you can add scripted EH to the vehicle
#arma3_scripting
queen cargo
getIn
tht triggers whenever a unit gets ijn
not sure about what you try to archive
but that should get you some steps closer to your goal as long as you understand some SQF
astral tendon
is this
helicopter is called med1
medi once there is a player inside cargo must go to a landing area
and repeat the cicle all over
{isNull gunner _x} was working
but i dont wanna the player have to be in the gunner seat to be trasported
{isNull gunner _x} count [medi1] == 0
this is the full activation
and {isNull cargo _x} does not exist
astral tendon
crew counts evreyone
queen cargo
<CODE> count <ARRAY>
_obj = gunner _x; {_x != _obj} count crew _x > 0
whoops
what is that
magic 😱
hf 😃
queen cargo
yes
{isNull gunner _x} count [medi1] == 0 is bs
isNull medi1 would be sufficent
private _obj = gunner medi1; {_x != _obj} count crew medi1 > 0
use that
astral tendon
now the tigger activate when there is not player inside
also, there is the AI pilot inside it
queen cargo
add more exceptions
alternatively you also can remove gunner etc. from the crew
think there was also a passengers command
astral tendon
i made it work
private _obj = gunner medi1; {_x != _obj} count crew medi1 > 1
that 0 in the end was what?
astral tendon
im gonna use that one you give me, thanks.
still forum
@queen cargo a object variable contains a pointer to a pointer. If the unit is deleted the second pointer is a nullptr.
queen cargo
but not in the SQF-VM
runic surge
if (_grpSide != west || east || civilian) then {_grpSide = resistance;};
Since there is basically no consistency with what is used to refer to the independent side, I am trying to use this to make sure that the player side gets set to resistance. It says generic error in expression before the first ||
little eagle
my
still forum
that's not how "or" works
you are doing
(_grpSide != west) || (east) || (civilian)
Which doesn't make sense
little eagle
if !(_grpSide in [west, east, civilian]) then {
_grpSide = resistance;
};
But that still is pretty nonsensical.
still forum
As you can see https://community.bistudio.com/wiki/a_or_b only accepts booleans.. you are giving it sides
It also doesn't give you a generic error in expression.. It shows what the real error is just above...
little eagle
Since there is basically no consistency with what is used to refer to the independent side
Yes there is. It'sresistanceandindependentjust like how NATO is onbluforandwest.
runic surge
the strings return "GUER"
runic surge
so it messes up my script format and complicates things
runic surge
it only works if I set the default value to be resistance
little eagle
Then fix your script.
private _fnc_serializeSide = {
#define SIDES [west, east, independent, civilian]
#define SIDES_STR ["WEST", "EAST", "INDEPENDENT", "CIVILIAN"]
params [["_side", sideUnknown, [west]]];
SIDES_STR param [SIDES find _side, "SIDEUNKNOWN"] // return
};
call compile str resistance // any
call compile (resistance call _fnc_serializeSide) // GUER
I posted this 3 times this week already. D:
robust hollow
😱
runic surge
I didn't know you could modify values like that
little eagle
I don't get the problem. This is very easy to do.
still forum
There is no value being modified in that script...
literally...
also.. @little eagle isn't the comment on the last line incorrect?
little eagle
How so?
little eagle
GUER IS independent
still forum
The function only returns "WEST", "EAST", "INDEPENDENT", "CIVILIAN", "SIDEUNKNOWN"
It can't return "GUER" or GUER
runic surge
so it doesn't change what the green side string is?
little eagle
GUER is how independent is represented.
still forum
I just assumed the comment is telling the returned value
little eagle
It does.
call compile (resistance call _fnc_serializeSide)
GUER
call compile str resistance
any
still forum
I see....
little eagle
Don't tell me this is too difficult for you too^^
still forum
It is
runic surge
params [
['_grpType',configNull,[configFile]],
['_grpCost',0,[0]],
['_grpSide',sideEmpty,[civilian]]
];
if (_grpSide == civilian) exitWith {hint "You cannot recruit civilians! That's a war crime, you absolute cunt";};
private _p_balance = [WEST_balance , EAST_balance, GUER_balance] select ([west, east, resistance] find _grpSide);
If I play as independent it always uses the default value, which in this case is sideEmpty, because the "GUER" string is what is being passed into the function
still forum
I already closed everything besides discord because I wanted to go to bed
just make the default value be resistance instead of sideEmpty
if you already know that sideEmpty can't happen as you always pass in a valid side
little eagle
There is nothing wrong with this script. The issue must be in how it's called.
still forum
yeah. He posted earlier today. He is calling this via a formated string
runic surge
Command is formed with the format string
little eagle
You can't do that. Change it.
If you really really have to "format string" (serialize) the side, then use:
private _fnc_serializeSide = {
#define SIDES [west, east, independent, civilian]
#define SIDES_STR ["WEST", "EAST", "INDEPENDENT", "CIVILIAN"]
params [["_side", sideUnknown, [west]]];
SIDES_STR param [SIDES find _side, "SIDEUNKNOWN"] // return
};
call compile str resistance // any
call compile (resistance call _fnc_serializeSide) // GUER
And it will all work out and there will be no problem.
...
runic surge
I kind of need to since it is a diary record and the part that calls the function is dynamicish
Anyway, thanks for your patience
little eagle
Yes and I posted how you have to do it
runic surge
Where should that code be?
runic surge
because I'm retarded I don't know what to tell you
little eagle
Post the line where you 'format string' the side
runic surge
could I post it in the init to be used in all of the format strings? There are several of them
actually nvm I can just post it where all the records are created
little eagle
... ?
astral tendon
is there a way to turn off a trigger? like, i dont wanna hes conditions to be meet until another trigger activate first
runic surge
triggerActivated triggername
or set a variable to true when the first one activates
use that as the condition
astral tendon
how i set a trigger to false?
runic surge
not sure what you mean
astral tendon
you said about set a variable to true?
runic surge
In the init or a game logic, put myvar = false and in the first trigger myvary = true on activation and the second trigger put myvar as the condition
second trigger won't activate until the first one
runic surge
yeah
astral tendon
in case of triggerActivated triggername && triggerActivated anothertriggername
i nedd to activate the 2 triggers or 1 at time and still works?
astral tendon
{isNull gunner _x} count [car1, car2, car3, car4, heli1, heli2] <= (4 - ({alive _x} count playableUnits)) max 0
this condition is supouse to work only if there is 4 gunners in total in those units but only alive players/AI slots
this is not activating
any idea?
they need to activated at same time?
robust hollow
// if there are 4 gunners this will return 2
{isNull gunner _x} count [car1, car2, car3, car4, heli1, heli2]
// if there are 4 playable units, this would be 0
(4 - ({alive _x} count playableUnits)) max 0
so your basically doing 2 <= 0;
2 is never less than or equal to 0
astral tendon
and if i put >=?
robust hollow
do u want it just players as gunners or any player or AI, as long as there are atleast 4 gunners?
robust hollow
then u could just do
{!isNull gunner _x && {gunner _x in playableUnits}} count [car1, car2, car3, car4, heli1, heli2] >= 4
which is if the gunner is not null and the gunner is a playable unit.
astral tendon
about the trigger, they need to be activated at same time or one time only?
runic surge
I am just going to start over with my better understanding of things now. I hopefully will do things right this time
astral tendon
{!isNull gunner _x && {gunner _x in playableUnits}} count [car1, car2, car3, car4, heli1, heli2] >= 4
this did not worked
3 AI gunners in and 1 player in and still not firing the trigger
little eagle
Well yeah, this only is true when 4 players are the gunners, not AI
robust hollow
playableunits includes playable AI?
astral tendon
i find another way around
quick question, if i add a DLC prop, units, vehicle people need to own the DLC to play my mission?
rancid ruin
i expect it will add said DLC to the requiredAddons list in your mission file so yeah
robust hollow
no. You can play still play the mission but if you use weapons or vehicles belonging to a DLC u dont own it will have watermarks (and maybe wont let you use some things, im not 100% certain on that).
waxen tide
is the biki dead slow to anyone else?
vague hull
dead, yes
vague hull
google cache?
tough abyss
Both the forums and biki are down.
waxen tide
stares at a wall, motionless
little eagle
What do you need?
waxen tide
a working biki.
waxen tide
😗 thanks commy2
rancid ruin
just use google cache like senfo said
or you could wget the whole wiki the next time it's online, i did that a year or so ago
waxen tide
somehow google cache isn't working out for me on biki.
halcyon crypt
rancid ruin
same
austere granite
Anyone aware of a way to figuring out if PIP setting is turned on?
rancid ruin
@Adanteh#0761 isPipEnabled?
austere granite
that works. Guess i sholudve checked the archive page
shut flower
Does anybody have a clue how I could get the path of the surface texture? With surfaceType I can get the surface type. When I remove the leading # I get the config name in CfgSurfaces. But there the paths to the textures are not defined, there is a wildcard for the file name. When just using maps from A3 I can parse this and get the path of the surface texture (\a3\map_data\<tetxure>_co.paa) but this doesn't works for community made maps. Theoreticially the absolute path has to be defined somewhere. But where is it defined?
tough abyss
Why you want the path ? Its prob baked in the wrp or rvmats for the surface mask.
austere granite
One example could be having custom fortifications with a hiddenSelection for the dirt mound, which you would then texture according to the surface it is placed on, by retrieving the surface texture for it.
Anyway from a terraijn making perspective you never specify the path in something that you would be able to read through scripting. The full path is specified in layers.cfg, but that gets baked in
polar folio
does anyone know if teh NVG goggles overlay is a dialog control or display that can be accessed?
austere granite
looking at my terrain files and scripting i dont think it's possible @shut flower
i think you'll have to go the route of either creating some config entries yourself for each terrain or using a switch for worldName and just looking up the paths that way
It would require you to manually add support for terrains, but dont think you got another option
astral tendon
thanks for point me out what is it
still forum
peak plover
hi, can I script here?
shut flower
Yes you can script here
astral tendon
script machine broke
simple solstice
special someone told me that you would like this
tickTime = diag_tickTime;
diag_log format["Time to complete: %1 (in seconds)",(diag_tickTime - _tickTime)];
that returns
"Time to complete: scalar NaN (in seconds)"
tame portal
@simple solstice its because the first "tickTime" is missing an underscore
simple solstice
I know that, but people were looking what causes a NaN error(?)
simple solstice
yup, x39 was looking for NaN errors 😃
little eagle
It's not really a NaN value though. It's a nil with the NaN type. Because all math commands can return either NUMBER (scalar) or NaN (NaN scalar).
All the math commands in SQF are set up to report NaN by default.
Other commands that can't report NaN, but only NUMBER do this:
str [_undefined ammo _undefined]
"[scalar]"
@queen cargo
little eagle
So careful, scalar NaN is a nil value and not a NaN value like 1+INF or QNAN
[nil, nil ammo nil, + nil, sqrt -1]
[any,scalar,scalar NaN,-1.#IND]
damn copy paste
peak plover
ugh...
little eagle
[nil, nil ammo nil, + nil, sqrt -1] apply {typeName _x == "NaN"}
// [bool,bool,bool,true]
[nil, nil ammo nil, + nil, sqrt -1] apply {isNil "_x"}
// [true,true,true,false]
all these typos, urg
peak plover
How can I setVariable on before initPost
little eagle
By doing it?
still forum
at init.
peak plover
I guess I might jsut add a sleep to my function that requires the variable
peak plover
I might just have to put a sleep for ~1 sec and then loadout
probably best to do a little randomization for the sleep as well, as loadouts might cause fps drops
little eagle
setVariabe public is broken for the object init events. It doesn't synch, no matter if the object is created at mission start or during the mission. And that is one reason why InitPost exists in CBA.
astral tendon
well, im having trouble with AI (once more), now i dont know what i did but they are not following correctly the leader of the convoy
they just stop in the middle of the road sideways and do nothing
any idea how to fix that?
little eagle
Delete them and create new ones?
astral tendon
by doing that im trowing a week of work to the trash
peak plover
That's impossible
You already worked, so you already spent it. Can't get it back and throw it away.
You could however throw the result of your work in the trash
they just stop in the middle of the road sideways and do nothing
Are they grouped together?
astral tendon
yes they are in the same group
actually, one of they stop in the midle of the road stoping everyone
peak plover
hmm
runic surge
is there any way to prevent an object from jumping up when being detached from a parachute?
I have tried several things, mostly attempts with enableSimulation and setPos
but everytime the object gets detached, it appears a few meters above where it was and falls to the ground
nvm setPosATL seems to be the way to go
astral tendon
threre is a command to make enginne invunerable?
tough abyss
Anyone who can help me regarding to script a weapon I'm making?
runic surge
you need to be more specific
tough abyss
well, I want to know what I need to do
My modelling artists are setting up the weapon model
how do I actually make it work
assign a sound to it
assign the animation they're making*
runic surge
You should do some research before coming here, since that AFAIK is pretty basic stuff, but if you have trouble, you should ask in #arma3_model
that is all just the config and model.cfg work
austere granite
@astral tendon No, you'd have to use handleDamage eventhandler instead
peak plover
Is there a way to avoid doing private do get to the previous scope?
because I'm too lazy to private ~20 variables
_var 1 = 1;
call {
_var2 = 2;
_var3 = 3;
};
_varTotal = _var1 + _var2 + _var3;
How to I do this with the least amount of effort?
I might just do global variables ...
peak plover
I'll try using global variables
and I used allVariables command to grab the varaibles
So I'll reset after every use
I hope this won't be slow
jeez. this all seems stupid..
little eagle
You're drunk.
peak plover
Nop.
little eagle
Is there a way to avoid doing private do get to the previous scope?
???
little eagle
Yes.
Well, solution 1:
_var1 = 1;
_var2 = 2;
_var3 = 3;
_varTotal = _var1 + _var2 + _var3;
Get rid of the scope.
peak plover
I need it because I'm loading a file
little eagle
"loading a file"
peak plover
call compile preprocessFileLineNumbers _factionFile;
private code = call compile format ["loadout%1Code",_role];
little eagle
🤔
peak plover
I'm adding a file that can be anything and then I'm getting a value that can be anything
Unknowns
little eagle
I'd rethink the structure of what you're doing.
peak plover
Before I had a huge array of private variables, but that ment if I ever wanted to add something I had to add another variable to private array
little eagle
Basing your code on carried over locals is really bad.
peak plover
Now I will just use global variables instead
little eagle
but that ment if I ever wanted to add something I had to add another variable to private array
Yeah, that's why we have the keyword now.
peak plover
Well now I can just _unit setVariable ["unit_loadout_role","ANYROLE"]; and it will get it from the proper file
little eagle
Object namespace variables are always a good idea.
peak plover
When I last tested, I think with 32 bit game
Took many millions variables to crash game
So it should not be an issue
Either way I got a script to delete everything
little eagle
Eh, all the variables you can think of will only ever make up a tiny fraction of what a single texture will eat of memory.
peak plover
I'm also going for more of "First do it, then optimize it"
Seems to be working out nice
Havind doubts 'tho:S
little eagle
It's not what I do, but the more "professional" programmers I know talk sometimes about that approach.
Very rarely do people actually use techniques to optimize code in SQF. It's more like rewriting it to compensate for the lack of a compiler that would do that stuff for you.
And stupid arguments like count vs. forEach.
peak plover
Yeah, while in reality all we need is 200 fps server 😂
I wonder if that benefits scripts?
In theory it should give more time per second to the scheduler
So scheduled scripts should run better...
little eagle
Quantum computers.
peak plover
Ok, in theory there are way better and faster script engines / languages... so there's no need to push hardware until we've peaked software wise...
little eagle
Don't think the problems of Arma with performance come from SQF.
peak plover
What's the biggest killer?
little eagle
I have no idea, maybe dedmen knows. I suspect animations and PhysX.
peak plover
ai...
little eagle
Eh, not sure. Many soldiers also means many animations going on.
peak plover
I think what's also a huge part of understanding is comparision to other games...
little eagle
To test one would have to place like 100 v 100 VirtualMan_F against each other.
Those have no model.
peak plover
VirtualMan_F The what?
little eagle
It's just like B_Soldier_F, as soldier classname, but without model.
No model = no animations. In theory.
peak plover
for "_i" from 0 to 1000 do {private _guy = "VirtualMan_F" createVehicle position player;}
//3fps
stabilized to 15 in a few sec
little eagle
You need createUnit
peak plover
12 fps with 100 createUnit
waxen tide
"Yeah, while in reality all we need is 200 fps server 😂 " That guy is so annoyingly stupid ...
runic surge
is there a way to set the high command commanders entirely through scripting?
waxen tide
The eden editor has a function to save compositions. they're saved in an extra file each, sqe format, and the contents look like this: https://pastebin.com/xtAG5CLi
i would like to use this format to save objectives i spawn dynamically by scripts during the mission. but i'm a bit clueless how to interact properly with these classes? i mean i know the basic concept of classes i've just never used them in sqf and my google fu failed me. maybe someone can point me in the proper direction. 😳
robust hollow
include them in ur description.ext (or config.cpp if ur doing a mod?).
for spawning them id just use soemthing like forEach ('true' configClasses <configDir>) to loop through each item in the composition.
waxen tide
it is just a mission, not a mod. okay i'll play around with configClasses. but it doesn't seem there is any real advantage in this data format?
i guess the advantages of classes (inheritage?) can't really be utilized in that case
robust hollow
its good for readability i guess. i tend to use configs instead of arrays of data cause to me its easier to understand getNumber(missionConfigFile >> 'myClass' >> 'myNumber') than _array select 4 or whatever
waxen tide
fair enough
tough abyss
getNumber(missionConfigFile) can be replaced if they fixed it using the BIS function.
@waxen tide
waxen tide
by what?
tough abyss
Less messy.
or
Not sure if it still works though.
It had issues with some config files I used #include config.hpp to use seperated config files.
Had issues where it wouldn't pick up the #include headers copy placed into the file
waxen tide
i'll check that out, thanks!
tough abyss
np.
rough heart
DutchMan.Hx - vandaag om 09:28
i never saw that channel
btw
quick Scripting question
i wanna have AI follow into a Player vehicle
multiplayer session
i have a netID
Major Mittens - vandaag om 09:28
Lol go #arma3_scripting 😛
DutchMan.Hx - vandaag om 09:29
Do i need to use the Netid from the player to dertermine what kinda vehicle it is or get the NetID from the vehicle so that i set the AI to follow it
tough abyss
backtick backtick backtick sqf
rough heart
so, all sorted
tough abyss
If you wanted to post the code you've got so far.
tough abyss
AI is handled server-side.
rough heart
I know
tough abyss
So making them follow the player isn't too hard. Even in MP
tough abyss
Do you need to know what kind of vehicle the player has?
rough heart
cant post the code in here
tough abyss
codeshare?
rough heart
some like that
tough abyss
tough abyss
Just waiting for Atom editor to install SQF syntax highlighter.
afk I've got do go do somethings.
rough heart
it exists
tough abyss
Okay so it's a flow logic error.
rough heart
it follows the player
tough abyss
But?
rough heart
but i need to let it check if the player is inside a vehicle and if
let it get in the car with the player
that it is following
but
with whatever i tried untill now it didnt work
so i took a snippit from a other script that i knew worked in the past, not sure if it still does but anyway, i figured adapting it to get the vehicle from the netid or something would be sufficent enough to make the ai stick to it and get in
runic surge
waitUntil {!isNil BTH_RTS_MissionInitComplete}; Shouldn't this wait until a variable called BTH_RTS_MissionInitComplete is defined? It gives me an undefined variable in expression error, which doesn't really make sense
nevermind it is supposed to be a string
btw that group spawning issue I was having a couple days ago is something I just decided to approach differently.
BTH_RTS_TEAM_BLU_Grp_List = [
[["B_Soldier_SL_F", "B_soldier_LAT_F", "B_Soldier_GL_F", "B_soldier_AR_F"], 550, "Fire Team"]
];
An array with a list of units to create a group from makes things much easier and allows much more customizability
lavish hamlet
i need to run a line of code whenever a player removes his backpack.. how do i detect this event. ?
astral tendon
You guys know any way to disalbe fiendly fire?
distant egret
@astral tendon like no getting damage or preventing friendly AI shooting you?
astral tendon
no get any damage by fiendly
still forum
@peak plover true. The more fps the more scheduler time. so scheduled scripts finish faster. That's about the only improvement I am sure about.
And yes.. There are faster script languages.. #Intercept
Biggest part of Arma frametime is usually simulation and rendering.
@runic surge The syntax is isNil "varName" not isNil varName
little eagle
Does simulation include animation and physics?
still forum
yes
little eagle
Now the question is which part of the simulation is the worst.
little eagle
I suspect it is really bad, but does it come into play when everything is stationary? Is it that poorly programmed?
I'm really suspicious of the animations of soldiers. They seem very complicated when you look at the config and how many classes there are and how convoluted they are linked.
tough abyss
Yeah. Some of the animations look pretty bad.
But again doesn't all this get piped again through SQF?
little eagle
No.
The only things that are SQF in vanilla are... let me think... Many ui things and stuff like modules and eden attributes.
little eagle
What exactly?
tough abyss
Can't recall and don't have A3 installed but I de'pbo'd a bunch of files and found a number of scripts.
I mean come on the cfgFunctions.hpp
is written in SQF :/
I've seen the code that manages it.
little eagle
Oh yeah. The initFunctions.sqf is SQF. But that runs at mission start and is done at some point.
Also the .fsm of the animals etc.
tough abyss
But a lot of the functionality as a whole the entire game is built on SQF I think that there is in part of the problem.
little eagle
Everything SQF is not part of what dedmen said is simulation and rendering though. It should be it's own item.
tough abyss
No that is true.
But his "FPS unlocking solution" according to dwarden will just improve some scheduled scripts execution
Doesn't unfortunately effect the Simulation loop
little eagle
I think that there is in part of the problem.
And I don't think so. Not because it wouldn't be if the premise was true, but the most complicated things are done outside SQF.
The FPS unlocking is rather pointless when your server already is below 47 FPS.
still forum
@tough abyss initFunctions.sqf us actually quite performant. I rewrote it in Intercept and didn't get much improvement out of it
little eagle
It's pretty ugly though.
still forum
@tough abyss "But a lot of the functionality as a whole the entire game is built on SQF" Not that true.. SQF is still a very small part if you look at the rest of the engine
@tough abyss Not according to Dwarden.. According to me. https://forums.bistudio.com/forums/topic/210492-server-fps-limit-testing/#comment-3233256 That is what Dwarden said.
It does affect the simulation loop. In that it get's double the executions if you run at 100 fps
still forum
and as you can see rendering is what.. 75% part of the frame?
little eagle
The vanilla debug console is so tiny^^
still forum
@tough abyss Can you even read what you post? wDisp 2ms rendr 16.5ms
tough abyss
Look at my dia_capture you could say sound is the performance hog.
tough abyss
Look at the "main thread"
still forum
I do..
still forum
You are trying to tell me that of 28ms the 0.6 and 2ms ones are the most performance hogs? but you ignore rendr which takes 16.5ms
preload. Load Models/Textures from disk
little eagle
Yeah, I never bought into this the "sound breaks the FPS". I believe people started thinking this, because a problem that causes bad FPS also affects the sound. But not the other way around. Also gunfire sounds, but the simulation of the fired bullet is the actual problem.
little eagle
I just don't see how playing a sound would be expensive compared to drawing a landscape and models.
still forum
It does sometimes though.. @little eagle There is some bug that causes sound to sometimes go to 90+ ms. It's processing sounds you don't even hear and that shouldn't even exist.
Very hard to reproduce so I don't know of BI already fixed that by accident
Particles are on CPU
tough abyss
There is lies a big big problem
Move the particles to the GPU
SpaceEngineers recently did this.
Eliminated a number of performance issues
still forum
Particles are also multithreaded.. except they need to be on main thread for some reason
little eagle
t does sometimes though.
Sure. Never exclude dumb bugs like that..
tough abyss
I really don't like the word multi-threading when talking about ArmA 3
rough heart
Space engineers is still a power hungry game regardless of those Little performance updates
tough abyss
It does a weak scaling multi-thread.
still forum
Also server still has to calculate particles. Server's usually don't have GPU's
tough abyss
Called... function parallelism if I remember correctly.
tough abyss
Everything is talking back to the main thread.
So you can clearly see this when you try to run 4 HC's
I've done it for a community CPU core 0 maxed out.
The entire server effectively got crippled, and when a HC DC'd oh yeah that wasn't good.
I even had the HC's locked to specific cores
to keep them contaminating the server loop
still forum
do a serverside captureFrame and check why that happens
little eagle
Oh, also MP is pretty poor compared to SP.
little eagle
If you have 4 HC's on your server, you have 5 different main threads.
still forum
@tough abyss Where did you come from btw? Your first message in this Discord was starting this conversation.
tough abyss
I'm the one you and RyanD were berating for being a retard
little eagle
I've seen this name before.
still forum
Why do you need a new account?
tough abyss
Seperation from personal-life.
As I said I am not a "noob" when it comes to this game.
I had to do a lot of RND for that community.
The servers just couldn't handle everything on 1 thread.
Even with HC's...
That needs to be fixed.
still forum
Again. Check what exactly is the problem using profiling build.
tough abyss
I was belittling your results I got shirty when RyanD was being an asshat
I apologize.
little eagle
48 cores is a lot though. And Arma is single threaded for the most part. Maybe the server just isn't suited for the game? Lot's of smaller cores shouldn't help at all.
still forum
Yeah.. Fewer higher clock/IPC cores are way better
Get a watercooled server. Overclock CPU to 5Ghz and profit.. This is getting #hardware_vs_arma again.. kinda.. or #offtopic_arma
little eagle
It has been offtopic from the start. But it's not like any meaningful discussion about sqf was going on here. So who cares. Don't make me switch between 50 channels-
still forum
We always go offtopic here anyway 😄
little eagle
Let's talk about how replacing forEach with count is dumb.
still forum
well.. it is faster...
BI should just add a forEach without _forEachIndex.. That will be faster again because it doesn't count
little eagle
I put that into the same category as saying
false || true
is faster than
false || {true}
It's technically true, but in practice worse.
A)
{
_x;
nil
} count [1,2,3,4,5,6,7,8,9,10];
B)
{
_x;
false
} count [1,2,3,4,5,6,7,8,9,10];
C)
{
0 = _x;
} count [1,2,3,4,5,6,7,8,9,10];
A, B or C?
still forum
all are kinda bad
Thinking of it... Turning true/false into constants instead of commands.... Would make that even faster again
little eagle
myTag_true = 1 == 1;
myTag_false = 0 == 1;
#define true myTag_true
#define false myTag_false
+100FPS
little eagle
No.
still forum
might even be worse than the command call
still forum
Which again.. nulars being variables or commands is evaluated at runtime.. so... another thing to fix
little eagle
nularies* I meant
still forum
ah yeah... the player thing..
calling player is slower than getting it from a local variable in lowest scope
which is different from a global variable again
tough abyss
#define is probably better than a variable assignment only if it's used constantly.
little eagle
Yeah, globals are worse.
still forum
as it doesn't check if that could be a script command. because commands don't start with _
#define doesn't make any runtime difference
tough abyss
Weird according to the wiki it saves memory.
still forum
What?
little eagle
Disk space?
still forum
where does it say that? 😄
little eagle
Doubtful.
rough heart
Geekguy, are you fracture?
still forum
@rough heart yes. he already said that before
little eagle
It's rather slightly worse due to it's usage and you having to make the macro callsafe.
rough heart
it triggers the shit out of me
rough heart
GAWD
tough abyss
Using a hard coded constant more than once? Use preprocessor directives rather than storing it in memory or cluttering your code with numbers. Such as:
rough heart
But, for real, he got banned?
tough abyss
#define BUFFER 1.053
_a = _x + BUFFER;
_b = _y + BUFFER;
still forum
no
little eagle
Macros do save disk space though. At least for SQF. But no one should care about that.
tough abyss
I think the most interesting article I dug out of the net was this one.
still forum
true. It's easier to code. But at runtime it doesn't make a difference.
It makes the uncompiled raw script smaller but after compiling there is no difference
still forum
Also a compiled script stores the preprocessed script. not the unpreproced variant
tough abyss
Compiled script is so confusing does anything actually get compiled?
little eagle
Using a hard coded constant more than once? Use preprocessor directives rather than storing it in memory or cluttering your code with numbers. Such as:
Macros are better than global variables performance wise, but so is putting the same number over and over again into your code. And doing that is exactly the same as using macros after the script is preprocessed.
still forum
@tough abyss Read what it is telling you. Storing something in a variable is using more memory.
Which does make sense
@tough abyss Yes.. SQF is being compiled.
tough abyss
Compiled in what way?
still forum
From string to binary instructions
little eagle
I have a way that could disprove this, maybe.
still forum
Correct. Arma's VM is not a Java VM.. which doesn't make sense... Because Arma is not Java
@tough abyss What is it then.. If you know so much better than me
tough abyss
Yeah but it uses a Virtual Machine to execute code correct?
still forum
I don't think the definition of VM is quite fitting here
rough heart
^^^
tough abyss
Not in that sense it's virtualized in the sense ArmA 3 understand s it
little eagle
test_code_str = "call { 'code with whitespace a compiler should remove' }";
test_code = compile test_code_str;
"{" + test_code_str + "}" == str test_code
???
astral tendon
is there a way to reduce gunner acuracy?
little eagle
Why not?
still forum
a CODE variable consists of the preprocessed string & the compiled instructions.
when you call str it returnes the preproced string. preproc doesn't remove whitespace
tough abyss
I'm betting that the ArmA 3 VM is just a string based interpreter mapping those strings to C++ insutrctions.
still forum
@astral tendon setSkill ? I guess
tough abyss
And if that is the case where does the "code get compiled" ?
astral tendon
i set to 0 and still sniping me
still forum
@tough abyss It's compiling the string to an array of instructions. Each instruction has a Execute function that does stuff.
peak plover
@astral tendon Using mods?
still forum
compile @tough abyss
astral tendon
RHS
tough abyss
But what are those instructions?
little eagle
a CODE variable consists of the preprocessed string & the compiled instructions.
That is really weird, but it would explain the issue. It also means that sending compiled code over the network is even worse than sending uncompiled strings.
still forum
Constant,Nular,Unary,Binary call for example
tough abyss
If for example you look at intermediate languages like MSIL or Java Byte code
astral tendon
this setSkill ["aimingAccuracy", 0.0]
tough abyss
is that what this is?
still forum
yes roughly
peak plover
I had the smae issue with tanks. They would never miss, no matter what accuracy. I guess they are using a weird way of doing shots for tanks
astral tendon
for the gunner and the vehicle
tough abyss
So what is that intermediatry language coded in itself?
still forum
coded in itself?
@little eagle true that. I think Arma doesn't even compress the string when sending
tough abyss
The programming language used to build the intermediatry language interpreter or compiler.
still forum
C++
little eagle
And C I believe
still forum
Well... C++ that is also C code..
tough abyss
No...
C is not C++ not even close.
C you have to manipulate memory addresses by hand...
still forum
C code is the same code in C++
little eagle
The result is the same after the game is build, no?
tough abyss
pointers and all that fun.
still forum
yes
tough abyss
No C is not the same as C++
C is far more efficient, if used correctly.
C++ is managed a lot of the time.
still forum
Write a piece of C code. Paste it in a C file. works. Paste it in a C++ file. works. exactly the same.
little eagle
You could say that every C code IS C++ code.
still forum
true
peak plover
I wonder, if #intercept would help with performance when doing loops there and for ex. remove ai FSM and do ai stuff there?
little eagle
Not every C++ code is C code though.
tough abyss
^
little eagle
Hence, Arma is C++
tough abyss
C++ is much easier to deal with.
Than C
C is very unforgiving of errors segfaults are a persistent issue you have to deal with. Also, buffer overflows are really easy too. due to memory constraints of data types.
little eagle
The question is, do we trust the compiler making the C++ efficient more than the programmer writing efficient C?
Depends on the programmer.
Now look at what we got.
: P
still forum
@tough abyss Same in C++.. Except you stick to modern c++
little eagle
We're not talking about Arma, not a kernel.
still forum
so @tough abyss If you know C++ I can also explain it to you that way.
SQF code is compiled into a tree structure of C++ Classes. Each class has a virtual execute method.
The levels of the tree are the scopes in SQF. If you use call you go into a subscope which is a level lower in that tree.
Arma goes through the tree then. So a SQF instruction execution is a pointer lookup in the vtable and a function execution.
It's not parsing the String live as in SQS
tough abyss
Yep understand and no currently focusing on studying Python and C
Been using Linux more and more so C is very useful.
lavish hamlet
were should i use CBA_fnc_addPlayerEventHandler need it to run a script when it detects loadout change.. i have tried initserver.sqf and init.sqf to no avail.
tough abyss
When you talk about classes yes I understand instantiation so this tree which is the root class?
little eagle
init.sqf
still forum
@peak plover loop performance https://www.reddit.com/r/arma/comments/6zvcpj/intercept_version_090_released/dmygvyu/
Senfo is writing Zombie AI on a Server using Intercept.
tough abyss
Tree data structures have a root is call that root?
still forum
@lavish hamlet Player Eventhandler only works on the players machine. Not on the server
The root is called "Simple" it does nothing besides holding a local variable namespace. It's execute method
lavish hamlet
so this ["loadout", player call '\RDF_Radioer\scripts\forceWalk.sqf', true] call CBA_fnc_addPlayerEventHandler;
in init.sqf should work ?
tough abyss
So in which direction is this traversed?
still forum
no. you are missing {} @lavish hamlet
still forum
no.. A tree is by definition not that.
little eagle
call expects {CODE} and not 'STRING'.
still forum
Wait I'll do a thingy for you
tough abyss
So this tree is basically starting at "Simple" this branches off to more tree branches
So which is processed first in this tree?
In terms of commands?
This is intriguing and might explain some of the screwy results SQF gives.
astral tendon
or any way to reduce the damage of a vehicle?
still forum
1.0 call {
diag_log _this;
}
Base
Constant (1.0)
Code ({diag_log _this})
binary func call (call) (call opens a new level and executes the code)
And here is what is inside that Code:
Base
Nular func call or varialbe (_this)
unary func call (diag_log)
little eagle
RDF_fnc_forceWalk = compile preprocessFileLineNumbers "\RDF_Radioer\scripts\forceWalk.sqf";
["loadout", {
player call RDF_fnc_forceWalk;
}] call CBA_fnc_addPlayerEventHandler;
As I said which part of the class tree gets processed first?
still forum
So when it executes that code in the end it does in this order
Constant (1.0)
Code ({diag_log _this})
binary func call (call) (call opens a new level and executes the code)
Nular func call or varialbe (_this)
unary func call (diag_log)(edited)
It is a stack machine internally. Constant and Code push to the stack
a binary func call pops it's left&right arguments from the stack
still forum
The "binary func call" class has an array of possible functions as member variable
in case of call that array has only one element. That element contains the pointer to the C++ function in the engine
tough abyss
so call.methods ?
still forum
that array has multiple elements for SQF commands that are different depending on which type you pass into one of it's arguments
tough abyss
No offense but this sounds like a screwy way to make a compiler.
tough abyss
Finite Automata?
tough abyss
So it is line by line instruction
So why doesn't BIS make an actual high speed bytecode interpreter?
Take all commands translate them to byte code execute?
barren magnet
SQF comes from a little basis of commands from the first arma, back then there was no need to have that few things on highspeed, as arma itself and hardware back then was still slow
still forum
A Interpreter by definition parses the source code while executing it. Which Arma is not doing. It's compiling it into a intermediary that's easier and faster to execute
barren magnet
then arma got bigger and then kept using the same engine again and again
still forum
SQF comes from SQS.
barren magnet
Now Enfusion is their attempt to make things better, and they did
tough abyss
Yeah I've open'd the DayZ pbos
Enfusion is different
It even looks far more akin to C++
barren magnet
i played dayz mod for ages, and when i look at standalone i got more fps everywhere, and the leaked enfusion scripting looks way better to use than sqf
lavish hamlet
@little eagle thanks works like a charm 😃
tough abyss
I just wish ArmA 3 SQF actually allowed me to run my little immersion scripts around the bas 😦
They weren't even processor hungry.
simple PlaySound3D compiled with a spawn'd loop
and a sleep command.
It was utilizing the builitin sound-files arma 3 had like the quad-bike to mimic a generator running at a base.
And the chatter from the radio sounds file to mimic radio chatter under tents
with sat-phones and computers.
little eagle
So what's the problem? That sounds feasible.
tough abyss
It was, till the mission had 30 - 40 people.
it was running scheduled and started to stall.
The generator sound I had running on a while loop
little eagle
it was running scheduled and started to stall.
I hate the scheduler too.
tough abyss
every 0.5; seconds
still forum
Running clientside? @tough abyss
rough heart
Move it clientside?
little eagle
[{
//playSound3D
}, [], 0.5] call CBA_fnc_addPerFrameHandler;
fixed
still forum
Then grab my 200fps patched server. If you run at 200 fps then you'll execute 4 times the scheduled scripts
tough abyss
PlaySound3D plays global
Can't limit it to CS otherwise the object wouldn't exist.
But I had all these little scripts I had for the community I'd written.
little eagle
Don't use the scheduler and it will work just fine.
tough abyss
I mean on Takistan I had to massively modify a script to work with a sand storm
biggest issue with it was, getting the scheduler to play nice.
It hated that script.
tough abyss
I was also playing with new weather ideas as I said did a lot of RND for that community.
Owner became an arrogant ass, and never went back.
little eagle
The scheduler just isn't suited for UI, effects, sounds etc.
tough abyss
I mean for I&A I re-wrote the entire destroyed objective manager
little eagle
So get used to not depending on sleep and waitUntil.
tough abyss
And used onDestroy to manage clean-up
Lightning fast.
I think one of the biggest problems @little eagle was that community was running a shit ton of mods
the LHD was one of them it's geometry was screwed.
Stretching models and bunch of other screwery
Including vehicles placed on it exploding at mission start
I had to write a workaround that by putting a script on every blufor vehicle on the LHD
Mods...
Are...
a pain.
still forum
Have to disagree. Mods made by bad scripters are a pain
tough abyss
I wouldn't know I protested many changes, such as adding lots of faulty mods.
"I was just the server and code monkey"
peak plover
Well can't really complain about performance when 100+ units active, because that's pretty impressive, compared to other games
Is there something good to compare the scripting to ? dayzSA ?
peak plover
Does that have any significent performance issues?
still forum
You could call everything that doesn't complete in 0 time a performance issue
tough abyss
No as they limit how many lines of code you are allowed to have
Atleast in the Programmable blocks
Outside that the game is fully moddable with C#
atomic epoch
Can you try 100 agents? See if the performance is better
queen cargo
The way scripting in space engineers works is by creating byte code out of any scripted thing
And then just executing it normally in the clr
peak plover
// Check last notification, so we don't repeat the same thing.
private _lastSide = missionNamespace getVariable ["mission_respawn_notifyLastSide",sideEmpty];
private _lastTime = missionNamespace getVariable ["mission_respawn_notifyLastTime",(time-5)];
// Quit if it has not been 5 seconds and side is same
if ((time - _lastTime < 5) && (_lastSide isEqualTo _side)) exitWith {};
// set new side/time
missionNamespace getVariable ["mission_respawn_notifyLastSide",_side,true];
missionNamespace getVariable ["mission_respawn_notifyLastTime",time,true];
What are the chances that 2 clients run this script at nearly same time and both get past the if
Also how do I avoid this ?
still forum
Let the server tell the client to execute?
little eagle
What are the chances that 2 clients run this script at nearly same time and both get past the if
Guaranteed over time.
peak plover
Hmm, yes remoteExe will bail me out here!
lavish hamlet
RDF_fnc_vehicleDenie = compile preprocessFileLineNumbers "\RDF_Radioer\scripts\vehdenie.sqf";
["vehicle", {
player call RDF_fnc_vehicleDenie;
}] call CBA_fnc_addPlayerEventHandler;
private ["_veh","_backpack"];
_backpack = backpack player;
if (_backpack == "RDF_M11_a_d") then {
_veh = player vehicle;
if (_veh != player) then {
player action ["getOut", _veh];
Hint "Antennen er foldet ud"
};
};
why doesent this kick the player out when the == is true
still forum
Because player vehicle is invalid syntax
You have a script error in there.
and Arma should tell you
little eagle
-showScriptErrors
Example 1
? vehicle player != player : hint "Player is in a vehicle"
Let's delete this.
still forum
That is SQS right?
still forum
Put a big (SQS ONLY) after the example and move it to be the last example
little eagle
The wiki should explain things. This example just confuses more.
lavish hamlet
i tried this
if (_backpack == "RDF_M11_a_d") then {
if (vehicle player != player) then {
player action ["getOut", _veh];
Hint "Antennen er foldet ud";
};
};
stil dosent work
little eagle
_veh undefined?
lavish hamlet
yes.. thanks 😃
peak plover
missionNamespace setVariable ["GUER",resistance,true];
missionNamespace setVariable ["CIV",civilian,true];
Why didn't I think of this before ?
still forum
isEqualType
little eagle
Why didn't I think of this before ?
Wow, this is so stupid, it might just work.
peak plover
It just did 😂
👌🏿 👌🏿
15:48:45 Error position: <GUER,"nigel"]]>
15:48:45 Error Undefined variable in expression: guer
15:48:45 Error in expression <ide","_name"];
!(
_targetSide isEqualTo _side
)
};
};
if (_target isEqualType >
15:48:45 Error position: <_side
)
};
};
if (_target isEqualType >
15:48:45 Error Undefined variable in expression: _side
15:48:45 File plugins\respawn\files\fnc\fn_srvDeadRemove.sqf [respawn_fnc_srvDeadRemove], line 259
No longer errors 😂
little eagle
You could've just used my function to stringify resistance instead though. Now you wasted a week for nothing.
peak plover
I had this issue in many different scripts, the last issue I had was with tasks and groups as well. But thanks anyway
Couldn't stringify grups like that, so I just use "group" and compile it when I need it (not before).
I don't think I wasted a week on nothing
still forum
You wasteda week on nothing every 2 weeks.. Sleeping is such a waste of time
peak plover
github says 23k additions. I doubt I wasted a week on resistance 😄
still forum
no strings attached
peak plover
Well If I don't convert the group it errors
little eagle
Rewrite it so it doesn't error and accepts GROUP.
peak plover
The reason It does not work is because I have a function for adding a task and a function that loops the tasks
adds _target to array
then tries to grab it
fail
little eagle
Rewrite it so it doesn't fail.
Converting to string is never needed outside from callExtension maybe.
peak plover
I did, it checks if it's a string and if it's a string, it compiles it in the loop
No error
Finish it Then optimize, right? right?
little eagle
Sounds like twice the work.
still forum
write optimized code to begin with
vague hull
this doesnt sound like an optimization case in the first place..
subtle ore
Anybody know why the animal fsms work for a little while then just stop?
Had a dog move around for a little bit, but the rest that spawned just sit there
Disregard last
For future reference for anyone who is running into the same siutation: the animals were created to early and didn't load their proper fsm
still forum
what means too early?
peak plover
During preinit?
subtle ore
Yeah during preInit
Hmm, now i am only getting a certain area to populate with animals.
slender halo
i need some help
i have an object and i want to find its rotation angles relatively to world
subtle ore
You could use zero as a base distance, and then getPosATL on the player
Trig will be your friend on this one
little eagle
(_vehice modelToWorld [0,0,0]) vectorFromTo (_vehice modelToWorld [0,1,0])
something like this?
little eagle
AGLToASL (_vehice modelToWorld [0,0,0]) vectorFromTo AGLToASL (_vehice modelToWorld [0,1,0])
Just to be safe on hilly maps.
slender halo
yep, something like that but the position doesn't matter
i juste want the angle according to world x y z
i went with a rotation matrix but i ended with what vectorUp and vectorDir returns...
little eagle
Well, that is what I posted. Add some sine and cosine and you have angles.
slender halo
ok, i try that
thick sage
what is the category in cfgFunctions for?
little eagle
The ingame functions viewer.
slim verge
? wanting to add a marker link in a briefing but the link is part of a string.... stripped down example...
_STR= "1) <a href='marker:EF_POI_1'>Place of Interest 01</a>";
player createDiaryRecord ["Diary",["> ENEMY Forces", _STR]];
Tried wrapping the marker in '', "", '"
Anyone any idea of the correct syntax ?
slim verge
direct from the wiki.....
You can use some <a href="marker:MarkerName">links</a> too.
https://community.bistudio.com/wiki/Briefing.html#Markers
indigo snow
Briefing.html isnt a thing anymore
slim verge
okay so whats the reference i should be using, havent done any scripting for a few yeats now
indigo snow
<marker name="nameofmarker">link text </marker> iirc
slim verge
ah ok thx for that
indigo snow
Might be a phone typo in there
little eagle
And what are the expected results?
slim verge
works as expected thanks
indigo snow
Click on link in briefing text, map zooms to maker coord, commy
indigo snow
Nah since that page is on briefing.html
The current 'briefing' system is in the diary pages
Its certainly outdated but theres text on top saying its not for arma 3 iirc
slim verge
thx for the link cptnnick
indigo snow
https://community.bistudio.com/wiki/createDiaryRecord for completeness
runic surge
anyone know if the apex uniform boxes create and apex dependency?
lavish hamlet
RDF_player_connect = compile preprocessFileLineNumbers "/rdf_deltag/functions/fn_player_connect.sqf";
RDF_player_disconnect = compile preprocessFileLineNumbers "/rdf_deltag/functions/fn_player_disconnect.sqf";
["RDF_player_connect", onPlayerConnected, RDF_player_connect] call BIS_fnc_addStackedEventHandler;
["RDF_player_disconnect", onPlayerDisconnected, RDF_player_disconnect] call BIS_fnc_addStackedEventHandler;
Error in expression <DF_player_connect", onPlayerConnected, RDF_player_connect] call BIS_fnc>
16:21:54 Error position: <, RDF_player_connect] call BIS_fnc>
16:21:54 Error Invalid number in expression
16:21:54 File \rdf_deltag\functions\fn_deltag_init.sqf [RDF_stats_fnc_deltag_init], line 4
16:21:54 Error in expression <DF_player_connect", onPlayerConnected, RDF_player_connect] call BIS_fnc>
16:21:54 Error position: <, RDF_player_connect] call BIS_fnc>
16:21:54 Error Invalid number in expression
16:21:54 File \rdf_deltag\functions\fn_deltag_init.sqf [RDF_stats_fnc_deltag_init], line 4
simple solstice
why not just add the files to cfgFunctions?
also the second parameter should be in quotes
lavish hamlet
how would you write it in cfgFunctions ?
lavish hamlet
this is my config.cpp
class CfgPatches
{
class RDF_Statistics
{
units[] = {"RDF_STATS"};
requiredVersion = 1.0;
requiredAddons[] = {"A3_Modules_F"};
};
};
class CfgFunctions
{
class RDF_stats
{
class Statistics
{
file = "\rdf_deltag\functions";
class deltag_init{postInit=1;};
};
};
};
still forum
@lavish hamlet onPlayerConnected is wrong syntax. onPlayerDisconnected is wrong syntax. you are missing the parameter to the unary function call.
Maybe you should read wiki.
simple solstice
class CfgPatches
{
class RDF_Statistics
{
units[] = {"RDF_STATS"};
requiredVersion = 1.0;
requiredAddons[] = {"A3_Modules_F"};
};
};
class CfgFunctions
{
class RDF_stats
{
class Statistics
{
file = "\rdf_deltag\functions";
class deltag_init{postInit=1;};
class player_connect {};
class player_disconnect {};
};
};
};
oh fuck the spacing on discord
lavish hamlet
dont know if it changing anything but im trying to make a serverside mod.. that writes to db on connect and disconnects to track attendance.
simple solstice
ok, what I wrote is that the function is compiled already
still forum
@lavish hamlet Read the wiki entry to addStackedEH and onPlayerConnected
little eagle
That is not the problem. CfgFunction sucks, might as well compile the functions yourself for a mission.
simple solstice
I know, but since he has it already why not just throw in there
little eagle
["RDF_player_connect", onPlayerConnected, RDF_player_connect] call BIS_fnc_addStackedEventHandler;
["RDF_player_disconnect", onPlayerDisconnected, RDF_player_disconnect] call BIS_fnc_addStackedEventHandler;
["RDF_player_connect", "onPlayerConnected", RDF_player_connect] call BIS_fnc_addStackedEventHandler;
["RDF_player_disconnect", "onPlayerDisconnected", RDF_player_disconnect] call BIS_fnc_addStackedEventHandler;
if I want a module to only exist if a player has the role (AI disabled) would it be better to check it the player unit is alive or if it is null
little eagle
Who is djotacon?
runic surge
who is who?
thick sage
@little eagle "The ingame functions viewer." thank you. So, no semantic change besides being easier to search on the editor
little eagle
Yes.
thick sage
good.
Another question that came up recently: is it possible to declare functions in cfgFunctions that can only be executed on the server?
(for missions, not addons)
so that e.g. some functions are not even initialized on the clients
still forum
afaik no
tough abyss
You will need to make up your own code to compile(final) the functions.
Or you could just add if !(isServer) exitWith{}; at top of the functions you don't run on client.
Or just split the server functions into a serverside addon.
little eagle
Just add
if (!isServer) exitWith {};
as first line. Stop overthinking this.
waxen tide
so private []; is bad
params is evil too
why not private param for each input individually?
yes.
YES.
subtle ore
you're telling me :
private _belh
private _meh
private _beh
is cleaner than
params["_belh","_meh","_beh"];
doesn't make sense.
waxen tide
no, faster.
subtle ore
By how much?
subtle ore
if it's by 0.000000001 ms then yes I do care, cleanliness is better than a unoticeable difference when I'm reviewing latero n
simple solstice
why did you compare private to params, two different things
waxen tide
i mean suit yourself and use params and private not as a keyword but don't come crying tomorrow when dedmen painted dicks on your face while you sleept.
subtle ore
Dedmen can paint dicks on my face all he wants.
waxen tide
you can't apply private to vars if you define them with params.
but if you do it with param you can
simple solstice
for '_i' from 0 to 10000 do {
private _a = 'a';
private _b = 'b';
private _c = 'c';
private _d = 'd';
private _e = 'e';
private _f = 'f';
};
//28.4444 ms
for '_i' from 0 to 10000 do {
private ['_a','_b','_c','_d','_e','_f'];
_a = 'a';
_b = 'b';
_c = 'c';
_d = 'd';
_e = 'e';
_f = 'f';
};
//40.32 ms
credits to @peak plover
subtle ore
why would you even do that? Wtf?
If you're going to be using params, then just use the values that are passed
little eagle
I think Katekarin mistakes private ARRAY with params.
simple solstice
me?
lavish hamlet
```sqf`
_playerUID = getPlayerUID _Player;
_playerName = format ["%1",(name _Player)];
_sql_res = "extDB3" callExtension format ["0:SQL:INSERT INTO rdfstat SET playerUID = '%1',playerName = '%2',connected = NOW()", _playerUID, _playerName];
};
No, asdfghj . I apologize.
simple solstice
i only referred what to what is faster, private [array] or private _var
lavish hamlet
https://i.imgur.com/UPlQ7X6.png why does it skip uid and player name ??
simple solstice
someone here have an crystal ball?
post logs, post the query you run, post the script that runs the query
little eagle
🔮
simple solstice
we arent wizards
little eagle
Yes.
simple solstice
its because you are blue
subtle ore
😱
little eagle
daba di
simple solstice
first of all you have invalid SQL syntax
or maybe its correct, havent seen it
nontheless
lavish hamlet
logs are not showing anything.. but it only logs connect and disconnect not name and uid.. .
_sql_res = "extDB3" callExtension format ["0:SQL:INSERT INTO rdfstat SET playerUID = '%1',playerName = '%2',connected = NOW()", _playerUID, _playerName];
};
simple solstice
_query = format["INSERT INTO rdfstat SET playerUID = '%1', playerName = '%2', connected = NOW()", getPlayerUID player, profileName];
and I suggest you to use something like an extdbasync file that you can grab from eg. altislife
to make your life easier
lavish hamlet
im new at this.. but ill check i out..
simple solstice
but if you just want to insert using extdb3
@lavish hamlet does your database initialize properly?
lavish hamlet
yes im using it for presistent milsim mission.. i have just added the rdfstat table to track attendece..
simple solstice
ok, so if the syntax you provided for inserting the query is alright, and works with other functions
_query = format["INSERT INTO rdfstat SET playerUID = '%1', playerName = '%2', connected = NOW()", getPlayerUID player, profileName];
_sql_res = "extDB3" callExtension format["0:SQL:%1",_query];
in my case, I use 1:SQL:%1 for inserting, updating
still forum
@waxen tide I prefer params :D
you can't apply private to vars if you define them with params. Yes... because params does it automatically 😄 Read Biki
lavish hamlet
@simple solstice https://i.imgur.com/4tjkOkb.png (and) should have been T. Andkjær
but still no uid
astral tendon
is there a way to force vehicle to NOT use the road at all?
simple solstice
@lavish hamlet anything in extdb logs?
lavish hamlet
[19:41:43:901909 +01:00] [Thread 5016] extDB3: SQL: Error MariaDBQueryException: Duplicate entry '' for key 'playerUID'
[19:41:43:901926 +01:00] [Thread 5016] extDB3: SQL: Error MariaDBQueryException: Input: INSERT INTO rdfstat SET playerUID = '', playerName = 'and', connected = NOW()
simple solstice
did you make playerid the primary key in the database?
if yes, switch the primary key to the ID
because playeruid will repeat
and do you have 0:SQL:%1 or 1:SQL:%1?
lavish hamlet
CREATE TABLE rdfstat (
id INT(11) NOT NULL AUTO_INCREMENT,
playerUID VARCHAR(20) NULL DEFAULT NULL COLLATE 'utf8_unicode_ci',
playerName VARCHAR(255) NULL DEFAULT NULL COLLATE 'utf8_unicode_ci',
connected DATETIME NULL DEFAULT NULL,
disconnected DATETIME NULL DEFAULT NULL,
UNIQUE INDEX playerUID (playerUID),
INDEX id (id)
)
COLLATE='utf8_general_ci'
ENGINE=InnoDB
;
simple solstice
unique index should be ID
you can have an index on playeruid, but its not necessary
lavish hamlet
["0:SQL:%1",_query];
simple solstice
unique index means that this value wont repeat
and the playeruid will repeat. cause its not unique to the table
if you insert it everytime someone connect
and try switching the 0 to 1
UNIQUE INDEX != INDEX
lavish hamlet
the plan i to have a new entry into the table every time a player connects so if i connect 3 times = 3 entrys
simple solstice
yes. I understand
switch the UNIQUE INDEX key to ID
if you leave UNIQUE INDEX on playerUID then that means playerUID CANNOT be repeated in the table
lavish hamlet
Ahh ill try 😃
lavish hamlet
@simple solstice tried both option 0 and 1 and changed the unique index.. now i get a new entry everytime i connect /disconnects.. but still on uid and only 3 letters in the playername
no uid
simple solstice
hmm maybe try a different syntax..
_query = format["INSERT INTO rdfstat (playerUID, playerName, connected) VALUES ('%1', '%2', NOW())", getPlayerUID player, profileName]
Oh @lavish hamlet is the table set to UTF-8?
and the database?
æ
lavish hamlet
yes it is COLLATE='utf8_general_ci'
simple solstice
try with the syntax I provided
and maybe try removing non unicode chars..
Well I don't really know how you connect with the database, how the other queries work.. I suggest you to take a look at https://github.com/RPFramework/RPFramework
and see how they handled EXTDB
once you have it setup, so you can just
lavish hamlet
the last syntax does nothing not even connect disconect.. but no errors in extdb log
simple solstice
_query = format["INSERT INTO rdfstat (playerUID, playerName, connected) VALUES ('%1', '%2', NOW())", getPlayerUID player, profileName];
_sql_res = "extDB3" callExtension format["1:SQL:%1",_query];
you have it like that, I assume?
simple solstice
When I submitted it without ; I was wondering if you are going to add it 🤣
lavish hamlet
need to learn so fine by me 😃 still dosent work.. in other tables it has no problem returning the uid and the correct playername
simple solstice
hope you solve it.. or someone else can help you.. I cant seem to find the issue. maybe player isnt initialized
you can try adding waitUntil {!(isNull player)} before the _query line
im going to bed now, maybe someone else can help you 😃
lavish hamlet
ill try.. thanks for helping
lavish hamlet
yup 😃
lavish hamlet
@simple solstice got it working with
_query = format["INSERT INTO rdfstat (playerUID, playerName, connected) VALUES ('%1', '%2', NOW())", _uid, _name];
_sql_res = "extDB3" callExtension format["1:SQL:%1",_query];
_uid and _name comes from the onplayerconnected...