.
Yes, all of those would be muzzle events so they'd be tied directly to whatever muzzle they're are configured to. It makes sense that SlotItemChanged could only really work on the "this" muzzle since secondary muzzles don't support attachments, but as you mentioned later, a muzzle EH is really the only way for that to work.
Thank you! I will see about going through and updating the wiki accordingly ๐
its a running joke atp because it doesnt work on units in dedi lol
^^
When I use unit1 ignoreTarget player; command on a player in a car, game instantly freezes. (Tested on current dev build)
vs vehicle player?
unit1 ignoreTarget vehicle player; doesn't cause freezing.
do you have to alt+f4 ?
on that note i think you also freeze the game if you do a for loop without exit condition, or is it supposed to have a max iteration limit?
Something like 10k iterations in the unscheduled code iirc
I think the iteration limit is specific to while loops?
Assumption is that if you do shenanigans with for loops then it's intentional.
while stops at 10k iterations as well
{_a pushBack 1} forEach _a stops at array size limit of 9999999
Alt+F4 does not work, I can barely enter Task Manager.
...
17:18:43 Creating action map cache
17:18:55 No alive in 10000 ms, exceeded timeout of 10000 ms
17:19:07 No alive in 21500 ms, exceeded timeout of 10000 ms
The last two lines are most probably the error.
Now I remember that it probably also freezes when you use ignoreTarget on a dead unit, I just didn't realize it when it happened to me before.
for "_i" from 1 to 2 step 0 do {...} hangs the game and goes to at least 200k iterations
for [{_i = 0}, {true}, {_i = _i + 1}] do {diag_log _i} hangs
Hmm, just toyed around load3DENScenario and sometimes it doesn't work even it returns true. Maybe loading into Eden through Startup Parameter caused it?
Yeah. I think I had some ticket for that or I pmed KK. Not sure what the resolution was though.
When a player gets incapacitated by an enemy AI unit, and the player leaves and rejoins the game as the same unit (still incapacitated), and then manually respawns, the AI unit who incapacitated the player gets bugged and does not attack the respawned player anymore. (Tested on dedicated server on current dev build with respawn on position of death). May be related to this: #arma3_feedback_tracker message
It probably is related, because when I set disabledAI = 1;, it works normally. I also got this in one of the rpt files, which might be the problem.
15:33:38 Error: Object(6 : 6) not found
15:39:45 Error: Object(5 : 6) not found
Another problem I have encountered, or rather an inconsistency, on a hosted server, when I add a FiredMan event handler for a remote (client) unit on the server, it does not trigger if the remote unit has a silencer on and is out of host player's view, but it triggers when the silencer is not mounted, no matter how far from the host player. On a dedicated server it triggers fine for any unit with a silencer on or off.
https://community.bistudio.com/wiki/Arma_3:_Event_Handlers#FiredMan
Special multiplayer behaviour: When added to a remote unit or vehicle, this EH will only fire if said entity is within range of the camera.
That range is determined by the fired ammo's highest visibleFire and audibleFire config value.In case of units, muzzle attachment coefficients are applied too.
Yeah, but the thing is that for unsilenced weapon it triggers even if the remote unit is 3km away from the hosted player. But maybe I have a higher view distance.
Gunshots can be quite loud. 3km may be within its notional audible distance.
When can we expect the next Dev Branch update?
asking for updates will push it back by 2 weeks
whenever they feel like, this game is in maintenance/extended support period and it's very old
I am well aware of the game's age, but for the last little while we were steady on every or every other Wednesday.
Wasn't pushing for an update so much as asking what the schedule may look like for the next one, as some features have requested testing.
In my experience, a pause in dev branch updates often means another release update is approaching, and devbranch will resume after that. I doubt it's 2.18 ("late summer" was last heard about that) but there could be a CDLC update going through. Or maybe people are just on summer holiday ๐คท
given lack of bohemians, probably the latter
ahh
or โฑ๏ธ
Must be taking a vacation to Limnos this summer
A skip of Dev can mean many things, from build issues to ~~pipeline ~~logistical issues to lacking QA availability to there just not being many changes to people absence. It makes every Wednesday afternoon an exciting moment of the week 
repro?
Positive.
2.17.151869, just put Mission file as a part of Launch Parameter, it doesn't work even if it does return true. If I launch Editor from Main Menu, it does seem to work
I just put a quoted mission file name like "test_HARM_FlyingJet.VR" at end of launch parameters but it doesn't open editor at game start
This?
oh full path
"logistical issues"? ๐
loaded into mission with start parameter.
Tried load3DENScenario "STrAtIs";
And it returned true and switched to stratis
loading a mpmissions...
also works
its a different terrain though
maybe only broken when loading different mission on same terrain
Oh interesting
it didn't work.
It switched terrain, and displays the mission name in top right. But it didn't load any of the objects in the mission
huh
so it's case sensitive, what's the cost of making it insensitive? is it worth doing?
Ah I got the problem.
Global state. Fun.
If you first load a "user mission" in eden, a "User mission" flag is set. And if that flag is set, then it searches in your my documents.
If the flag is not set, it uses the path as a global path, without prefixing my documents folder, and fails finding the mission
lol
fixed next prof/dev
On the current dev I was able to lock (manpads, vehicle A-A, A-G missiles) target but all locked missile was not even try to follow target, after shot the missile fly direct vector, as like no-lock shot
perf build same #perf_prof_branch message
load3DENScenario fix is not listed there
also on unrelated note setFace not working with custom faces should work in next dev
Does it still do sounds with custom faces?
Wait what does this even mean, i thought A3 had 2 million lifetime sales?
65k today
ah 65K just today
Lol makes more sense
The tweet said "so far" so it was confusing
Yeps
Yes, 90% today
there is a reason to give Arma 3 a rebirth and release a major addition, for example with Chernorus + ๐
How much Arma 4 dev time are you willing to sacrifice to make that happen?
all of it
Make ๐ Tanoa ๐ And ๐ Livona ๐ Playable ๐ Without ๐ DLC ๐
that would be stupid
Why? Have the usual nagging to buy the DLC when you play there.
This, always said this! Make the maps free and treat the rest like the other DLC stuff.
you're giving away a lot of access with the terrain
plus editor
Could be restricted to not work in editor, only have missions and servers playable with nagging?
Either way, both Tanoa and Livonia are so underutilized, such a shame
interesting solution @winter crescent
i dont use livonia for anything but screenshots because it runs like crap IME
but this is getting off topic
if it's still in the regular price the budget for Arma 4 will explode
yeah indeed
lol
i see, not able to use editor on that terrain at all nor the 2d one.
Yeah Livonia could really do with an optimisation pass
Maybe Leopard can be rehired now ๐
BI did this with DayZ's Livonia DLC with DayZ's latest patch didn't they?
So it is doable, even by BI's standards now
DayZ's newest upcoming map (winter) ported to A3 too? As a new DLC
Well I'm assuming there's going to be no new first party content so...
we cant even get hiddenselections fixed on certain lods lmao
That's intentional. I considered it too unimportant for changelog
I see
so sad. every fix matters!
I just visualized Lexx patting the load3DENScenario fix on its head, saying "Don't worry about what the others say about you. To me, you matter!" ๐
Has anyone else noticed that Eden placement is so weird?
On rocks, like wtf is happening here
If I just comment out 40 lines of code that I don't understand why they are even there, we get this.
Someone must've noticed this over the last 9 years right?
Its not scripted placement. Its the mouse cursor handling
I wonder what will break when I just make this change.
To me the last 9 years state is just obviously broken
scream test time
pretty sure i know this issue for years
cant remember when this happened, because i usually dont place objects on other objects like that
maybe it had something to do with small objects? tables etc? because that placement is very awkward as well
Okey I found one maybe problem.
The old one only places things "on top" of objects.
The new one also lets you place it on bottom of object, if you aim up against a ceiling for example.
But it also lets you do this, before/after
Just a bit annoying that it uses the orientation of the ground below and thus they don't stand "on" the object..
Also how about this..
Is there no way to turn off that placed objects are aligned to terrain below?
Like here, its orienting to terrain below, but I want it to just be zero rotation straight up
I tried the buttons at top but doesn't do anything
yeah you have to orient them in the transform menu every time afaik
ร button on german keyboard
Yeah that's what I would've expected to do it but it doesn't
it makes it so it doesn't follow the terrain while you move it around. But it doesn't influence initial placement
Yea not for initial placement
I want people to be able to do this:
With how great Eden is it surprises me that you cannot do this
That would be so helpfull. No more placing things down on the ground first then trying to adjust the height and if its off by a millimeter and everything would end up flying because of collisions.
is stuff in the viewpilot lod of objects only loaded into vram when the lod is required or is the entire p3d and textures loaded at once?
LOD's are generally only loaded when needed. Some like fire/geometry are probably always loaded
I'll put this in for testing on dev/prof only and we'll see how it handles.
Most likely gonna skip it for 2.18 stable.
Maybe also with that.
It would be more neat if it would just orient it to the object you are clicking on, instead of terrain/sealevel gradient.
Probably doable...
Mh.. Proobably indeed
Have you ever wanted to place things on ceilings?
I don't know why you would do that.. but hey... No judgement.
this is simultaneously great and also annoying ๐๐
wait its a keybind to turn on/off right
Are you checking the geometry LOD instead now?
View and fire lod
It always did that. It then just ignored the position and manually did some roadway checks with the object it hit
Well that's not a good idea. Objects that do move should be placed on roadway
I mean soliders
Also placing physx objects below the geometry/physx lods is also not a good idea
Soldiers and vehicles yeah. Putting a soldier onto a shelf that doesn't have roadway, makes them float.
But I think I'd rather leave that decision to the mission maker, than simply it not being possible to put any objects onto non-roadway things
Soda cans are being pushed out of the shelf if they are not simple object. But why is it a not good idea?
I just mean in terms of physx interactions
Well it will bug around
but so will placing two vehicles inside eachother, and we let people do that too.
So it not being good isn't a reason to not allow it, you can do lots of good with it. But you can also mess it up sure
Can any of this improved positioning logic also be applied to Zeus placement? (Especially to the alt+drag vertical positioning...)
Zeus is oofie
Zeus placement is ๐.
We already discussed that in the past. But Zeus doesn't have UI buttons for place along surface or not, like Eden has.
So it would be a UI change which is hard
Also as I mentioned here
#arma3_feedback_tracker message
It also has a very bad camera movement, especially over water
Can this be changed to AGL?
Last time I used Zeus I was like " Why doesn't if have the Eden placment tools"
That should be doable without any issues I'd think yeah
Changing how camera moves surely can't break anything ๐
I was thinking mostly of the f'ed up placement on top of rocks and stuff, rather than the ability to place things on ceilings
Yeah iirc it doesn't use any intersection with objects
If its just that then probably.
Can you make me a example, screenshots or video so I can see what exactly you mean
At least having roadway intersection would be nice
(I would be very happy just with alt+drag only moving the object vertically instead of also moving it horizontally)
Like when you want to place units on a carrier
They just end up in water
Yeah that's the thing that would need a button to toggle it
Does it? I can't imagine anyone actually wanting the object to move horizontally while they're specifically holding the "move vertically" button
well, local Z and so on
Ah alt drag is the up down only, I thought you meant sideways :u
I don't think I've ever seen height editing doing anything other than height editing
i mean "I've placed a garden gnome on the side of this sphere, now i want to move it 15 cm inside" would be a valid usecase 
I myself encountered it a month ago, but I assumed it was "intented" behaviour. Called it "the Arma jank" we are so used to by now
Who doesn't want to spend 7 minutes trying to place something on a bridge, because every time you try to move it up onto the bridge deck it also goes several meters away from it at the same time. ๐
I need to see it in a video to understand that
I've just recorded a short demo, trimming for upload now
[world's longest copyright check for a video under 2 minutes]
I didn't have much of a test plan so please excuse the slight aimlessness. https://www.youtube.com/watch?v=V0L0zvhVECc
You can see when I'm holding Alt when the cursor changes to the vertical arrow
It definitely feels a lot worse when you're trying to actually do something in a mission. I wish we had a proper 3D translation (and rotation) widget for Zeus, but that's probably out of scope
Ah! Thank you
Yeah that's stupid, should be easy fix
RE: zeus would it be possible for a keybind or something to turn off snapping to ground/objects
similar to how we can turn off snapping in 3den
are there any commands to hide "iconplayer_ca.paa", the red ring on players, i know its linked to gps and difficulty, but trying to use RscMapControlEmpty, but it shows the player
this but it also hides AIs: disableMapIndicators [true,false,false,false];
no im trying to do it local to the new rtt map
i think its being drawn by a func so i cant disable it
Without a UI indicator to see what's current? meh
Make it a command and let ZEN handle that
iv been checking all the cfg that uses the paa, and evan when i remove all of it, its still drawing, so i think it needs a cfg entry or command to disable it built, otherwise its stuck on
and when using that function it still drawing the grid position lines
literaly need a thing with nothing
pretty sure there is a ticket for it
||and Dedmen wanted to look into that some point|| ๐
honestly ive got no clue
i dont think that first one would fit
the second one could work
im trying to do it on rtt dialoge map, not the player map, but that would work
if to be make customizable/togglable for tank, should be done also for infantry
im trying to do it for a helicopter rtt map
but evan a basic funciton like that could do what i want, and i could just script it so when the main map is open, the gps comes back
so its only hidden on the rtt
doubt that will get added though, its more that a year old
p:\a3\3den\config.cpp
8564 iconPlayer = "\a3\3DEN\Data\Cfg3DEN\Object\iconPlayer_ca.paa";
8567 iconPlayable = "\a3\3DEN\Data\Cfg3DEN\Object\iconPlayer_ca.paa";
p:\a3\ui_f\config.cpp
16779 text = "\A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa";
17893 text = "\A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa";
20619 text = "\A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa";
58524 text = "\A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa";
59117 text = "\A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa";
88536 icon = "\A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa";
95491 iconPlayer = "\A3\ui_f\data\igui\cfg\islandmap\iconPlayer_ca.paa";
97899 icon = "\A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa";
p:\a3\ui_f_curator\config.cpp
248 texture = "\A3\ui_f\data\igui\cfg\islandmap\iconPlayer_ca.paa";
250 textureRemote = "\A3\ui_f\data\igui\cfg\islandmap\iconPlayer_ca.paa";
257 texture = "\A3\ui_f\data\igui\cfg\islandmap\iconPlayer_ca.paa";
259 textureRemote = "\A3\ui_f\data\igui\cfg\islandmap\iconPlayer_ca.paa";
6547 text = "\A3\ui_f\data\igui\cfg\islandmap\iconPlayer_ca.paa";
if you havent, you could look up those
maybe one of those help - if you can do a config mod
i tried those
the 3den one dosent apply to my context, neither does curator
the ui_f is a definition for the class RscDisplayMainMap;
i tried making my own without it
but it still shows
think its being drawn with this funciton
(findDisplay 12 displayCtrl 51) ctrlAddEventHandler ["Draw", {
if (BIS_WL_playersAlpha > 0) then {
{
(_this # 0) drawIcon [
"A3\ui_f\data\igui\cfg\islandmap\iconplayer_ca.paa",
if (_x == player) then {[1, 0, 0, BIS_WL_playersAlpha]} else {[1, 1, 1, BIS_WL_playersAlpha]},
getPosVisual _x,
20,
20,
0,
if (_x == player || !alive _x) then {""} else {format [" %1", name _x]},
2,
0.05,
"EtelkaNarrowMediumPro",
"right"
];
} forEach (BIS_WL_iconDrawArrayMap);
};
};
so i cant stop it
if the scripted, you can either disable it with the mission parameter (BIS_WL_playersAlpha == 0) or do ctrlRemoveEventhandler / ctrlRemoveAllEventHandlers
that's ideal for chandeliers and such :P
or inverted light poles
nothing is ever easy with zeus code.
In eden, when you height edit it creates a vertical plane, on which you then move the object up and down.
Zeus instead uses a horizontal plane, and moves it up/down based on mouse movement and recalculates position on the now elevated horizontal plane 
delete zeus code, copy&paste 3den code. ez fix
can i ask what is currently holding up dev?
There was a devbranch update last week. It's common for dev to be every-other-week, so not having an update this week doesn't mean there's a holdup.
oh thanks, sorry that i didnt pay attention to that
I just found out this exists. And has existed for years ๐ค
https://partner.steamgames.com/doc/api/ISteamHTMLSurface
Steam has a integrated webbrowser, that via steam API you can render directly in game ๐ค
Webbrowser in ui2Texture
With Javascript
So many bad ideas
Let's exploit 
...But why? It is just because you thought of something out of nowhere without exact usecase? 
I've been scrolling through SteamAPI and stumbled upon that
hmm. make custom website, host on your server, integrate on laptop scenery hidden selection in your mission file.
can i abuse this for custom mission statistics with checking how often a certain website gets pinged ๐ค
External websites would get messy very quickly. You would still need to go through Arma's whitelisting, same as htmlLoad.
But html file with integrated CSS and Javascript, in a pbo?
I wonder how that works when a Javascript tries to load a external site. Would it just go through unchecked, or would we need to approve it.
yeah, dangerous. easy to abuse i guess
it should follow CORS, its still chromium behind the scenes
If you're looking into steamworks stuff, do you think it would be viable to add the new Steam Timeline APIs as a commands for mods?
Would be neat to for example be able to add markers about doing unconcscious, dying, respawning, objectives etc. to the timeline.
I was thinking about making a mod that uses that via DLL but having it integrated in the game would be way better.
Adding a replay marker every time the sniper misses a shot
The events need configuration in the Steam back-end (icon / localization), so I don't think custom modded events are feasible easily. We're exploring the feature for the future generally though.
Although perhaps the set of pre-defined ones could be useful in a limited way. We'll see.
there's icons provided by steam already, you could add more ofc.
as for the events:
doesn't this mean it accepts whatever the game gives to it?
Even if detailed custom event types aren't doable, I think a generic "something happened" marker that can be triggered by script would be a neat thing to have on its own
https://partner.steamgames.com/doc/features/timeline The icons here are probably already plenty for mods, I didn't see that that was a thing :3
so it's ideal for mods ;D
yeah, I guess you could upload all of your task icons too
and that would be plenty times two
ah yes, but its supposed to be a slightly limited cef or so i thought. other games use that such as tf2. can do like html motds like source engine has(?)
wait nevermind, source engine webview is more win32 land instead, or used to be
That would be hard to do with a mod. You would also need to somehow update the steam api. We ship a older SDK version that doesn't support the new API.
can't you just load steamworks in your dll? 
i hope that's not a big deal
The game already loads steamworks though. Can you load it twice? 
Or rather steam_api. You may be able to bypass steam_api and call into steamworks directly, which might load the latest one from steam?
its preferred to link against a newer sdk instead of loading a newer dll
you replace the steamworks steam_api.dll that is provided and hope everything is fine when the game loads it
but i'd rather recommend making the upgrade in source code and linking against the newer steam_api
that'll break anything depending on a static interface list
so if you'd like to possibly break steam emulators it'll be fun to push this
You cannot replace steam_api.dll with a normal mod
im aware, the option is to literally replace the file yourself on the game install directory... or incorporate the update in the source code side
wouldn't it really do much harm to update steamworks api to benefit from newer features, no?
or you really fear adding work overhead on the source code because of a "minor" dependency
Lately I feel that anything dedmen touches and fixes, it shows another 5 20 years old bugs
God knows what dark magic they had to do 12 years ago to get it all working
i am positive that updating steamworks sdk wont put a big maintenance burden on the source code
I have no idea what you are talking about at this point
You're not making sense to me
Get Biki rendered within Arma itself? (Or any other custom wiki for a server, mission, gamemode etc)
basically, motd = message of the day in a web view
or a cinema gamemode in arma 3 xD
things like that
are one of the usecases of having a webview
and steamworks sdk would need to be updated then mods can implement web screens and other newer features from steamworks
A extension can, as of right now on dev/profiling. Or 2.18 release. Already use and implement the steam browser thing.
But they have to render it into a texture or screen overlay, and manually handle mouse/keyboard inputs.
It would be better if it were an actual UI control

Totally unsuspicious

You see? He's not green
former developers doing some a3 updates huh?
former developer, new changelog poster
another posted a changelog a few months ago
If you mean DnA, Joris-Jan Van't land is still known A3 dev
Raz was on the A3 frontlines before, and he's back for more 
woot
B-but where is our โtechnical issuesโ
Logistical issues delivering the technical issues
Next dev: new param for vehicles 'disableDriverAttenuation', should be true for open vehicles so that when driver gets is external sounds are not attenuated. Complements disableSoundAttenuation for turret and attenuationCargo for cargo.
@lucid totem could you add to wiki please?
I uh what where?
no, former developers, neilzar on march 20
Probably https://community.bistudio.com/wiki/CfgVehicles_Config_Reference along with all the other things missing from there
(I don't know what nor where ๐ฌ)
ah
Possibly here as well https://community.bistudio.com/wiki/Config_Properties_Megalist (but more ideally the individual config pages would be updated to include all that information - hell of a lot of work though)
๐ฉ please someone tell me in #community_wiki - I'm no good at config
if you can make a section for vehicle sound attenuation I will fill it
remoteExecutedJIPID <- ๐
Does that have the same issue as isRemoteExecuted(JIP) where it lies if it was an HC -> client remoteExec?
@unreal arrow Since you're looking at RE, any chance for https://feedback.bistudio.com/T171307 ? I guess RE does entity netid match execution, perhaps it can also do the same for netid+space+anychars so you can stack multiple things for each entity
no change, it is a nightmare
Sad, could've been a mega useful feature
How does it work right now anyway? Regex check?
\d+:\d+ on JIP handle?
it checks if string is netid and treats it like an object, otherwise as string id
look at the example i added
Can't you just change it to check if "beginning" of a string is netid?
no
_args remoteExec ["func1", 0, format ["%1", netid player]]
_args remoteExec ["func2", 0, format ["%1 my RE one", netid player]]
_args remoteExec ["func3", 0, format ["%1 my RE two", netid player]]
ok
Even if having all execute at all, instead of one entity overwriting another
also sleep doesnt work in jip exec have to look at it
Make a note for each mission/mode to add their own postfix to avoid collisions
what do you mean
If my idea was implemented, we should add a note for each mission/mod maker to have their owner postfix when adding JIP RE to an entity
Not right now
with remoteexecutedjipid you can encode object into id and remove message even when object is no longer
is there a ticket? am not aware
I assumed it was intentional.
IIRC client -> client doesn't do it, which suggests there's code specific to headless clients.
you can even have a generic function and all jip call one function then dispatch calls based on jipid
and have one routine for removing redundant messages from queue
calls come in, object linked, is object good? proceed, no good exit and remove from jip
proceed according to second id encoded
no idea
hmm, there's a note on isRemoteExecutedOwner:
in a remote-executed context received from a Headless Client will return 0 by design. To check if the context is remote executed, use isRemoteExecuted or isRemoteExecutedJIP.
Although my recollection is that isRemoteExecuted returns false in that case.
hmm not sure then if remoteexecutedjipid will return "" or id
We actually already had script value to json, and other way around from Argo scripts, it was just never included in Arma 3...
I needed it for testing hashmap things.. So might aswell?
Only bool, float, string, array, hashmap (With only the support types as key/value)
Json to and from CallExtension, instead of barbaric strings? ๐ฅน
jason is not string?
Json is less barbaric string ๐
Having something that allows you to convert Arma stuff to json and pass that to the extension (instead of doing str() in SQF) lets you decode that json using state of the art json libraries instead of writing your own custom "Arma string decoder/encoder" functions in the extension ๐ฆ
Soโฆ no additional doc to do, just add an Arma 3 icon? ๐
Is it already on wiki?
Behavior also changed slightly because hashmap support instead of key-value pair arrays
I also thought I'd change the names to just "value" instead of "gameValue"
Yeah due to the array/hashmap change, the examples on the argo wiki pages also won't work
ok, copy/paste with some edits it is then ๐ ๐
I don't know, you tell me.
Do you like the "gameValue" name. IMO it doesn't make sense, game-value is internal naming for script values/variables. I don't think that is otherwise visible anywhere publicly. So it doesn't really make sense to put here?
You'd still have to use callExtensionArgs, and pass it via parseText.
Otherwise it will still do the double-quoting mess.. right?
Ah seems not.
"fake" callExtension gameValueToJson allExtensions that doesn't do any extra escaping
"fake" callExtension ["test", [gameValueToJson allExtensions]] that does.
"fake" callExtension ["test", [parseText gameValueToJson allExtensions]] that does not
that allExtensions command returns alot of data 
is it possible to do as in DAYZ: for the player's camera to zoom in when the player is at the fence and he could not peek at the fence from a third person?
My use-case is that I'm roughly doing _result = "Pythia" callExtension (str _this); and then parseSimpleArray _result ( + some edge cases because of the 10240 chars limit, etc...).
If I could replace str with gameValueToJson and parseSimpleArray with jsonToGameValue, that would simplify a lot of things
no
I'm not using callExtensionArgs because I have a dispatcher on the extension side that decides which (Python) function to call, since the users provide their own functions, so I don't know how many args there will be, ahead of time
However I'm not sure if having jsonToGameValue return a hashmap (if I understood correctly) would not break a few things ๐ค
Need to go afk, sorry
JSON object will return hashmap, JSON array will return array, string returns string, number returns number, bool returns bool.
So you'd probably just return a JSON array, and you will get a script array out
Oh, so it would basically be backward-compatible with parseSimpleArray but with the additional object->hashmap. Cool ๐
(I mean... besides that it takes in json as input, now)
"Added: valueToJson and jsonToValue script commands" it is
@Lou
valueToJson or valueToJSON ?
second
valueToJSON valueFromJSON
or just toJSON fromJSON
me gusto โฌ๏ธ
"Copy" button of the code benchmark in debug console is partially broken for me in the dev build. The result in clipboard is ```Result:
%2 ms
Cycles:
10000/10000
Code:
getText ((configOf backpackContainer player) >> "assembleInfo" >> "assembleTo") isNotEqualTo ""``` (notice %2 instead of actual execution time). Ticket-worthy?
๐ข
Put comment on https://feedback.bistudio.com/T182822
Very interesting. That shouldn't be possible unless the code says %%2
I concur
Cc @full sonnet
althoughโฆ is the "from" only from hashmap or?
to hashmap
its neither to hashmap nor from hashmap
all json types are supported
im confused
ah ok
I prefer these
(not all caps JSON)
The 11-year-old niche mil-sim Arma 3 somehow managed to sell nearly 700,000 copies during the Steam Summer Sale. https://t.co/9WWo4jwfHb
it just so happened. nobody knows why. it is a mystery
1/3rd of a Kebab might be a hint
so niche
popularityโข๏ธ
toJSON/fromJSON it is!
are those able to replace the CBA functions for JSON parsing and encoding?
You check what CBA does specialful and tell me ๐
This does hashmaps for objects, I don't know if CBA does. if CBA does then it's probably equivalent
https://github.com/CBATeam/CBA_A3/blob/master/addons/hashes/fnc_encodeJSON.sqf
Description:
Serializes input to a JSON string. Can handle
- ARRAY
- BOOL
- CONTROL
- GROUP
- LOCATION
- NAMESPACE- NIL (ANY)
- NUMBER
- OBJECT- STRING
- TASK
- TEAM_MEMBER- HASHMAP
~~ - Everything else will simply be stringified.~~
So.. I'd say no? It might be if you manually iterate the whole structure, and manually stringify unsupported types before passing it to toJSON
Your parseJSON also uses CBA namespace by default. So it would not be a drop-in replacement.
It somewhat works, but not with default arguments
Found some weirdness with remote explosives' damage falloff behaving backwards...? Closer to the model center the less damage is being done. Originally found on some of our explosives in SPE, but also was able to repro with the vanilla satchel charge. ๐ตโ๐ซ
You think this is new issue? Or a hidden gem
toJSON/fromJSON won't support nil/any, isn't it? ๐ค
Not directly, in array/hashmap values I don't know.
I wanted to test that but didn't yet. It should though
Oh wait, json does support null values. Nevermind, carry on
I got bamboozzled by the fact that you can't have null as the object key so for a moment I assumed that you can't have nulls at all.
>>> print(json.dumps({'x': None}))
{"x": null}
>>> print(json.dumps({None: 'x'}))
{"null": "x"}
I tried the latter first, hence my confusion
Mh but hashmaps can have nil keys 
Fix the hashmaps, then ๐ ||/s||
Not entirely sure. Wouldn't surprise me if it's a hidden gem though. I took a video comparing vanilla satchel vs vanilla demo block.
We have two explosives in SPE that are very similar in size to the satchel and demo block. We played around with the service lods on our Satchel equivalent, (Config wise they were the same except for the indirecthit value) if we shrank it down to smaller than the visible part it started to work. So it might have something to with exceeding some bounding sphere limit? ๐คท
Sure.
Y U NO GREEN
At last, recognition! https://x.com/ArmaPlatform/status/1816134727977349312
heh neat
holy Gun Jesus
hmm would be kinda cool if we could change sound attenuation by script. this way could imitate open doors / windows
cant remember right now if it's a single value or related to all sound effects, though.. so might not be as easy
That would be really useful for convertible vehicles, e.g. GM Iltis, UAZ
Speaking of sound attenuation, seems to me sound attenuation when move out from building/vehicle should not be instant
https://feedback.bistudio.com/T181711
Here you go:
https://feedback.bistudio.com/T183118
Mission is in the ticket.
What does that actually mean? mode is always 2 regardless of mode setting in mods/mission?
Any progress / statement on this? Would be awesome to have populated servers running Tanoa and Livonia 
possible fix for special characters in paths?
i get a message when opening and then rpt spam while the player cards are in display:
Warning Message: Picture c:\users\ not found
CfgRemoteExec does nothing atm but remoteExec filter should still work
sounds like a mod issue?
#dev_rc_branch message
so... we gonna have RscBrowser
?
UIs with CSS and JS
Nothing concrete is in the plans right now, but we will discuss such ideas for the long-term future of A3.
No. Because you can only load from such paths, with -filePatching enabled. Which is a developer-only option that is not set during normal play.
uhhh, this is not directly related to filepatching, this is a part that seems to load user images from a path and fp is disabled
Might be some bad mod script
With -debug you should get context, if it comes from a script
no, it happens when i open the player profile (right shift+p) and as long as i have that open the warning is spamming the rpt, my user profile is called '
nah its a vanilla thing
Make a feedback tracker ticket with repro steps
okay
basically what fails to load is the images that should be there representing the player(s)
Tweaked: Regenerated a lot of Eden Editor previews to fix glitches with various assets
What were these glitches?
I know a few were wrong and some had incorrect aspect ratio
some of them were related to outdated loadouts in preview pictures, some of them contained older state of the structure from development and sometimes they were using wrong picture due to incorrect config
Cheers
never noticed any of that ๐
does the helmet nvg thingybob support hiddenselections?
don't expect it
roger
os there any chance of that scripted depth of field postprocess effect for 2.18 or is it too late now
late
yay
is it expected for selectionPosition with syntax 2 to take 500ms at 10000 iterations? This is incredibly long
Edit: The culprit is the use of "AveragePoint" or "BoundingBox" as the return mode... I get that it probably has to check all the vertex positions to get this number, but still. That feels long
500 milliseconds per call?
or you mean 500 microseconds, or 500ms for 10k iterations (50 microseconds)?
The second one, 500 milliseconds per 10k
did you use code performance? if it says 500ms at 10000 iterations that is per call, as the resulting time is divided by number of iterations
could you just copy the result and post here, so confusing
I think Arma doesn't cache bone transforms so every time you call that command it reanimates the bones (and it probably does it once per vertex in avg mode?)
Yeah. It reanimates each vertex 
should be doable to just do it once for all vertices? 
Yeah. It reanimates each vertex
But it depends on what type of object it is.
But yeah we can animate the whole object, before we pull the points, instead of doing it for each point.
But, not soon, and with ticket repro/benchmark
This is the code I ran (13 is firegeo lod).
selectionPosition [player, "Spine3",13, true, "AveragePoint"];
Result: Execution Time: 0.0511 ms | Cycles: 10000/10000 | Total Time: 511 ms
Using something with less vertex's, like "Head", results in a total of about 220ms, proving it runs more for more vertex's
My use case has been the player object. It would be great if this could be improved for others in the future, but for now I can use the "Hit-Points" Lod which has a similar variety of points while only have 1-4 vertex's per selection. Making the time much less
0.05ms isn't that bad
And I would expect the Spine3 selection to only contain one point? ๐ค
Why not use FirstPoint?
It contains 30+ points in the firegeo lod. I can't use first point because I want to get atleast close to the centerpoint of the selection. on the dev branch using diag_captureframe, that one line of code is resposible for 6ms per frame
Ah right you are doing that extreme objects follow thing
Yes, like I said a few messages though, I've found a solution thats good enough for my case. Just here to try and improve it for later down the line now
for commands which have already been implemented and need changes do we just put it on the ft ticket?
ticket's already been closed as resolved and assignee (leopard) removed 
May I ask what you are trying to build up?
My use case it a dynamic loadout builder (attaching objects to the player)
You could, but you need to let someone know
https://feedback.bistudio.com/T160321 latest comment
I would really like the
switchCamera
command to work with the parameter set
thirdPersonView = 0;
@full sonnet
Is this https://feedback.bistudio.com/T183118 making it into 2.18? If so do you want any testing after it reaches dev branch?
Next dev and prof tomorrow (not dev tomorrow)
Fix looks pretty good. If you want to test you could on internal >152082
Hmm last time I tried updating my internal (two weeks ago) it crashed on boot. So I got scared. ๐
But I will give it a go.
Any rough idea about when 2.18 might ship Y/Q guess there are certain schedule windows throughout the year when it might drop (sorry newbie here no idea about BI release processes) ?
The last official word I can recall was that the target (not guaranteed) was "late summer".
Recently there have been some mentions of certain changes being "too late for 2.18" so I think it's going to be reasonably soon.
Probably see RC for it this month I would imagine.
oh hmm, only bi and cdlc devs get internal access?
Yes
would it be possible to symlink/hardlink/junction most of the pbos that don't change with dev to save on space with the p drive and maintaining a dev install?
Anything can change on dev branch. It's perf/prof that maintains data compatibility.
im aware, im asking since largely there's no changes if a link based setup would work better to save on storage or if it wouldn't work due to the amount of changes
ummm sorry to ask here but are tester positions open at all from time to time nowadays?
usually you get invited by them
gotcha
We'll start branching / build snapshotting in about a week, then public RC opens in the weeks following, so could still be this month.
Release of 2.18 may be in September, but it's still being puzzled in the BI release schedule (we need Publishing QA for example).
Thank you very much!
oh good, what data changes are there to expect?
Check back Dev-Branch change log since the last update. Don't expect new content.
non-local objects don't use PhysX iirc so angular velocity probably doesn't do anything for them
its being run locally
but the interpolation should probably still be there given it interpolates for velocity
You missed configfile >> "CfgVehicles" >> "Land_CzechHedgehog_01_old_F" >> "editorPreview"
It's using a none existent preview with suffix _old
I think its old issue https://feedback.bistudio.com/T157167
It's already fixed. It didn't make it to devbranch
I added some questions/feedback on its behaviour on the ticket https://feedback.bistudio.com/T180721
I wouldn't worry about the perf cost of the EH. The artillery shell itself is likely to be far more expensive on networking.
Especially if your MLRS is firing cluster rockets :P
But i am thinking how useful it is to know all that information for every shell in the fire mission, because it'll all be the same except the shell, and the shell you can get from Fired EH anyway
I guess that shouldn't happen... Happens every time I start editor preview after DEV update... ๐
17:15:41 Invalid path (only relative paths supported): 'C:\Users\xxx\Documents\Arma 3 - Other Profiles\Xeno\missions\co30_Domination.Altis\init.sqf'
17:15:41 Warning Message: Script C:\Users\xxx\Documents\Arma 3 - Other Profiles\Xeno\missions\co30_Domination.Altis\init.sqf not found
17:15:41 ScriptVM file C:\Users\xxx\Documents\Arma 3 - Other Profiles\Xeno\missions\co30_Domination.Altis\init.sqf cannot be opened!
Dedicated Server may be unable to load mission files
Looks to be the same issue for that too. CC @earnest mural
Fix for profiling branch tomorrow, dev-branch next week maybe.
I can give you last weeks dev-branch exe if you need it?
All good, no hurry
is the angular velocity stuff likely to get looked at before release
no
sad
the big shells are network synced so can trigger EH globally without loss of performance. they are also seldom events in game not like firing vulcan
I see, thanks!
In hope seeing more data fixes for 2.18 want again send list of the issues: https://pastelink.net/a3configIssues
although i would say, is there a way to expand the EH to obtain the fire mission too? like do fire missions have an ID or something?
In my current implementation, I was handling it by adding a Fired EH and then removing it using the code executed by the EH (so that it only fired once), and then adding it again later, which i think is super clunky.
I just need to answer the question "where is the artillery firing this time" but having this information per shell in the fire mission is not useful to me (but it can be for other implementations, so it's good to have), and would just add more processing steps (which i'm not sure of how to handle with this EH really).
So this EH solves the problem of having a unified EH to track artillery fire events and targets for players and AI (hopefully ignores commander turrets?) but still will cause complications for tracking per-firemission info. Like i could say, if the target is the same then it's the same fire mission, but that's not necessarily true at all. I could impose a time limit but artillery units in the game have significantly different salvo firing rates. So i'm a bit unsure of what the solution is, short of removing and re-adding the EH, but since this is global and tracks all artillery units, it would be chaos :/
It could be as simple as adding a counter to the vehicle (unit namespace?) that increments from 0 on each fire mission, and is returned by the EH
(Although that would still require some processing in code to prevent things running per shell fired)
Maybe ArtilleryMissionFired could be a variant EH of this, that just execs per fire mission?
Here are two more for the list:
https://feedback.bistudio.com/T138393
https://feedback.bistudio.com/T174468
Nice! added https://pastelink.net/a3configIssues
Someone said awhile back that there was going to be a method of protecting texture and model content within pbo's coming in the future, is this still a thing? (I think dedmen said it, but I cant find the message)
Binarisation is already a thing for model.
That there are tools that will unbin it is another thing.
And tbh, whatever you ship to client can and will be stolen. The game has to get the data somehow. If someone is motivated enough, he'll get it anyway.
It wasnt about Binarisation
iirc it was just pbos in general
but as ilbi says, theres no way to stop people ripping what needs to be loaded to be displayed
not textures
Having troubles with ocap exporter extension, saying it could not be found.
It was reported working before the last update, any ideas?
(diag exe)
๐
In dire need of feedback!
Thinking about adding a WebBrowser as a UI control. Question is security/safety.
All URL's need to be whitelisted (same as htmlLoad, or clickable links on Button/StructuredText controls. Whitelist is controlled by mod config or description.ext)
The differences to existing website opening (via button that opens steam overlay):
- Can run Javascript inside the page, from SQF (And the javascript inside the page, can trigger SQF eventhandler to allow bi-directional communication)
- Can be opened without user input
- Can be displayed ingame (as opposed to steam overlay or tabbing you out of the game to open in your main webbrowser)
What I got so far:
What if someone just displays porn that could get twitch streamers on the servers banned.
The URL needs to be whitelisted, so the mission maker would specifically need to whitelist the bad website to allow for this to happen.
But what if the mission maker is malicious and wants to do it.
Then the mission maker could already include bad images/videos in the mission file, and already display them in UI.
Also we had to hardcoded block Steam websites, because you would be logged in on them and via Javascript a SQF script could be editing your steam profile. So no visiting Steam in these.
For now the idea is to show the user a popup when a webbrowser control is used.
"Arma 3 wants to show web content, allow?
Yes / Yes - Always / No / No - Always"
The "Yes " option would remember it for the current game session, so you only get one popup per mission.
The "Always" option stores it into your profile and remembers it forever.
Another safety option could be, to simply not allow at all to open external websites.
Everything to open, would need to be a .html file inside a pbo, or just generated as a string in script.
But that also opens the question about iframe's. You could whitelist a website, that includes a non-whitelisted website in an iframe.
Though if you have the power to whitelist a site that contains a malicious iframe, you might aswell just whitelist the malicious website and open it directly.
Also looking for ideas of what you would consider "malicious" things that could be done with this.
Displaying bad images, videos, sounds is covered above.
It could be used to grab players IP address, but that can already be done with htmlLoad (if website is whitelisted).
Running a in-browser bitcoin miner, well I guess? I'd put this on the same level as the mission maker whitelisting porn websites.
The browser is a embedded chromium, same as in Steam overlay. With all its security features, and being kept up-to-date via Steam client updates.
For discussion purposes, making this a thread. Please reply in there so we can keep it together.
