#arma3_feedback_tracker
1 messages · Page 30 of 1
I do that a lot
I've never seen you sleep. You're here 24/7!
lies!
@untold sky Are we going to get a fix on the VON slider not working? Currently it's basically an on/off toggle, no in-between. For some reason you can change the VON volume with effects though? Pretty broken stuff.
Seems like there is a ticket for it as well.
There is a fix on profiling branch so far I know. I dunno if a hotfix is planned or not
Hope so, I mean it's very annoying not being able to turn down people's volume haha
just switch to profiling branch. Youll get your fix, and free performance improvements on top
@lapis osprey If you dunno how: https://community.bistudio.com/wiki/Arma_3_Steam_Branches
would have been cool if one could simply move existing markers, instead of deleting, putting again and naming it/them again
Just use ACE then.
mod...
Yep. Have fun
https://feedback.bistudio.com/T154453 (Old Man: Virus Task) Any chance someone can look into this? I had this error now in two playthroughs. Others seem to have the same issue. See #old_man_feedback
Afaik old man guy is currently too busy
Ah too bad. Hope he finds some time at some point.
It’d be nice to have a functional VoN slider in the volume options, considering it has been working perfectly fine since the release of the game (and even in Arma 2). It only became bugged out in the recent 2.00 update when one of the devs “unified the VoN volume handling” for an undisclosed reason. It really is a mystery why the bug wasn’t fixed within 24 hours of it being reported, considering how major it is. Every other game I play has the ability to modulate voice chat volume independently of effects - and for a very good reason. I’m not going to pretend to know everything, but it doesn’t sound too complicated to revert back to the previous VoN handling that never had any issues to my knowledge. Telling people to switch to and download a separate branch of the game that is incapable with the majority of servers that people play on, and recommending they just install mods to make the issue irrelevant, sets a dangerously low standard for Bohemia Interactive. I hope that someone will take responsibility for the issue and commit to making Arma 3 great again by setting higher standards for the development team. I shouldn’t have to rely on mods to make the game playable. Mods are meant to expand upon or improve and already working game. They should not be required in order to fix a broken voice chat. Incase anyone still doesn’t know what I’m talking about, I will link to the feedback tracker post and explain it here very briefly:
VoN volume slider in vanilla Arma 3 is broken because:
- VoN slider at 0% mutes VoN
- VoN slider >0% enables VoN
- VoN slider between 0 and 100 does not change voice chat volume
- Effects slider changes all voice channel volume.
https://feedback.bistudio.com/T154335
So, I think by unifying it (still confused why this was even done), the dev in question linked the volume of all voice channels to the effects slider, which is what direct chat was always tied to. The reason I view this as a major issue that needs to have a higher priority is because there are many people who play on Official and other un-modded servers, which rely on the default voice chat system. Because of this bug, people have taken a break from Arma 3 because they don’t want to damage their hearing in order to hear others in voice chat. You have to turn your effects volume up higher to hear people, but this obviously makes other sounds in the environment really loud. I think at this point the community is due an explanation of why this was done in the first place, and why they are so reluctant to fix the problem on the main branch...
It’s fixed on main branch?
and recommending they just install mods to make the issue irrelevant
mods cannot fix this issue, its a engine bug
It really is a mystery why the bug wasn’t fixed within 24 hours of it being reported
it was fixed. Right after it was reported.
a separate branch of the game that is incapable with the majority of servers that people play on
incapable? its perfectly compatible with all servers, and even has some additional performance improvements and bugfixes.
back to the previous VoN handling that never had any issues to my knowledge
Direct chat was always on effects volume slider, instead of the VON volume slider, people have been reporting that for the last 4 to 7 years, but devs never noticed that issue nor saw reports about it.
I hope that someone will take responsibility for the issue
I broke it, and I fixed it within hours after seeing the report.
You have to turn your effects volume up higher to hear people, but this obviously makes other sounds in the environment really loud.
This was always an issue with direct chat
Should direct chat have it’s own slider then?
I think at this point the community is due an explanation of why this was done in the first place
Quite sure I explained that already... Cleaned up the code, didnt expect that VON volume would ever be controlled by effects volume, because I expected (as is logical) it to be controlled by the VON volume slider, wasn't the case, but didnt expect or know and thus didnt check for that, neither did the people who reviewed and tested the change.
No why? VON volume is VON volume. All VON volume no matter what type, should be on the VON Volume slider
Alright thats cool, so now von controls direct chat as well?
yeah
Sounds like an improvement, good work I suppose
Im gonna report another issue here if thats alright
It has to do with the A3 Units
When you have a long link as the website of your unit, it will glitch out the ui, making the buttons for mute, vote kick, vote admin, and kick off dissapear
ingame? in the map screen?
Yes, on the player list
Make a ticket and I might look into it. This channel is just for discussion not for actual bug reporting
Well at least know u know about it
It’s especially powerful for trolls when coupled with putting an emoji in their name, since emojis dont show up in-game and nobody’s gonna guess which one it is... making chat commands like #kick innefective as well
This makes the only way to kick them off the buttons on the lobby screen
Keep in mind i’m talking vanilla here, so no admin menus of the sort
I think #kick should also accept the steamid of a user, not sure
Is there an eta for when the von fix will be on main branch? Does it need more testing?
no eta
But yes we also pushed it to profiling branch, to get some testing and feedback for the change
Ok cuz i had to google the access code for that branch. Anyways, the reason I would recommend a separate slider for direct chat is because you might want to hear people better in direct chat at a distance without turning up the other channels. Better yet, have sliders for all channels because sometimes there will be people talking over your group members. I guess I could just mute them individually, but it would be convenient to have a feature like this. Thats my feedback
Was something changed in the Scenario Menu? Scenarios don't seem to get marked anymore as played. (Profiling Branch)
negative
@daring wagon was it changing on profile branch before?
@solid marten That I don't know. I just did another test on dev branch and calling BIS_fnc_endMission does also not mark the mission as played. I will do some further testing if you want.
could you try it on 2.00?
Did, doesn't work. Downloading legacy now. Also switched to a new, clean profile and no mods of course. I keep you posted.
thanks
Okay, on Legacy the following happened:
- Local mission still didn't get marked as played
- All my Steam mission which were not marked as played in the other builds suddenly became marked as played.
So I don't know what's wrong. I'll ask around if someone else has that issue.
I'm having trouble hearing people speak because effects volume now effects VON as well
known, fixed in perf branch
#perf_prof_branch
i'm not sure, if it is bugtracker in this discord, but i need to vent and report - please, remove delete action from Enter key in Publish to Steam Workshop modal window in the Eden Editor
I accidentially deleted my mission with 4k subs and have no chance to recover it 😠
But you have to press enter twice to actually delete it.
But yeah, Enter should never focus on delete.
ideally the second Enter should be on "Cancel" yes
Or like that
Just checked, there is also no indication what button is focused when the GUI is opened. Arma GUI 🤷♂️
The game crashes when I delete a controlsGroup using a child button in one frame (as in: onButtonClick = "ctrlDelete ctrlParentControlsGroup (_this#0)")
But if I spawn it everything is fine (even if I put isNil around it to force the code to run in one frame: onButtonClick = "_this spawn {isNil {ctrlDelete ctrlParentControlsGroup (_this#0)}}")
Is this a bug?
jup. and its not new either
you are deleting stuff while stuff is being processed on it
So why does spawning the code work?
Maybe because it happens in the next frame(s)?
spawn does execute the code a bit later than its execution, if my knowledge is correct
yes because it happens later
and in a different section of the frame
outside of UI processing code
in general deleting controls inside events
imagine a control with 3 eventhandlers
the 2nd EH deletes the control. So engine goes "okey Im done with the second, lets get to the third, I already have it here right next to... wait.. where is it.. I had it here! I just printed that out! where.. wtf.. Bridget!!! Wheres my handler! I know I put it here! Bridget come help! Bridget!!"
Interesting. Same thing happens with ctrlEnable!
Children should never mess with their parents! Lesson learned! 😉
Okay, dedmen goes more mental, than usual. Something is odd.
when placing any of the vr soldiers (or civillian) in 3den on diag branch you get this rpt log:
Selection missing in CfgModels
class vr_protagonist_f
"camo3",
Sounds like a ticket for reyhard 
On dev branch. When switching from a mission which has function in CfgFunctions defined to an empty mission, the .rpt gets spammed with:
14:24:03 Warning Message: Script functions\fn_alarm.sqf not found
14:24:03 \/Context: [] L291 (A3\functions_f\initFunctions.sqf)
[] L299 (A3\functions_f\initFunctions.sqf)
[] L304 (A3\functions_f\initFunctions.sqf)
[] L314 (A3\functions_f\initFunctions.sqf)
[] L323 (A3\functions_f\initFunctions.sqf)
[] L328 (A3\functions_f\initFunctions.sqf)
[] L333 (A3\functions_f\initFunctions.sqf)
[] L338 (A3\functions_f\initFunctions.sqf)
[] L374 (A3\functions_f\initFunctions.sqf)
[] L375 (A3\functions_f\initFunctions.sqf)
[] L102 (A3\functions_f\initFunctions.sqf)
[] L129 (A3\functions_f\initFunctions.sqf)
[] L132 (A3\functions_f\initFunctions.sqf)`
The file exists and definition is correct*
in v2.00 too
I get a "hey, file not found" when trying to get a fresh new start
Yeah, it's been always like this, but wasn't it supposed to be fixed?
Need to check the ticket for the status
Also I am not sure if this kind of rpt log is not a bit too much
14:26:30 Attempt to override final function - bis_fnc_om_modulefasttravel
14:26:30 \/Context: [] L1 ()
[] L3 ()
[] L9 (A3\functions_f\debug\fn_recompile.sqf [BIS_fnc_recompile])
[] L11 (A3\functions_f\debug\fn_recompile.sqf [BIS_fnc_recompile])
[] L291 (A3\functions_f\initFunctions.sqf)
[] L299 (A3\functions_f\initFunctions.sqf)
[] L304 (A3\functions_f\initFunctions.sqf)
[] L314 (A3\functions_f\initFunctions.sqf)
[] L323 (A3\functions_f\initFunctions.sqf)
[] L328 (A3\functions_f\initFunctions.sqf)
[] L333 (A3\functions_f\initFunctions.sqf)
[] L338 (A3\functions_f\initFunctions.sqf)
[] L381 (A3\functions_f\initFunctions.sqf)
[] L382 (A3\functions_f\initFunctions.sqf)
[] L396 (A3\functions_f\initFunctions.sqf)
[] L392 (A3\functions_f\initFunctions.sqf)
do the functions work despite the error? or is CfgFunctions in missions outright broken?
Let me check
it is too much, it will be hidden behind a new -debug command line option
it is already hidden behind -showScriptErrors
Seems broken, functions viewer shows empty text
it is too much, it will be hidden behind a new -debug command line option
@untold sky Awesome
Something I am not sure how others think about is the default stuff that is always logged into the rpt. Right now I have 700 lines telling me what pbo was loaded, that some base classes are being updated and other stuff. Would it be possible to hide that behind another debug command as well?
this info is important when you get issue reports as mod developer
the "no geometry or visual shape" I'm hiding. The list of what PBO's is loaded is important for analyzing crashes, so has to always be on
base class errors should just get fixed ┬─┬ ノ( ゜-゜ノ)
I am fine with the addons but yeah, these base classes are just annoying
I have no idea about your file not found issue. Maybe @solid marten knows what it could be.
Best would be if you can provide repro mission and FT ticket
It's also kinda ironic that it doesn't tell me that the functions don't exist when loading the scenario they are defined in, but rather when I load a differerent scenario
oh thats... interesting
repro needs 2 repro missions probably.
"First load this mission, then switch over to that mission"
Maybe when switching to new mission, it somehow reads old missions CfgFunctions
This is the scenario without functions. See how the entity list is also not updated.
yeah FT ticket, good find 😄
Maybe when switching to new mission, it somehow reads old missions CfgFunctions
that popup was always a thing afaik 😄
ticket please with repro
@solid marten https://feedback.bistudio.com/T154708
And the second requested ticket https://feedback.bistudio.com/T154710
Is there a script error logic error in line 21/ 22 in BIS_fnc_sandstorm?
_density = if (count _this > 1) then {_this select 1} else {0.07}; if (_density < 0) then {_density = 0.07};
_colorCoef = if (count _this > 2) then {_this select 2} else {1}; if (_colorCoef < 0) then {_density = 1}; //Shouldn't that reset _colorCoef to 1 instead of _density?

Dedmen do no squieff
@daring wagon looks like copy paste error can you make ticket as well? thanks for the other 2
params 
lol yeah. I've seen so many bad practices in many BIS functions.
@sinful kettle This functions is really old. Was created before BIS_fnc_param and params etc.
@solid marten https://feedback.bistudio.com/T154714
I'm not talking about this alone. But yeah most of those that I've seen are old
@solid marten Is it possible to change the suppression so that the unit could also be suppressed by rockets/grenades as well? Right now they only react to bullets.
It is possible but it is not easy and will require quite a bit of change around the engine possibly affecting performance. I’m not doing it
-
I have updated the BIS_fnc_sandstorm ticket with an additional suggestion. Maybe you have time for it while you are at it, KK.
-
I might have a bug with ctrlSetURL but I am not sure, maybe it's intended.
When just using_button ctrlSetURL "someURL";The tooltip does not show at all.
When setting an URL and a custom tooltip it shows the URL in the first line and the custom tooltip in the second line.
Ideally I'd prefer ctrlSetURL not forcing a tooltip by default and then not even showing it. Maybe as optional parameter, or the option to overwrite the tooltip by using setTooltip after ctrlSetURL.
easy to check if this is ctrlSetUrl or not. use url config property to set url instead and check behaviour @daring wagon
Okay, tested. From my observations:
config URL:
Tooltip is visible, custom tooltip is added underneath
ctrlSetURL
Tooltip is not visibile by default but becomes visible if ctrlSetTooltip is used. Custom tooltip is added underneath
@solid marten
you mean you defined tooltip in config too?
I’m not sure what is the problem, both script setters should just set relevant config property with exception that url one checks if url is allowed
No, I just defined URL in config, not tooltip. I understood it would behave the same as using ctrlSetURL
I will create a repro later today.
you confused me. What tooltip is visible in config variant and what custom one is added if you defined no tooltip in config?
Ah ok. So if I use config property "URL" and nothing else. The tooltip displays the link. If I then use ctrlSetTooltip, the new tooltip is added on top of the already existing one.
With ctrlSetTooltip and no config property, the tooltip is not visible. It only becomes visible if I use ctrlSetTooltip. Which then also causes the custom set tooltip to be shown on top of the URL tooltip (which was not visible before)
you should see url tooltip when you set url with script or otherwise. does it not show it when you use ctrlseturl?
Nope, doesn't show.
It does, when defined in config. Not with ctrlSetURL
Fixed in 146861 @daring wagon thanks for reporting
Glad I could help
Given the fixes quickness, I wonder if there are unit tests on commands now :D
If you set the mission's name as "" in the description.ext, in local MP the empty string will be empty and you'll think the mission is not showing up.
does this classify as a bug/undesired behavior?
well you got exactly what you wanted.
And as long as the list entry is still there, and works. Sounds fine to me
is empty nicknames also allowed?
afaik no
ok
Is it a known issue that 1 call BIS_fnc_recompile; in 2.01 can crash the game? Happened to me twice now without any mods. I can provide reports and dumps if needed.
Is it a known issue that
1 call BIS_fnc_recompile;in 2.01 can crash the game? Happened to me twice now without any mods. I can provide reports and dumps if needed.
@untold sky ^^^
@daring wagon with -showScriptErrors yes.
Fixed a while ago, but dev branch was built just before I fixed that
to be clear, when its spamming errors in RPT, together with script callstack
ok
https://feedback.bistudio.com/T154721 can be closed
ticket for Edit Terrain Module doors lock not working for server clients:
https://feedback.bistudio.com/T154752
should be an easy fix.
also this can be closed: https://feedback.bistudio.com/T126757
Thank you dad 
https://feedback.bistudio.com/T154769 - Suppressors compatibility issues since 2.00
I think reyhard is already on that
bumping that one, freshly mentioned in #arma3_scripting
https://feedback.bistudio.com/T123883
enableWeaponDisassembly fudges the get in action / angle
Yup, I'm on those slots already
@gray wharf did someone check that the bug is still actual?
Well no repro on dev branch, nor internal...
wtf 🙃
this is really frustrating and drives a lot of players crazy
https://www.youtube.com/watch?v=b6bWaeSN9iM
Leave Feedback! https://feedback.bistudio.com/T119992
Arma 3 bug that was recently discovered. This bug is from the feature that makes you shrink into the ground in Arma 3. The player shrinks but the hitboxes do not shrink with the player. This feature also causes problems whe...
at distance the adversary sinks into the ground, but only visually. hitboxes don't. so when you think you're shooting into his head, you actually are shooting into his chest and if caliber is not high and the guy has armor, you won't kill him
yep, but not a bug - a design decision
I would agree that an improvement would be to have the hitbox move as well, but…
nothing is actually physically moved afaik. its only rendering
doesn't matter how it works
Does. Not a bug.
it's the result that matters
need to shoot above "visual" head to hit the "real" head (hitbox). so stupid
best would be a mission option to disable it if desired 👼
why not sinking hitboxes as well
also if you check the video until the end, you will see that when you have somebody behind a high fence, you can't see him, since he sinks partially into the ground, so is lower than the fence itself
BUT!
why not sinking hitboxes as well
because thats a completely different thing
he can see and shoot at you, since from his perspective he's not sinking into the ground
I see the problem with it. But won't fix.
so it's like cheating from devs. really bad design decision
it's either that or no more sinking into the ground
yeah was a bad decision 10+ years ago. and still is. But won't be changed now
maybe an MGS-like "transparency" could be an option - but that would be a big rework.
@gray wharf
https://feedback.bistudio.com/T84222
https://feedback.bistudio.com/T142928
https://feedback.bistudio.com/T139095
https://feedback.bistudio.com/T126951
https://feedback.bistudio.com/T121797
https://feedback.bistudio.com/T117665
https://feedback.bistudio.com/T116794
https://feedback.bistudio.com/T85836
https://feedback.bistudio.com/T85784
https://feedback.bistudio.com/T84963
https://feedback.bistudio.com/T84949
https://feedback.bistudio.com/T84908
https://feedback.bistudio.com/T84864
https://feedback.bistudio.com/T84433
https://feedback.bistudio.com/T84381
https://feedback.bistudio.com/T84361
https://feedback.bistudio.com/T84224
https://feedback.bistudio.com/T83565
https://feedback.bistudio.com/T81214
https://feedback.bistudio.com/T81202
https://feedback.bistudio.com/T74711
Can be closed 😉
can only close the first one :(
Okay, that's it then.
Requesting moar FT pawarrrs from BI!
send me a link to one that doesnt work
i guess the tickets you cant close are too old? Guess moderator permission was added later on so old tickets dont hav
Those he can't close are wishlist (marked in blue)
being able to close has nothing to do with priority tho
can't close this one either (not a wishlist)
maybe when it is wishlist or assigned
https://feedback.bistudio.com/T65678
its probably just age
ticket permissions are set when its created
maybe :u
I guess
yeah well, if I am "hired" to clean up and I can't, it kind of defeats the purpose doesn't it :p


I would appreciate it if a dev could look into that. It's an easy fix (really) and would make this attribute control useable. (Incorrect height set for control inside controls group)
fallen trees and fences were fixed in an appropriate manner that AI can't see and shoot through them, but they still don't have physics when fallen, so one can drive/walk through them and it was a genius decision.
no perf costs and fixes the problem.
so disabling visual model of adversary from sinking into the ground at distance, like it was in OFP/A1 is also a very very good solution that a lot of players would certainly have welcomed and not posing any problems in any of SP/MP PvE/PvP missions.
I think it would be so natural to do exactly this move
It is not a solution. I already told you. And i also already told you that won't happen.
And you were also already told that this is not the channel for that
well, there is a ticket for that, so this is the appropriate channel
it's certainly isn't offtopic at all
yeah was a bad decision 10+ years ago. and still is. But won't be changed now
nobody asks to face and solve problems related to physics and other possible problems, since "simply" disabling this from being done only for one part of the players but not for the other is a really good compromise.
since both sides would be able to see each other and 0 problems with AI as well. everybody would be happy
whether it is ideal or not (it isn't), whether you provide a solution or not, this isn't going to change anything in this matter
(((
accept the "no" on this aspect and carry on other issues - and it doesn't mean that we find the current solution ideal; just that it won't change
I'm certainly not a programmer, but I thought it would "simply" have required it not being done/happening, so "simply" disabling this...
I am a programmer and i already said no 🙃
man, with next title expect me to come and directly check things like this day one and then don't say first there are other more important things to fix and several years later, well it's EOL now, so don't expected it to be fixed now neither...
I regret not having switch from A2 to A3 directly in 2013
4 years ago when that ticket was active and we had tons of resources available, yeah maybe. but now its too late (even back then it was said "technical limitations, won't be able to fix),
I told you no multiple times, I've closed the FT ticket, now please stop beating on that, thanks.
would have reported a lot of stuff much earlier, when it still was actively worked on A3 by the team
@gray wharf you watch one punch man?
no
Oh okay 🤣
I would appreciate it if a dev could look into that. It's an easy fix (really) and would make this attribute control useable. (Incorrect height set for control inside controls group)
https://feedback.bistudio.com/T154820
@daring wagon Could you explain more what is the problem? Treat me like I know nothing about this.
Sure. I'll update the ticket later
I have updated the ticket with two videos and a bit more details @solid marten
@untold sky Your repro you requested https://feedback.bistudio.com/T154837
https://feedback.bistudio.com/T154841 Seems like the demine waypoint is broken. The condition is missing some ( ) from what I can tell.
doh, I will take care of it
Annoying bug:
https://feedback.bistudio.com/T154844
that's… what include does? and how is it "high priority"?
No it's not. The preprocessFileLineNumbers command works fine
how is it "high priority"?
No proper error reporting.
And this is highPrio because of?
not "high priority"
bugging, annoying yes
CTDs are high priority
??
How do you know that?
It's very easy to fix tho
I mean for the devs. Simple, all they have to do is put two#LINE 1s (one at the start of #include and one at the end of it). Similar topreprocessFileLineNumbers
quite sure I already had my own ticket with that line number issue.
We fixed it in ACE/CBA by moving where the include is done. Either before or after block comment don't remember where anymore
but I would even call that low priority
is there an EH/way to detect when VON gets activated (and log who spoke)?
if not, meaningful to make a FT request for it?
no
dunno how easy..... I guess it displays the name in bottom left right?
then yes, should be easy enough to be done. meaning FT yeah
what if include file contains more code and it errors there?
include file drops #line at start of it.
We just need to drop a #line of the parent file, after exiting the include.. I think
surely it is technically correct behaviour that cfgfunctions (preprocessFile) doesnt correctly count lines. otherwise it would use preprocessFileLineNumbers 🤔
correct
CfgFunctions was never supposed to have working line numbers, and thats not engines fault
We will fix that with a thing I'm working on
setSliderSpeed command causes the slider page step to become extremely small - https://feedback.bistudio.com/T154871
Is it intended that savingEnabled returns true even if the attribute configFile >> "Cfg3DEN" >> "Mission" >> "Scenario" >> "AttributeCategories" >> "States" >> "Attributes" >> "Saving"in Eden Editor is set to false?
dont think so
I guess I should make ticket?
https://feedback.bistudio.com/T154879
Seems also be the case with saving = 0 in description.ext.
so it seems to only check about enableSaving impact
yes
Fix already in dev branch
https://feedback.bistudio.com/T154021
Close ticket, please
closed
https://feedback.bistudio.com/T154878 @gray wharf Spam
@untold sky*
I can't delete 😬
you delete tickets by editing them, and removing the "view" permission for normal users
now you can
gone!
Is there a quick way to unsubscribe from mutiple tickets, or do I have to unsubscribe one by one?
https://feedback.bistudio.com/T84356 can be closed
https://feedback.bistudio.com/T84400 can be closed (fixed)
done
You were able to close them?
I iz powerfull nao
Lou is the fancy ticket man now 
before me were ticket men, but no that fancy ones 😎
take your time
Yeah I'm a bit slow sometimes 
you understand quickly! you only have to think about it a lot 😄
I iz powerfull nao
@gray wharf Yeah, but why are all the ticket I reported are not closed yet?
Good boy 😄
Once you are done, let me know I send you more
not sure I will then! 😄
Maybe I'll actually be able to see older tickets when some newer ones have been deleted 
I got my Dashboard set up to watch you. So I will know anyway
Maybe I'm using the command incorrectly on video in this ticket?
https://feedback.bistudio.com/T154810
Or is it a bug?
no no, at the very least a command should always have the same effect, not a random one like here - bug!
https://feedback.bistudio.com/T85377
https://feedback.bistudio.com/T85070
https://feedback.bistudio.com/T84940
https://feedback.bistudio.com/T84898
https://feedback.bistudio.com/T83720
https://feedback.bistudio.com/T78926
https://feedback.bistudio.com/T85723
https://feedback.bistudio.com/T85791
https://feedback.bistudio.com/T85886
https://feedback.bistudio.com/T86048
https://feedback.bistudio.com/T86057
https://feedback.bistudio.com/T86081
https://feedback.bistudio.com/T86181
https://feedback.bistudio.com/T86192
https://feedback.bistudio.com/T86290
https://feedback.bistudio.com/T86309
https://feedback.bistudio.com/T116783
https://feedback.bistudio.com/T116819
https://feedback.bistudio.com/T125232
@unborn acorn are you sure it's not just video settings? Perhaps you have reached maximum number of dynamic lights
oh! didn't know they were impacted by the dynamic lights setting. also, isn't it distance-based as well?
@sterile nova it was new mission with 7 AI-mens and me. Dynamic lights i have on Ultra
they dependent on distance, strength & video settings
smaller/low power lights disappear faster
All settings ultra
try to move closer to those AI which appear to be without light
can reproduce in VR
pretty violently, actually
(moving closer is not showing lights - even with only one on-screen unit)
so its some bug then it seems
yep
re-executing "forceOn" is working as well (I guess because of "chances" to get them on)
same in console or spawned btw
any other repro needed?
[] spawn {
{_x enableGunLights "forceOn"; sleep 1 } forEach (units group player);
};
``` turned everyone's lights on immediately 🤨
I can reproduce
ah
However, if I get closer to the unit, the lights turn on
With how many units have you tested it lou?
a group of 8, so 7 AI
also, it's a group setting if used on the leader - player enableGunLights "forceOn" results in everyone (trying to) turn their lights on
Please note that, this command does not turn on the player's flashlight
it's mentioned here it's for AI only https://community.bistudio.com/wiki/enableGunLights
Okay, nevermind. Now it doesn't work for me as well with 9 units
but if I don't use the leader as argument, it does not work? 🤔
it did for me
nvm, it works
https://feedback.bistudio.com/T79816
https://feedback.bistudio.com/T81203
https://feedback.bistudio.com/T83985
https://feedback.bistudio.com/T84856
https://feedback.bistudio.com/T84865
https://feedback.bistudio.com/T84967
https://feedback.bistudio.com/T85167
https://feedback.bistudio.com/T85274
https://feedback.bistudio.com/T86138
https://feedback.bistudio.com/T124726
https://feedback.bistudio.com/T139805
https://feedback.bistudio.com/T148504
https://feedback.bistudio.com/T154711
https://feedback.bistudio.com/T84953
@gray wharf These are the last ones for now 🙂
It wasn't, there is a solution in the ticket though. I don't think at this stage inheritance things will be fixed which go so deep
Although a command allExplosiveswould be nice similar to allMines
true! make a Feature Request ticket
Leave that ticket open then. I might use some of the infos there
Although it's so old I can't remember most of it 😄
woops, just when I closed it as won't fix xD
Sorry, my bad
one fresh FR ticket might be preferable? it would stand out
that's the plan
Awesome.
https://feedback.bistudio.com/T71017
to be closed?
https://feedback.bistudio.com/T78223 language @formal dome 👀 😄
this took me like 20 minutes to find the source from
like literally, discord did ping me but did not showed where
(muted this channel)
😱
https://feedback.bistudio.com/T154837 I have added a short video demonstrating the issue with CT_TREE. I might be able to provide that GUI if neccessary (dedmen)
https://feedback.bistudio.com/T154837
Damn I noticed this yesterday. I was forced to create a completely new "tree" using controlsGroups! 😣
I had like 10000 elements in it (config viewer), so it was getting very very slow (~15 FPS when scrolled to the bottom)
@untold sky
_d = findDisplay 46 createDisplay "RscDisplayEmpty";
_c = _d ctrlCreate ["RscTree", -1];
ctrlSetFocus _c;
for "_i" from 0 to 10000 do {
_c tvAdd [[], str _i];
}
no ping reee
You seem to have become so fond of these! 🤣
hard to find stuff that fits
KK, can you do a mass replace in the support module functions and change:
getVariable format ["BIS_SUPP_limit_%1", _type]
``` to
```sqf
getVariable [format ["BIS_SUPP_limit_%1", _type], 0]
It throws errors if those vars are undefined.
#arma3_scripting message
I can make a ticket if you're fine with it
@sinful kettle yes please, then tag me with it
https://feedback.bistudio.com/T154837 I have added a short video demonstrating the issue with CT_TREE. I might be able to provide that GUI if neccessary (dedmen)
@daring wagon This has nothing to do with ct_tree, this happens to any multi line control with too many lines, they are still rendering even if not visible. No provision for that was added when classes were written. It is a huge job to rewrite all those and even bigger job to test that new implementation is backward compatible. There are certainly other matters more important to look at.
https://feedback.bistudio.com/T145276 There a a few spam comments there it seems.
@solid marten I guessed that it has to to with rendering stuff which is not visible. I just added CT_TREE because it's a good way to repro it. I created the ticket because Dedmen asked me to do so.
Why would you post about dayz ticket? they have own moderators
@solid marten https://feedback.bistudio.com/T154904
"they have own moderators" not really, usually dwarden or me take care of these bots.
And he posts here because Lou can take care of it and Lou reads here, Lou is currently the only real "Moderator" on FT. The others are just QA who do it on the side when they happen to see it.
can someone with markman DLC source look for the script with STR_A3_Revive_MSG_INCAPACITATED_BY
it has a faulty "name _unit" check
it can show "any"
0 hits for "STR_A3_Revive_MSG_INCAPACITATED_BY" in any marksman or basegame scripts
p:\a3\language_f_mp_mark\stringtable.xml
401 <Key ID="STR_A3_Revive_Status_Incapacitated">
402 <Original>INCAPACITATED</Original>
411 <English>INCAPACITATED</English>
588 <Key ID="STR_A3_Revive_MSG_INCAPACITATED">
589 <Original>%1 was incapacitated</Original>
598 <English>%1 was incapacitated</English>
605 <Key ID="STR_A3_Revive_MSG_INCAPACITATED_BY">
606 <Original>%1 was incapacitated by %2</Original>
615 <English>%1 was incapacitated by %2</English>
622 <Key ID="STR_A3_Revive_MSG_INCAPACITATED_BY_FF">
623 <Original>%1 was incapacitated by %2 (friendly fire)</Original>
632 <English>%1 was incapacitated by %2 (friendly fire)</English>```
this is where the strings are defined
and it happens when you get killed using the A3 revive system
but the script/fsm seems not available in PBOs
you get the <no unit> output which is when name _unit is applied on dead
and logged to rpt
WARNING: Function 'name' - 1d3c1ccd280# 128279: if_soldier.p3d REMOTE has no unit
- network id 5:5
- person XXX
- dead
Cannot create object D:DDDD with type[AISubgroup], param[subgroup], NMT code[107]:
Cannot create object D:DDDD with type[AIUnit], param[unit], NMT code[112]:
any info available on these?
and logged to rpt
WARNING: Function 'name' - 1d3c1ccd280# 128279: if_soldier.p3d REMOTE has no unit
- network id 5:5
- person XXX
- dead
@desert trench this should not be happening any more, what exe version?
Quite sure the name fix is not in 2.00, it was in first prof changelog, which only listed differences to stable
This is why I asked. Bugs in stable could be already fixed in dev or profiling
mission.sqm/Mission/Entities/ItemDD/Entities/ItemDD.type: Unknown enum value Artillery
mission.sqm/Mission/Entities/ItemDD/Entities/ItemDD.type: Unknown enum value Demine
mission.sqm/Mission/Entities/ItemDD/Entities/ItemDD.type: Unknown enum value SeekAndDestroy
mission.sqm/Mission/Entities/ItemDD/Entities/ItemDD.type: Unknown enum value TransportUnload
anyone knows why these happen even when built with Eden? or what it actually means
Error in expression <7500*(power interpolate[0.1,1,0.4,1])>
Error position: <power interpolate[0.1,1,0.4,1])>
Error Undefined variable in expression: power
dont get how this errors out if the native definition also uses "power"
class CfgLights
{
class SmallFireLight
{
brightness = 376;//10;
intensity = "7500*(power interpolate[0.1,1,0.4,1])";//"3500*(power interpolate[0.1,1,0.4,1])";
};```
a missing include? thought maybe #arma3_config
@desert trench it all depends where do you use such effect I guess
https://community.bistudio.com/wiki/Arma_3_Particle_Effects:_Config_Parameters no idea if such page exist for light sources
its used by the flamethrower ammo
class LIB_Impact_Flame
{
class Light1
{
simulation = "light";
type = "SmallFireLight";
position[] = {0,0,0};
intensity = 1;
interval = 1;
lifeTime = 30;
};
with
explosionEffects = "LIB_Impact_Flame";
natively its used Land_Campfire_F and Land_MetalBarrel_empty_F
guess will try intensity or damage instead
yeah, power is exclusive property to "fire" simulation (like fireplace)
#define EVAR_SET_FIRE(XX) \
XX(burning) \
XX(power)```
Any update on when VON will be fixed? The only way you can hear players is by changing effects.
it was already fixed weeks ago
Is it only for the perf branch for now, or will it eventually be added to the main platform?
It will be in the next update
cool thanks
unfortunately no hotfix :(
wait there won't be any hotfixes before 2.02?
doesn't look like it
not planned no
https://feedback.bistudio.com/T83687 Can be closed. Thanks for that command =)
and thanks for the quick fix https://feedback.bistudio.com/T154841, can be closed as well.
*boop* done
Thanks
This one can be closed as well. Most of the stuff was cleared up, fixed https://feedback.bistudio.com/T85378 and the remaining issue is probably mentioned in a separate ticket (which I can't find right now)
https://feedback.bistudio.com/T71690 can someone close this one as invalid? Seems like I don't have permissions for that 🤣
done 😬
https://feedback.bistudio.com/T85927 can be closed?
Yes
https://feedback.bistudio.com/T150882 "funny", satchel charges still have some impact on unit reveal I think
yeah but I remember a ticket that fixed that (in A2?) already 😂
Technically makes sense, since AI is chasing the unit who "fired" the explosive, not the explosive itself. So in short; the AI should go into search mode when something explodes without knowing who fired it.
it never was fixed
in search mode sure, I agree
but not doing so if the unit is behind cover, not so much 😁
it was fixed indeed at some point
probably was forgotten to be merged to future titles
Cannot create object DDD:DDD with type[AISubgroup], param[subgroup], NMT code[107]:
Cannot create object DDD:DDD with type[AIUnit], param[unit], NMT code[112]:
getting quite a spam of these these days - possible to get info what they are about please?
Disengage
oukej: Cancels out Engage at will (makes the AI stay in formation and not pursue new targets (the one currently being engaged is not affected)). Subj. to merge with Return to formation.
from https://feedback.bistudio.com/T85378#1125881
worth making a FT request on this? it would improve and simplify AI commanding usability
Commanding menu should only show weapons and equipment nearby instead of showing all weapons on the map.
@daring wagon is there a separate FT task for this?
I was sure there is one, otherwise I make a new one.
@desert trench https://feedback.bistudio.com/T75458
IMHO there should be a radius set as limit for the rearm actions and the distance should be displayed next to the entry and they should be sorted (closest first9
https://feedback.bistudio.com/T83895 (This one can be closed (fixed))
Can be closed
https://feedback.bistudio.com/T154364
https://feedback.bistudio.com/T153399 spam @gray wharf
muh
😄
https://feedback.bistudio.com/T151119 @untold sky Was that related to the new topography command you added? If so, can this be closed?
(all done but that one ^)
that was fixed
closed then!
reyhard, I checked the current DEV and Structures (Livonia) -> Village - Wooden House (Small, v3) Land_House_1W03_F is still not fixed
despite changelog of 2.00 saying it was all fixed
https://cdn.discordapp.com/attachments/108187245529268224/710908155986706442/arma3diag_x64_2020-05-15_19-20-49-894.png
https://cdn.discordapp.com/attachments/108187245529268224/710908105294086174/arma3diag_x64_2020-05-15_19-20-58-701.png
still not assigned to anybody:
https://feedback.bistudio.com/T148411 - List of vehicles with commander turret not returned back by AI to vehicle's direction, when player switches to another seat
https://feedback.bistudio.com/T154433 - Tank AI commander with 0 ammo, makes player as gunner reloading 7.62 mm coaxial machinegun after each shot, if/after enemy AI get spotted
https://feedback.bistudio.com/T154445 - BTR-K Kamysh has no possibility for player as commander to cycle through different armament types (Ctrl+F), when turned out, unlike AFV-4 Gorgon with same turret/armament
https://feedback.bistudio.com/T154446 - Left Mouse Button and Ctrl+LMB don't work to manually fire or order the AI gunner to fire from 120/125 mm cannon, when player is turned out as commander (using Arma 3 Apex keyboard preset) in T-100 Varsuk and Rhino MGS (UP)
https://feedback.bistudio.com/T137283#2019162 - possible to see through deployed by Strider smoke when looking through Strider's windows
Not sure if it's helpful or not but I updated both configs to fix all the config errors + some other improvements
https://feedback.bistudio.com/T139827 (RscTestControlTypes)
https://feedback.bistudio.com/T154650 (RscTestControlStyls)
yes please ^
https://feedback.bistudio.com/T155096
"EntityCreated" mission EH request! 🙂
i thought that was already a thing
it might fire ALOT though. If you also count bullets (which are entities)
yeah… well is it a perf issue, and if so, what about limiting it to create* commands? objects, humans, animals, not particles/particleSources/bullets?
(PS, I see it a lot recently, but it's a___lot ^^ "alot" is not a word)
you can build it yourself via config init events
grrr, not script-wise!
you probably want a PostInit event then, because pre/init some things like setobjectTexture might not work yet
KK recently did stuff in that area, he'll know where to do it
go KK!
btw: can https://feedback.bistudio.com/T154215 be closed?
I just move things to Feedback state, while they are dev-branch or not released yet
because you want feedback, maybe people have issues, or it doesn't work as desired
invisible walls and glass, while on finished Industrial Shed everything is fine (at the end of the video)
https://www.youtube.com/watch?v=6Q7qpGbNn8U
one day i will finish this in form of a ticket
already half of objects/buildings checked
…I remember that script 😄
thx again for it - helps a lot
don't hesitate - also, I wonder if it collides with everything, e.g if sometimes grenades wouldn't pass while bullets would 🤔
bullets pass - yes, while grenades not
there is also a real problem for a lot of buildings, where for some of them grenades simply fly through glass of windows (without breaking it) and for other buildings, so the grenade can fly through a window, you need to break the glass by firing on it first, only then you can throw a grenade through it
so it's inconsistency - should be same for all buildings with windows that have glass, simply fly though glass or one needs to break glass first and not 50/50, depending on building
it's simply impossible to remember how it works for each single building and thus you reveal your position/presence or even die, if grenade falls under your feet, because it couldn't pass through the glass
it would be much much easier to remember on of both behaviors, if only one of them would have been applied to all buildings with windows with glass
Seems like menuSetURL is not working.
This works
// Create main folder
disableSerialization;
private _ctrlMenuStrip = findDisplay 313 displayCtrl 120;
private _indexMain = _ctrlMenuStrip menuAdd [[], "URL Test"];
private _indexURL = _ctrlMenuStrip menuAdd [[_indexMain],"URL"];
This doesn't add the second entry at all
// Create main folder
disableSerialization;
private _ctrlMenuStrip = findDisplay 313 displayCtrl 120;
private _indexMain = _ctrlMenuStrip menuAdd [[], "URL Test"];
private _indexURL = _ctrlMenuStrip menuAdd [[_indexMain],"URL"];
_ctrlMenuStrip menuSetURL [[_indexMain,_indexURL],"https://community.bistudio.com/"];
https://feedback.bistudio.com/T154395 Fixed
Closed
https://feedback.bistudio.com/T140183 this can be closed as duplicate of https://feedback.bistudio.com/T140973
done
@daring wagon and if you try setting "https://arma3.com/" ?
Also when you say doesn’t add second entry, what do you mean the code is identical to the one above except extra menuseturl
I noticed that the width of a controlsGroup's VScroll impacts the scroll speed.
Is it the intended behavior, or is it a bug?
nvm, my bad
@daring wagon and if you try setting "https://arma3.com/" ?
@solid marten Same result
Also when you say doesn’t add second entry, what do you mean the code is identical to the one above except extra menuseturl
Forget about that. After a fresh restart of arma, both codes add the entry. Just URL does not work.
https://feedback.bistudio.com/T85180 (Can be closed. There are many other tickets with more details about this issue)
https://feedback.bistudio.com/T120210 (Maybe reyhard can take a look at this if he has some time)
https://feedback.bistudio.com/T85180 (Can be closed. There are many other tickets with more details about this issue)
done
https://feedback.bistudio.com/T120210 (Maybe reyhard can take a look at this if he has some time)
@daring wagon I think you'll find that one's not an easy fix because it's by nature of how the game engine renders the viewPilot LOD on top of everything outside the model. You get the same sort of issue with attachTo'd objects
🤷♀️ No clue how it works or how much work it is, I can just forward it.
@daring wagon could you make a ticket with menuSetUrl not working repro please
Yeah, as da12thMonkey mentioned, bullet rendering is not really fixable via some model tweaks - some engine solution would be needed
as far as I know, BXBX "fixed" at some point in A3 rendering of bullets on infantry weapons
and that fix involved rendering of bullets behind the weapon - previously they would be rendered as a top layer
Thanks for the explaination reyhard. 🙂
https://feedback.bistudio.com/T120210 Can be closed (won't fix)
@solid marten Here is the ticket. I tried again this morning and had again the issue that the entry would not be create if menuSetURL is used. https://feedback.bistudio.com/T155161
R3vo I found similar performance issue to your issue in edit control.
I fixed edit control now so I'm sure I can do the same for your list stuff.
I got edit box performance from 180->24 fps to 180->150
sweet. No more game freeze if one opens BIS_fnc_arsenal in the functions viewer 😄
I didn't test vertical multiline rendering
😄
its very laggy 
fix fix fix ahhh
but yeah freeze is fixed, and selection behaviour should also be smoother. But rendering on that one I forgot
I am sure you can fix that one as well 🙂
Okey I found it, I'll fix that too. Probably the exact same thing as your issue thingy.
Good guy Dedmen fixing all the stuffs
Let me know when the fixes are live on perf branch. Then I will test all my GUIs and how they perform
next build, monday or tuesday. But not with your issue fix yet. I'll do that only next week
That's fine. It has been like that for 7 years. I can wait a few more weeks 😄
Hm, Multiline Edit should have check for that, but apparently its not working right
I like the color balance of that reply.
the orange and blue and stuff
https://feedback.bistudio.com/T155172 anyone know if this was reported before? I can't imagine this going unnoticed
ah indeed 😄
But yeah I can remember having that happen before when I was dragging objects around
it exists yeah
Ah I found the thing with vertical text R3vo...
They have a check for it.. but the height per line uses ui coordinates height, and the total height limit uses display scaled height
118 height fits 47 lines. so that means each line is 2.5 high,
but it thinks every line is only 0.03 high 🤣
😄
AH. I see the problem.
The text control doesn't know that its in a scrollbar, and thinks it has its full height
Its resolution is 2444x72089 pixels... oof
Oo
WTH
Very easy mistake to make.. But I think tons of other controls will also suffer from this
And I assume same issue for tree/listbox/combobox and others.
sooo perf++ 👀 😍
Basically, we can now load the full configFile in a tree view with addon icons and have an engine driven instant search.
tadaa..
it thinks its height is 117.8, while its being rendered in a scroll container who's height is only 1.5
I dont have a clue what any of that means but it sounds bad 
I just tested it and I don't remember that this has always been happening
always afaik
ok
Okey thanks R3vo, function viewer fixed too. Please don't give me more now, I need to get back to fixing actual bugs instead of performance stuff
CTD > bugs ←→ perfs I assume 🙃
problem is, perf is more fun than bugs
I'll try my best. Thanks for the fixes 🙂
Thank you Dedmen
Do performance! It is more fun and more noticeable... bugs we are used too... bad perf. we will never get used to... just accept it 😬
The performance isn't that noticable for most the players, except some few specific things.
that said, the bugs also aren't.
And fun is bad, fun makes me not wanna stop and work too much overtime :/
get paid for it!
"Many developers on Cyberpunk 2077 have been working overtime hours for months if not years now. But this call for six-day work weeks, "mandatory" crunch time."
Don't let those devs get all the crunch time! 🤣
Besides, you were the star on the last SITREP, those are big shoes to fill 👏 🤞 
@solid marten If I use an incorrect key combination with menuSetShortcut, it adds UNXPECTED_KEY_ID (missing an E) as shortcut text. I am not sure if that typo comes from the command or the menu strip control.
finally, I identified the UPnP issue:
https://feedback.bistudio.com/T155202
UPnP checkbox having no impact when hosting from Eden Editor
Would it be possible to add an Eden Event Handler (config and scripting command) which only triggers when Eden Editor is started? (From Start Menu, Multiplayer lobby)
@daring wagon Thanks will have a look
ThingX 2768fe16b00# 282271: dummyweapon.p3d was falling under terrain! Position corrected to ground [1698.88, 7.95, 1356.66]. Disable simulation to keep object under terrain.
we are getting this spammed in out rpt but my team says its not from IFA3. is this coming from A3 data/system itself?
Ref to nonnetwork object 284972: <no shape>
that one with <no shape> i also cant remember seeing in the past
Duplicate weapon Throw detected for SoldierClassName
Duplicate weapon Put detected for SoldierClassName
this one I also never could figure out what its about
yes arma
arna
ooh that's a hot name!
https://feedback.bistudio.com/T74615
Does anyone know if the module still does this bug?
Or is it superceded by the dynamic sim feature
Is it a bug that pushBack gets slower for arrays with nested arrays?
_someArray pushBack _arrayWithNestedArrays
the more times you try that the slower it gets
I don't think so
_a = [];
_b = [];
_a pushBack _b;
_b set [0, 1];
_a
gives [[1]]
well, when the nested array is (relatively) small: 6 ms
when it gets big: 270 ms
I can send you the code to test if you want
6ms? for a single pushBack?
no like 3000
I think
I know you might think : "why not append?" but I have a good reason
No I don't think that
yeah sure, have fun
_a pushBack _a
check
It gives me Recursive array error
ah I see. that's unfortunate
better check than BSOD 😄
Wouldn't comparing both argument pointers suffice?
So that's illegal because pushBack doesn't do anything with the element?
"big boom", if you like
I mean it could make a (deep) copy once before pushing it back, or not? In theory at least.
Or rather: It could have been designed and implemented that way.
backwards compat. not it cant
I'm mostly interested in the software engineering behind this.
So it was a design decision, I see.
also that would generally be expensive if you copy everything you pushBack
no it wasn't.
It wasn't thought of, and then ducttaped on afterwards when people's bad code caused problems and we couldn't change behaviour
Makes sense
so anyway, the reason my array gets big is that my mod has a "tree" view (which I made myself), and the elements are like:
[configName, configPath, Classes];
and classes is an array that contains similar arrays for the config classes of that config.
I think if I did something like:
treeLevels = [level1, level2, etc.];
and instead I changed my array to:
[configName, configPath, treeLevelIndex]
it won't be as slow right?
less nested arrays, more faster
hashMaps for that would be even better tho 🤣
omg.. you can nest hashMaps...
AHHHHHHHH
The editor displays vanilla units when searching in "Compositions (F2)" mod @'modName'
https://ibb.co/3pWWmky
did you type that in manually or select it from the arrow to the left? @unborn acorn
from the arrow
I tried it out today and it worked fine for me
https://feedback.bistudio.com/T155249
Anyone repro?
Can confirm, it's not working (in current Dev)
https://feedback.bistudio.com/T154366 Can be closed
set to feedback per usual dev process
I suppose "Closed" arrives on 2.02 ^^
https://feedback.bistudio.com/T152790 how realistic the idea is that these commands would be implemented to the game? With a quick thought the system would be similar to animationstate or animDone EH?
5 realistic
Isn't 5 enough? 😄
5/5 much realistic

that one was easy, tho 😛
I think Arma3RetailProfile_DX11_x64.exe hogs the performance
kill it and you're good
shred the file
and to think that all these lazy devs have to do is rename DX11 → DX12 to have all the latest DirectX support…
yeah, let's publish our own version of Arma
with blackj…
that screams of multithreading.
Well food is done now so back to lunch and then I'll fix that
something tells me your natural rythm is shifted by a few hours!
@daring wagon
https://feedback.bistudio.com/T154837
script like the tree view one but for listview? :prettyplease:

I retested your multiline test with 50k lines. Game froze for two minutes while creating the UI and text.. but...
Once its there it runs just fine 🤣
This is scrolling through 50k lines of text
that's weird. did you fix it? I can still reproduce this issue
you don't have the fix yet
dunno if I'll push this week. Probably wait for the other fixes too
what are your fps with 50k lines? 🤣
0.3? 
the old code also copies every line it renders 3 times or more
A bit more and I demand payment 😄
I thought you meant it didn't need fixing!
if they pay my bills

I'll do the listbox in a moment.
I only tested a few thousand lines. And I was getting ~10FPS!
72k pixels are hard to render 😄
btw Dedmen, that fix applies to all text controls in a ControlsGroup right? (RscEdit, RscStructuredText, etc.)
no
I fix stuff inside the specific control. Currently only CEdit is fixed
Tree and List come next, for the rest I'll need seperate repros to see what the issue with that specific control type is, they are all different
I think I have the same issue with multiline RscStructuredText in a controlsGroup
I use two overlapping controls (one RscEdit and one StructuredText) so I'm not sure
Feel free to add it to the same ticket with a repro
@untold sky Done https://feedback.bistudio.com/T155320
❤️ thanks :u
Listbox rendering with 40k entries is fast for me 150fps. So that one seems to either not have the issue, or my fix on Edit textbox control also fixed List, which is unlikely
I personally always had the feeling that listboxes were faster, but that's subjective of course.
the unit list tree on the side in 3den, takes longer to render than the listbox with 40k entries 🤔
Probably. They are both so small in impact that its hard to measure where a problem might be.
Besides that treeview one that I already know about
Btw, where is the perf branch update with all the goodies, it's Tuesday 😄
I want to finish this stuff first
btw, any way to adapt some UI elements to higher resolutions like 4K? I know you don't do UI itself, but e.g @ 4K or 8K the Eden cursor is very smol
…pwetty pwease? :3 😄
Guess it makes sense to put all the fixed related to UI out first
CT_CURSOR incoming ?:D
so we can scale it with pixelGrid
the treeview item adding itself, is only 10%
setData is 5.5%
that doesn't seem that slow to me, don't think I can do much on that.
The rest stuff is just script.
16% for creating that array, 13.5% for getting some variable.
hmm too bad
adding 800*40k tree items, is "only" 60 seconds
I don't think config is that big right?
probably more like 40k total?
With sub classes?
more
at least 190k
oh wait
TBH I can't tell. I mistook it with properties I search through in one of my GUI
thats still under 1 second to fill the tree itself.
with setData/setValue, maybe 2-3 seconds. Atleast with that tight loop from the benchmark
I can probably shave off 1-5% but.. not worth the effort I think. Your main overhead is the script itself
I guess it's best to wait until perf branch was updated with all optimisations. Then we create a test GUI and load the configFile into a tree view
Then we can see how fast it is
what would help would be something like a tvFill, give it one big array with all the name/data/value for all the items in one command call
but building that big array will then still be the bottleneck
I think the main problem is that filling that tree view would need to be spawned since a 3-4 sec freeze everytime when opening the config viewer is probably not acceptable for most ppl.
And if spawned I don't wanna know how long it would take
how often do you open the config viewer tho?
And you can do it over multiple frames. only do like 2k entries per frame
Add a EachFrame handler on open, and remove it again when tree is filled
you would want to block search though, while its doing that
with the new diag, just make a basic extension to open the AIO export with some external editor 😉
or that special engine implementation for that config tree
I personally would not enjoy that freeze at all.
Is it not possible to modify the search function so that only the currently visible level needs to be loaded?
(As I understood it the reason for loading the full config is searchability.)
you prefer a 5 second loading screen that prevents you from doing anything and a repeated loading screen everytime you click on something over a one-time 2 second freeze?
So searchability is not the (only) reason?
current config viewer doesn't let you search at all, even with the horrible loading screens
searching is not what's slow. Filling the tree is slow.
And if you want to search subclasses, you need the full information to be present
We could also use a little trick.
By default don't load all 13 levels but only like 8
Or just one level per frame? that would get worse the more levels you get tho
I personally prefer 2 second full freeze, over 3 seconds 2fps though.
You don't open config viewer that much. And everything is better than the current loading screen and having to wait 5+ seconds for stuff to become usable
Yes, that's why I was thinking that, by not properly filling the entire tree but just the currently visible level, we could boost that filling speed, trading it for a different search function that will then take longer as it has to crawl the config some other way.
I don't see any problem with just waiting for it to fill, if its really only a couple seconds
Can you "cache" that so that you only have to wait once per editor session?
We could also make it a setting. Either spawn or call, if spawn is selected we disable the search till everything is filled. If called then it freezes
The NVisibleItems thing seems easy to fix..
Its iterating through all items one by one, even though it just needs to pull the number of children, which it already has in a variable
The NVisibleItems thing seems easy to fix..
What's that?
TreeView rendering speed
But if my idea is not that viable, forget about it; having a config viewer that takes a few seconds to open and runs smoothly after that would be a big improvement, especially if it comes with search functionality.
The expensive part of rendering a tree with thousands of items, is actually updating the height of the scrollbar 
The scrollbar is updated and re-counts the number of tree items recursively every frame.
Could just cache the height, and only refresh when some item was expanded or added or removed, but thats probably too much effort
Unless I find a easy way to let the tree root know that something changed..
Though.. the root will know of item add/remove, and it knows of user input. Just refresh scrollbar on every click on the tree 🤔
hope you'll work on next title as well, to not let certain people write stuff the way it shouldn't be written
like it's the case in A3, where you have to redo a lot for it to work how it should have worked from the beginning, since it was/is possible and you prove/show it
@untold sky
17:44:05 191263 //Total count of classes
17:44:05 1.35498 //Time it takes to iterate through all of them when called
_time = diag_tickTime;
_counter = 0;
_fnc_goDeeper =
{
params ["_config"];
{
_counter = _counter + 1;
//systemChat str _counter;
//systemChat str (diag_tickTime - _time);
_x call _fnc_goDeeper;
} forEach ("true" configClasses _config);
};
configFile call _fnc_goDeeper;
diag_log _counter;
diag_log (diag_tickTime - _time);```
code used.
Not sure tree view is feasible for config viewer. I mean, this it not even tested with ace or cpu or rhs
so adding tree add onto that, 3x the time maybe?
Hard to say
"true" configClasses _config thats so bad, as it iterates over every item to call "true" is there no better way to get all subclasses?
I can probably make "" work as "gimme all" though
I'll test that and see what difference it makes
This command is probably the same, but the condition is optional
Maybe you can check what engine does in this command when the condition is omitted
lel
Mouse wheel scrolling in big tree's is also very slow because it also checks how many visible items the tree has.
It counts the items to calculate a value to check something and set a variable that..... is never used 
Were the tiny up/down buttons on top/bottom of the scrollbar always broken? they don't scroll 🤔
I think so
maybe that should be fixed 
i just clicked the down button, and it just reset me all the way to the top.. interesting
actually, I just tested one controlsGroup.They're a bit clunky but they work
clunky? I think it should just scroll by one item per click
controlsGroup
were you referring to something else?
Probably treeview has its own scrollbar?
yeah I think they do
Okey I got tree view with 50k items at 120fps
I think before I had about 5 fps with that
but after scrolling half way down, I'm back at under 0.1fps
actually I think tree view never slows down when scroll=0
not sure
it's when you go down that it gets slow
it does, see my results above.
every frame it iterates through all items (in a bad way) to count them to update the scrollbar
My basic fix was apparently good enough, I thought it needed more but that part i fixed atleast
and the drawing code iterates through items in the same bad way so I need to find a different way to do that.
and left clicking onto an item does the same bad iteration
not sure how the game does it, but this is how I do it in my tree view:
https://sqfbin.com/ivijetuloyulenaregey
I save all expanded tree branches.
when I have to update the view, I just use a simple for loop that only iterates in the view limit based on the height of the tree elements and scroll value
I'm not sure if it can be done in the engine though
It is super fast
no lag at all
I don't need to save/cache.
My new way to count items is very fast, and all it does is iterate the items differently.
I just need to adopt it to skip out of view items when rendering/clicking
is all of the chinese stuff discussed here only related to SP/editor or are there also benefits for MP for normal players?
We are talking about UI rendering
Including Zeus interface. But most normal players probably don't have list/tree views with hundreds of items on screen
(((
I should be able to get it into a profiling build late tomorrow. 
Atleast if the benchmarking tomorrow works out and I don't find anything else important that I need to fix
That's not fair! Dev branch should be updated first! 
https://www.youtube.com/watch?v=OiWY7OHexd8&feature=youtu.be First proof of concept works.
Only issue right now is the sorting of the tree view. A command would be nice which sorts a tree view including all sub trees similar to tvExpandAll vs tvExpand
Since the editbox is getting some attention maybe https://feedback.bistudio.com/T85981 would also be interesting to adjust
oooh yes :3
Someone actually started a "tree sort by value" implementation, but didn't finish it.
How would you want to sort exactly? by name? @daring wagon
Also one thing that bothers me is wrong double-click selection behaviour in edit box. You'd probably want the same for ctrl+backspace.
Also.. Today I learned, ctrl+backspace is a thing
Yes by name is fine
Lol, same here
Just tested it. Never heard of it before
But I also didn't know CTRL+ENTER in a browser would add .com to the adress 😄
thats probably browser specific
It is, just saying. There are so many shortcuts we probably don't know 😄
the ctrl+backspace is a tiny bit more work though, as I can't just go back to next word limiter ( .,-/...) but also need to skip all, till encountered a word and then find next limiter..
Actually that's easy with new stuff I implemented last week 🤔
But I think support for that is a designer decision
btw, Discord has an issue with Ctrl + ←/→ when there is a user @ tag
Ahhh more hotkeys
more hotkeys
to be implemented :D
sooo a Ctrl + ←/→ in a text field would be nice indeed :3
but how many people know about these shortcuts :U
Fixing double click behaviour should probably be first
but how many people know about these shortcuts :U
…plenty? just you two didn't 😄
You know what WIN + NUMBERKEY does?
aaah, the numbers above the letters
yes, starts the apps in the taskbar
I have my browser on Win+6, don't ask why
literally nothing happens on win 8.1 🤣
Might be a new one dunno
might have been introduced with Win10 yep
Tree fix benchmark failed btw. so maybe no prof today if I don't get it fixed fast
for those who don't know: Ctrl+Shift+V = paste without formatting - for when you copy a big text from a webpage into a Word doc)
Oh, this one is nice
Need to remember this when pasting stuff into word. It always ducks up the formatting
We need a new channel in discord #ShareYourShortcuts 😄
arma itself also has shortcuts that you can only know about when reading the functions :D
I only know ALT + F4 and that's my favorite one 😄
eg the current config browser does support some basic filtering when you just start typing on the listbox
Yes I know that one. It even shows what you typed in the top left
yes - I believe I only found out last year or so 😄
(almost not using the config viewer though)
Yesterday I got told that the config viewer automatically booksmarks weapon config and magazine config + something else if you preview a vehicle from within Eden Editor with it.
Might be a feature for your Config Viewer @finite anvil
bottom left yes - it changes when you change weapon as well
I never noticed, just wondered why I can't remove some bookmarks 😄
i only know from the config viewer function that it sets some default entries for you favourites
if you have none
You STUPID Arma UI that doesn't let y… ooooooooooooooh!
Yes.
Btw, how high is the chance to see engine search for listbox and maybe even listnbox?
R3vo what new tree commands you wanted?
only tvSort? if multiple then probably a ticket.
If you make me a ticket, and show me how config setup of a searchable Tree looks like (you have some entry that defines what the search textbox is) then sure
probably also want config property to choose whether only display text or also data/value thing
What list do you have tho thats big enough to need engine search 🙃
Like ACE Arsenal does just fine even with the huge magazines list
It's not about the size of the list
It's about convenience
With ctrl tree one just adds a CT_EDIT and an IDC to the tree view and boom
With listboxes it's so inconvenient.
only tvSort? if multiple then probably a ticket.
Only tvSortAll
Maybe ask @finite anvil or @tulip shard They do GUI stuff a lot too.
is there a tvSort already?
yes
It needs a path though
And for consistency tvSortAll is probably better since we also have tvExpand and tvExpandAll (The latter doesn't take path)
fine if I do no idc parameter?
no idc?
alt syntax
you mean no path?
I don't want to overhaul all the UI stuffz. Just looking for work I can do while waiting for stuff to compile
I made ticket for tvSortAll
non-resizing cursor would be neat :p
https://feedback.bistudio.com/T155341 Ticket for search
tvSortByValueAll too? oof
I can't think of any commands that I'd need atm. I am very happy to have a fast tree search :)
Well, I could think of a few more UI commands 😄
❤️
Thats 50k items.
No tree: 170fps
Tree open scrolled to top: 150fps
Tree open scrolled to bottom: 60fps
it still needs to do the calc to skip the items, so I can't do that much better without multithreading, and that would be sooo complex I'd rather not
but muh 120hz monitor :U reeee
60hz only 😢
tvSort is done, I'll do a quick copy-paste job for the ByValue and get profiling build done.
your listbox filter is no prio yet, but I'll make it a ticket
Awesome, much appreciated
Anything else I forgot? We'll probably see when profiling is out if there are still problems around
tvSortAll is slow, I can make it fast if needed. We'll see then
I guess it's still faster than looping through all levels with sqf
definitely.
But not as fast as using all your CPU cores to sort everything in parallel
dev of the year award 👑
https://feedback.bistudio.com/T155341 Here is the ticket in case you missed the message.
I saw it but now I don't need to scroll up :3 thanks
https://feedback.bistudio.com/T123362 any chance for this to be implemented?
I don't see a need
_unit modelToWorldVisualWorld (_unit selectionPosition "pilot")
is a thing. It would only be a slight perf improvement
https://feedback.bistudio.com/T139484 Is this considered intended behavior?
lol wut. Don't think so
might happen if vehicle player == player because code was executed too soon there
it is not affecting the driver. I don't think it affects gunners of turret either, but I can't remember properly. But passengers for sure. just spawn an empty vic and a soldier, get on a seat and execute vehicle player setVehicleAmmoDef 1 and bam, contents of vest and backpack gone.
do the same with setVehicleAmmo and it does not happen.
just did a test again with a hmg prowler and an ammo barrier: Gunner, commander and driver are not affected. As a passenger you lose contents except for faks? Even the gun disappeared.
but svda 1 should not remove anything anyway
Btw I think this ticket can be closed https://feedback.bistudio.com/T67343
Also
dedmen questioning why a ticket is still open in 2017, and its still open now https://feedback.bistudio.com/T65635
Why is this issue still open? Priority urgent, Severity major. For an issue that is only shown via unavailable youtube videos and that couldn't be reproduced by anyone shortly after it was created.
Yes please. 🙏
Can caching help?
yeah, but need to clear cache when any item is expanded/unexpanded/added/removed/reordered or scroll is changed
I think this ticket can be closed aswell https://feedback.bistudio.com/T66260 , the hunters damage model seems pretty good to me (from memory) ( @gray wharf )
aswell as https://feedback.bistudio.com/T64729
done @restive bone
Thank you
(Also this one I mentioned earlier #arma3_feedback_tracker message )
Also https://feedback.bistudio.com/T129139 , it was already marked as wont fix for a bit but got reopened
all done
Huh this would be nifty https://feedback.bistudio.com/T128495
@gray wharf You develope now? 😄
Just reading up and noticed tvSortAll/tvSortByValueAll - is that to replace having to do
{
_people tvSortByValue [[_x],false];
} forEach [0,1,2,3,4];
Does tvSortByValue sort every sub tree from [_x]?
No, I'd have to have a path greater then just _x so like [_x,0] etc, however my tree doesn't go that deep
my tree doesn't go that deep
If that's the case, then yeah.
Because tvSortByValueAll sorts the entire tree view.
Speaking of which, wouldn't it have been better if both of these commands also accepted a path and sorted the tree from there?
Like
_tv tvSortByValueAll [[0],true];
to sort all sub trees starting from [0]
and:
_tv tvSortByValueAll [[],true];
to sort the entire tree (what the command does right now)
Better maybe, but work, and if we do that, we'd have to do the same with expand/collapse
if I find time tomorrow I can modify the command to take array instead
Yeah thats a good idea, saves a lot of faff with getting certain bits to sort. Maybe tvSortAllByValue tho 😉
I might have found a bug with https://community.bistudio.com/wiki/lnbSetPicture
It seems that once a picture is set it cannot be removed by ctrl lnbSetPicture [[_row,_column],""];
However I can replace it without any issues if I set a different texture.
Can anyone confirm that?
oh now you remind me: CT_COMBO can not be unselected with lbSetCurSel -1 like the listboxes can

Can confirm.
https://feedback.bistudio.com/T155366 Ticket for lnbSetPicture bug. If more information is needed ping me.
@gray wharf Spam https://feedback.bistudio.com/T155362
https://feedback.bistudio.com/T123348 Last note is spam as well
I wonder what the spammers intention is
None To click the stupid links. But the spammers are "bots"
I know its bots but I'd think the bi feedback tracker isnt really the first thing to come to mind when you set up a spam bot 
done (and disabled said users)
Good guy Lou
as I said, bring'em!
sorting 50k tree of random sorted config classes (repro from R3vo) is ultra fast. Even in debug build its just "blip"
though thats only one-level deep tree. but should be fine
Very nice.
Here's another bug I found 
I saw, I'll do configClasses first
Before:
classes: 189508
Time: 3.216
After:
classes: 189508
Time: 1.89301
just added check if condition == "true" then don't run any SQF and just pass everything straight through
nice optimisation.
I thought I'd do if condition is empty. but that changes behaviour from all true to all false
Is it faster when the condition code is a string? (I'm assuming the answer is no)
If so, why isn't that a code in the first place? Like:
{ true } configClasses configFile
it makes no difference
but writing the code is easier
because of syntax highlighting
But its effort to add new syntax, and i couldn't do this optimization that easily with that
On thing I just saw in ACE code, they do "isClass _x" configClasses which is useless as configClasses only returns classes
I wondered too, why string instead of code, if it is to put code in it?
compile usage would add a little overhead no?
a tiny bit yes.
It takes string because its old
A3 v1.24 🤨 it's not like it's ArmA 😄
I dunno, but I'll keep it that way
unless its super easy and we have a StringOrCode type
event handlers support both don't they?
We don't
addAction takes both?
you can make it a feature request for some other day if you need it so dearly. But that will not have my optimization probably because it would be quite a bit harder to do, so I don't see the sense
more in the "nice to have" column
yeah. not really high priority for me
I could do other hardcoded optimization, like I saw in ace getNumber (_x >> 'scope') > 0 but really, that's not used often enough to be worth the specific inclusion as a hardcoded condition.
"getNumber (_x >> 'scope') > 0" configClasses (configFile >> "CfgWeapons")
Before:
8.564ms
8.529ms
After:
1.692ms
1.579ms
maybe worth it afterall
"true" configClasses (configFile >> "CfgWeapons")
Before:
5.8655ms
After:
1.517ms

noice!
configProperties too?
yeah that one! We ❤️ you! 😉
oof
when you leave the optional condition out, engine will automatically insert a "true" and run the useless condition everytime :/
Had slight hope that someone maybe thought of that and optimized that case
configProperties [configFile >> "CfgVehicles", "true"];
Before:
37.037ms
After:
13.263ms
33 minutes for these 3, I consider that fast enough 
you have a ticket for that yet?
i dont think there is (yet). I'll make one
I'm still on lnbSetPicture for 10 minutes 🙃
https://feedback.bistudio.com/T155369 after testing it again i found out it actually selects the first entry of the combo
Is unselecting a entry in a combobox valid?
doesnt combo always have a value as soon as some entries are available?
[] spawn {
_d = findDisplay 313 createDisplay "RscDisplayEmpty";
_c = _d ctrlCreate ["RscCombo", -1];
for "_i" from 0 to 50 do {
_c lbAdd format["Entry %1", _i];
};
};
``` will leave you with an empty combobox (nothing selected)
huh, interesting, weird.
Other UI programming stuffz usually select the first element by default
But if Arma allows that... okey.. lbSetCurSel shall do too (unless I find code that's specifically written to not allow it.. in which case lbAdd should default select first element)
https://github.com/7erra/Terra-s-Editing-Extensions/blob/master/TER_editing/gui/scripts/TER_3den_RscDisplayConfigViewer73.sqf#L845 this is a workaround that i currently use
a note on the lbSetCurSel page:
Posted on November 29, 2014 - 23:40 (UTC)
DreadedEntity
Prior to ArmA 1.90 lbSetCurSel -1 had no effect while the listbox was populated. You needed to use lbClear first, then lbSetCurSel -1, then re-populate the listbox.

for the combo or the listbox?
combo
i suppose listbox can select -1?
yeah
Combo has special implementation that doesn't allow -1 :u
so like I thought you aren't supposed to have no value selected.
So I should really fix lbAdd on combo not setting selection to 0
but Mr. Backwards compat doesn't like that idea
Why do you need to unselect?
the config viewer can favorize config classes but if the currently selected config is not a favorite no entry is selected. ofc i could add a dummy item like "--"
from my experience with combos it is more often the case that i run the lbSetCurSel 0 command afterwards anyway
I would say either won't fix, or add special case for -2 means unselect in combo, to not fix back compat stuff.
but a dummy item is the most logical solution, as most other languages also don't allow no selection in combo
so tell me if you're fine with "won't fix" and can just do your dummy entry 👀
