2.22 and next prof RVFeature_ArgumentNoEscapeString can be used to prevent strings from being escaped.
Making it much easier (and more efficient) to pass and parse large strings, for example json as you don't need to fix the double-quoting.

Was your issue that your stringtable only contained <Original> language entries? That is something we fixed in todays hotfix

Was fixed in todays hotfix

2 of your 3 crashes are inside your graphics driver.
The last one is weird, it looks like a bitflip? Maybe you have RAM stability problems?

Difficulty settings are alread fixed on profiling branch and will be fixed by the 2.20 hotfix

Thats a issue we suspected, thanks for confirming.
But stringtable.bin should work the same, so it should not actually be a problem?
We'll fix it soon-ish that it keeps both xml and bin

You are loading CBA mod 4!!! times
You have "A3_GPNVG18" twice, I don't know from which mods
You have "WW2_Assets_c_Characters_Americans_c_US_Rangers" twice
You have "re_rhsusf_hide_stuff" twice

Can you explain what you mean by "not binarized correctly" ? What does "incorrectly" mean?

Won't do

We have feature flags.

We had to undo this fix, we may revisit it for 2.22

Please do send the crash log files https://feedback.bistudio.com/w/ft_a3_howto/gamecrash/

We also have toJson now to generate a giant string.
But, passing it to extensions is very annoying because it double-quotes the string. I want to fix that anyways.

The wiki is wrong, the default is not nil.
The default is nothing, if the value is not there, _this is not set.
It behaves the same as call.

Signatures is possible, BattlEye is not

Problem is the game does not have the languages.
At game/mission start, when loading stringtables, it only reads the configured language and skips over everything else.

Spam is not appreciated.

Should not be any slower, I also made the sorting multithreaded. So it should be much faster too, and in my tests it was.

BattlEye had an outage.

Action scripts are recompiled everytime IF they just contain a script fileName, in that case they essentially do a execVM.

It'll be a simple issue in the shaders. But the impact is probably too low to invest time into it

The backpacks in vehicles (not just quads) are generally hidding to prevent clipping issues.
For example sitting in a seat with back support, you wouldn't want your backpack glitching through it.

https://github.com/WolfCorps/gf_scripts/blob/master/addons/gf_scripts/functions/fn_p_makeUnitCurator.sqf
I use this script to give units zeus access.
It spawns ModuleCurator_F in script, same as you show it.

There is no way to reproduce this problem.

But what if both muzzle and magazine have different soundTypeIndex, how do we choose?

It was a BattlEye outage

A proposed fix was pushed in Profiling branch v31
It will reappear in 2.20 v1

The 16000/18000/19000 hardcoded limits I assume you mean.
Yeah we could probably raise that a bit, it'll still be limited by video settings anyway, but you can change them in config then
particlesSoftLimit/particlesHardLimit in config somewhere, they are just currently clamped by hardcoded in-engine maximum

BattlEye as having outages. this looks like it was caused by it.

Send the RPT/crashreport then: https://feedback.bistudio.com/w/ft_a3_howto/gamecrash/

Well if the server crashes, then we need the crash report files from the server.
Of course you loose connection when the server crashed.

Unlike ViV loading (where this was done), its alot harder to do with slindload. Because it literally just stores the memory point index. So that it can later look it up and consider animations moving that point.

The rope is actually unwound, back to 50cm minimum length.
But its stretched by up to 21x. The problem here is that the rope is elastic and stretches itself, not that ropeUnwound is returning a wrong value.
ropeUnwound is working as designed.

prof v31, 2.22

Map icon draw crash.
Old issue, we don't know how to fix it. So far. There are dozens of tickets for it

DOF parameters [scale, focusDIstance, blur coefficient]
Sharpness parameters [scale]
Fisheye parameters [aperture, scale1, scale2] (aperture is between 0-180)

Solved in https://feedback.bistudio.com/T163827

Fisheye: https://feedback.bistudio.com/T188336

// Would have to stop pylon commands from working within here to prevent recursion

Fixed in next dev-branch update

ACE team is aware.

Oh my good friend the stack overflow. That's not nice.

If there is a timeout, its in battleye config. If there isn't then it won't be added

and it wrote to me (0xC000001D - STATUS_ILLEGAL_INSTRUCTION)

Your CPU is too old.

We just use Microsofts API to access the Microphone.
For most users it works fine on 64-bit. I looked into it and couldn't figure out why it would fail for some users.

The problem lies in automatic updates being pushed through that break one or more mods and potentially de-stabilize an ArmA unit's ability to play missions until the issues are trouble-shot and fixed(could take hours or weeks).

I'm firin mah lazor!

Indeed. "AddCrewWeight" internally.
When get-in process is finished, it adds 100 to the weight.
When get-out process is finished, it removes 100 from the weight.
There are all kinds of ways this can break, teleporting a unit out of the vehicle should do the same, a player disconnecting and their body being deleted too.
This is unlikely to be fixed because of low impact and complexity in fixing it.

Fixed 2.20

The list of whitelisted addons is not stored in savegame.
When the Curator is recreated when loading a savegame, the addons are default initialized with only the addons that are currently active in the mission.
So it lost any that were added after curator were created (like is the case when you set it to all addons)

You could make it a giant string, the limit for that is 9megabytes? With 2048 arguments, you can pass 19GiB in one extension call.
That should be plenty.

There are lots of odd things about Arma AI :D

It behaves the same as
call x vs [] call x

Stuck on loading screen is a common issue with badly made mods, we introduced the ENDLOAD cheat code to bypass that. But official/vanilla, I've not seen it happen in vanilla yet
https://community.bistudio.com/wiki/Arma_3:_Cheats#ENDLOAD

Also need to consider the mineDelay for mines that do not have a timer.
A tripwire mine with mineDelay = 2, would show timeToExplosion=2, and be stuck at it until something actually triggers it.

You can always fetch the time delay by yourself.

timeToExplosion is only implemented for Bullets with "explosionTime" config entry, basically UGL grenades that explode after certain time.

1/2 were already done with callExtensionArgs and ExtensionCallback.

3. is mostly done, allExtensions command returns signatures. But there is no way to enforce signature checking and only allow loading signed version of extension.

Scratch some part of that.
I can use the same connectToServer hack also for the direct connect UI.
So you have to specifically enter domain and port 0 (or manually empty it (it will be default filled with 2302)) in direct connect.
Then it will try to resolve the port via SRV record, if that fails it falls back to 2302 (same as would've happened before if you left it empty or entered 0 manually)