#linux_mac_branch
1 messages · Page 11 of 1
why waste time with those
I'd watch that
Because cheater vs human is boring
Cheater vs cheater can be fascinating
anyway you drifting way too offtopic for this channel
Fair enough.
lol
https://github.com/KoffeinFlummi/armake Found arkmake thanks @gritty sage
Epic games found the best anti-cheat is suing the little cheaters lol
as someone working on the edge of modern machine learning, too many people are grossly overestimating them. Results are far from 100%, and even if you did manage to attain 99.9% accuracy and that's beyond most models at present, that remaining 0.1% tend to be not just nearly wrong, but entirely, spectacularly wrong. Would an automatic anti-cheat that randomly perma-bans completely innocent players really be a good idea?
Neural networks are in a way dangerous, because we cannot understand how they reach a decision, we cannot know what bugs are laying hidden within the algorithm
all the training data in the world cannot eliminate the potential for random screw ups. It's all about percentages at the moment, what percentage of failures you're willing to accept in the system.
Well, we could go into the massive topic of AI safety, but probably not on this channel? 😄
Not on this server
@wispy breach uh nice, do BI have documented their file formats somewhere?
I don't know I'm a noob.
on BI wiki? 😂
yes... ?
it is funny (and sad in the same time), on one hand it is nice that BI didnt delete it from their wiki (even tho it is legal) but sad is that community had to do it themself
it depends if they signed EULA
in most cases you sign it while installing the game, but you dont need to install
and... I think reverse engineering file formats (without cracking the game) is legal
reverse engineering is perfectly legal. Doing so having signed a contract not to do it is a breech of that contract, breech of contract is illegal under civil law, not the actual act.
Building an implementation using the knowledge gained from reverse engineering may infringe on patents, that would be illegal - again, civil law
Of course, the law varies from country to country and reverse engineering may be illegal where you live.
OpenMW, WINE - thats all legal stuff
@scenic pendant Just so you know, I don't think we can discuss AI safety on this server, or most servers. It crosses pretty heavily into racism, which is nearly universally off-limits.
Um, this server isn't loading the image of the WAT that you are linking.
You get the point tho
I get the point that you are confused by what I just said.
I'm not willing to get into some huge argument right now, but couldn't not not react to the sheer riduculousness of what you just said
I'll leave it at that tho.
Well, if you think it's ridiculous that AI will probably be used as the basis for a fully sentient program that will be treated as the next civil rights case, I guess you probably have a few years to come to grips with that inevitability. I already have.
Hmm. AI's the same as on Windows, right? I don't see any known linux-specific bugs, hmmm.
Relating to AI, I mean
Mine is suddenly very weird O_o
New Arma Sitrep #00222 have BI words on Experimental Port, looks an answer about what u ask here.
Looks like the same as it's been so far
heh i saw the mech https://steamuserimages-a.akamaihd.net/ugc/954086927750233437/7883DFBC4F5FF8F04E12996ADE174EB92D9EC05F/
First thing i thought was "get away from her you BITCH" 😃
You haven't seen Macross, have you?
no not really into that sort of thing
Well, it's your loss. It's classic sci-fi.
so is Aliens 😃
Hey, I can appreciate that franchise too (aside from Ressurection). But you're still missing out.
Well sh*t. Far from the vote being about better support for the linux port, that sitrep pretty clearly states they are evaluating whether they keep releasing updates at all. No plans for us to get 1.84 at present.
@smoky crater I know you can't tell us, but did VP raise their prices?
It looks like VP is still working on it though. So doesn't seem that they temporarly stopped it
Imo, thay just say no news for us, do'nt expect change in near future.
@scenic sleet yep that one reads like “right now, we think it's not worth it”
Well I wouldn't expect them to just close shop and go home mid-release. If Linux is dropped, 1.84 will probably still come out.
Also, the SITREP all but states that outright: they state that 1.84 is delayed, and as for future updates — that's unknown.
My prediction so far: Linux is not worth it, 1.84 is the last available version.
those experimental port mentions in SITREPs have always had the same wording
@molten pulsar no, this isn't the same old statement attached to the releasenotes, this is a Sitrep in which they actually call into question the future of the port because of the cost involved.
"As for future port updates, even though we'd like to make Arma 3 available to as many people as possible, we're currently evaluating the expected popularity of these ports versus how much they cost to prepare. We'll make sure to let you know if and when new port updates can be expected in later SITREPs."
If i understud well Arma 3 did'nt born to be ported so this make more hard for VP to do thair job
not sure whether they would get enough takers, but I'd be willing to chip into a fund to keep them working on the port.
BI already got already more than 100€s from me so I contributed already 😛
Thats not really important if they start to lose money on port
the problem is that we didnt send tux cookies to their office
Let's do it
@hidden grail i have the oven but not the stamp
and a Tux decal for Edina
who is edina?
sorry, typo, Edita
Edita is the name of BI's T-72 Tank - https://twitter.com/bohemiatank
@covert quiver last time you told me they have T-90
😄
that would be a really costly decoration
actually i said T-100, thats why my friends call me Warship guy and not tank guy I guess
btw a bit of a missed opportunity not to include Naval Working Uniforms to the Encore Update
@scenic sleet must have skipped over that part.. 😦
that doesn't sound all too good :/
I'd be up for crowdfunding it if it ever comes to that
don't play that often but I do use it for modding purposes
That sucks, my guess is that they are in a situation where the ports in their current state are not profitable, and the current state is due to the porting having intrinsic delay
Hopefully the vote above shows that there is a lot of interest in the port if the porting job is sped up a bit :)
I guess +-175 interested people isn't enough
To be fair, they say they are evaluating the value, not that they have decided to stop maintaining the ports
Also, 175 interested people isn't 175 sales.
AlsoAlso, 175 people, if they all bought the base $40 game, would ammount to $7k
...if they all bought all the DLC (without the bundle, so we're being generous here), that comes up to a whooping $25025
You can pay a junior sysadmin for a year with that money! Wow!
Of course I forgot that Steam takes a cut. What's the cut? 25%? 33%? Let's be generous and misremember it to be 25%. $19k.
I'd be extremely sursprised if VP got paid under that for the port.
Overall
per update, maybe
I also forgot taxes btw 😃
Czech Republic has an income tax of 19% for legal entities.
So that would be $15200 after just the income tax
But salary in czech rep is much much smaller than in germany/us 😄
on the other hand, VP is not in czech rep
VP is based in the UK, London rates too ... I've no idea how many devs they allocate to a project, but I don't think they are exactly busy so just to keep the lights on they would have to be charging in the region of £10k a month for their services minimum. Each port takes a couple of months ... that already more than the $25k pre-tax, pre Steam cut
I doubt that even there, $15200/y is more than one decent specialist
I make more as an ops in Moscow
About twice that
a good C/C++ programmer with specialist experience in Linux, OpenGL etc in London could make six figures, but let's say lower end you're still looking at maybe £75k a year
rate drops some the further from London you get
And then! Then BI probably want to actually make some money, so not all of your money for the game goes to VP. But that's beside the point.
And I think I've hammered mine well enough
If anyone's counting btw, I'm prepared to buy the game to every single friend of mine who would be interested to play it (that would be about 4-6 people), but that would also be Russian prices, so about $40 per person with all the DLCs. Still not a lot 😃
Then again, we're basing all of this on how many people knew about and responded to an informal poll in a Discord channel. That's only a tiny fraction of all the potential linux customers ... so who knows?
Yeah, obviously
I was addressing the 175 people claim specifically
I'm pretty sure that if ~200 people reacted actively, BI have probably a few times that in sales at least, if the port gets parity and they handle the release right
But even if you take a thousand guaranteed sales, that's stil not really that much money
You know what would be nice?
Opensourcing the ports if BI decide to stop paying for them.
There's also crowdfunding.
worldwide, I'd expect at the very least twenty times that figure. I'd hope a few hundred times that, there are certainly the numbers of Linux gamers out there but it really is a question of marketing and the games appeal, plus you have to discount that number who already bought the game for Windows
and it's a 5 year old game, not a newly released title, so you're looking at the tail end anyway.
the ports will not be open sourced. VP would never agree to that
besides, the "windows" side of the game code is not changed that much
also its not like we get the source to everything. we dont get source to middleware like physx for example. we dont get the sources to BattlEye, only libraries prepared
Yeah, yeah, just wishful thinking on my end
i dont know how well wine supports xaudio2 for example, i know ARMA 3 makes heavy use of it. It's one of the things we spent a lot of time on building
arma3 actually had an openal backend, but it was too old for us to use it well
Wine has xaudio2 implemented through openal
then it has a poor implementation of it
xaudio2 is much more powerful than openal is
we know, we tried to do it that way too
in the end we hired a guy who wrote it entirely from scratch - the only thing openal is used for is the end output of it
I have zero knowledge about audio and all that but if they implemented xaudio2 through openal then it should, in theory, be the same wouldn't it?
well I guess it's mostly about the API and not so much about the quality
xaudio2 is a very powerful engine in terms of how you can chain effects, submixes etc
openal doesnt have that power - its pretty much buffer -> source -> effect (limited support) -> output
for example xaudio2 allows you to write your own effects and chain them in
sure, but that should be part of the xaudio2 implementation
thats kinda my point
I don't get it, if WINE has implemented xaudio2 then they should be able to do all that
no matter what they use at the end
not if they have implemented it just using openal to do all the work
our implementation does all the mixing and processing itself.. it only outputs to an openal buffer at the end
I'd imagine that's what wine does as well
if they have a proper implementation anyway
actually, it only outputs to an openal buffer at the end on linux.. on os x it outputs to Apple CoreAudio
so much of what Wine does is a basic implmentation with much of the API stubbed out and returning what Windows expect it to in terms of types, but without the actual functionality present
indeed
true
most games either use directsound which is quite simple, or they use a third party sound middleware engine which does all the work itself, and either outputs to dsound or wasapi
can't really find any details on their implementation besides going through the source code without any audio knowledge 😛
I guess it's a matter of getting the damn game to run in wine and listening to the audio ^^
tbh i think if people took to getting it work on wine, it'll probably make BI think "fuck it then"
yeah 😦
and tbh the only way i can see Wine implementing BattlEeye will involve breaking it's EULA, which is likely to lead to Wine users being listed as cheaters
tbh its not likely anyway, the windows version of battleye is heavy on the kernel side
Late to this party (been on holiday), but I've added my reaction to the pinned message. Would love to see better parity between Windows and the unofficial ports.
Source code of Arma 🕵 🤐
won't help you porting it if that's what you're after
You could use it to go the full port route instead of the wrapper route that VP is going
If BI publish the source code of RVEngine, 50% of the community it's capable to help BI for release a native cross-plateform version. But it's a dream.
50% of the Arma Linux community are C++ engine programmers? With knowledge about graphics and sound engines on linux?
Would probably be easier to work on including Linux support in a new engine imo. Not sure if Enfusion(?)'s devs have taken that into account, though
Enfusion runs on xbox already. And afaik playstation is also planned. Linux/Mac onto that shouldn't be that much more work
@gritty sage Not i talk about all Arma community.
@warm dune Enfusion is a new engine, you have less chance to shared the code. RVEngine is end of life, you have more change with this.
You need linux to do a linux port tho
About half of Enfusion is still RV
That's the "fusion" part
Yes, but Enfusion doesn't have source code leak and he have some total rework i known. And he was the earn bread of BI. All players of Arma play on Windows because Arma is release on Windows. But they can have MacOS X, Linux, ect... and known how opengl, openal, ect.. work.
Enfusion is not used in DayZ ?
Enfusion is used in DayZ.
Microsoft open sourced the Directx shader compiler (it uses LLVM/Clang ). Google already got it working on Linux. Then all is needed is to use mesa as library and convert or map hlsl to sprir-v for OpenGL/Vulkan. Then sdl2 makes a windows. Something like that.
There's a lot more in DirectX than just shaders
True but I have not read up on that stuff. sdl2 has audio and input stuff
It does but as both Dwarden and Jaycee have said, SDL2 lacks a lot of functionality which Arma 3 currently uses in the MS apis
But then that's the trap of using Direct*, MS want that lock-in
I'm a beginner with this stuff. I know vavle has steam audio and Yaakuro got it working with ue4.
sdl's audio stuff is useless
its input stuff isnt very good either in practice. we ended up dropping it and hitting udev
there were also some window manager related issues that we could only solve by going around SDL2 and hitting X11 directly
same on the Mac
Bring arma 3 to the mac baby!!!!
heh, well we do... but you need a very good mac 😉
problem with Macs is their gpu's are usually shit
This is why the Mac version runs on Mantle now.
I thought Mantle's dead, isn't it Metal API?
Oh. I meant to say Metal. I got that mixed up. Thanks for correcting me.
Np comrade
Metal is used because Mac's OpenGL is stuck in the dark ages
and slow
however most Macs have Intel GPU's which are just not capable of good gaming performance in most cases
Even Linux has better OpenGL support on Mac than MacOS
heh. well we switched to metal on MacOS for another project and it just carried across nicely to Arma 3
heh interesting. the server browser issue is looking like it's a Steam client bug, but only on Linux. the Mac version works correctly now
interesting
@smoky crater Something changed in how that was implemented between 1.80 and 1.82 though? Is it not simply an option to revert those changes?
Even if the root cause is a bug in the steam client, the older code was working and it's often faster to go back than push through, especially as whatever workaround you come up with may actually cause a new issue when the steam bug is resolved
Is it even known that the bug wasn't present in the old code with the new client?
Steam has had a very recent update
Yes
A few days before Arma 3 1.82 for Linux, I think.
They briefly and accidentally rolled us back to 1.80 a few days later and the browser was working
1.80 is still available in branch if you're interested in seeing for yourself
Discord is making a store that will compete with Steam. https://blog.discordapp.com/the-discord-store-beta-9a35596fdd4
well, considering Steam stepped on Discord's turf with their new chat feature, only seems fair that Discord fights back
@smoky crater, when you say that SDL audio and input is useless, have you ever considered running it past Iculus for improvement? He posts quite regularly on his Patreon on the work he puts into SDL2. I can't understand how it's used so widely, but useless in this case?
Intel have very good driver. He use all capacity of hardware. I can run Arma 3 on Pentium.
@scenic sleet @scenic pendant it is not our bug. It's on the client side of Steam - 1.80 previously worked, and it now exhibits the same behavior im seeing in 1.82.
@silver coral in my experience sdl audio is not used widely at all. Only by very simple things. And well, theres not much reason to change it - there are tons of better things out there such as openal-soft
my main complaint with oal-soft is that he wont implement a method for finding out what the output format is
so we cant tell for example if the user has it configured for a 5.1 system
and Linux has no standard way to tell what the speaker configuration etc is as Windows does
Ta. I had no idea! Interesting to hear real-world experiencies of using these technologies.
the problem with sdl's audio is it is pretty much only a callback for supplying pre-mixed audio
but if thats what you want, theres better ways to do it such as PortAudio
and no, you really dont want to start working directly with ALSA or Pulse because theyre horrible 😉
What's the code to unlock the 1.80 branch?
Probably Arma3Legacy180
That would be for the Windows branch, and that version doesn't exist for Windows anymore, but the port_180 branch does still exist
anyone know a fix for sound issues with thunder for example?
I vaguely remember there being a fix
it's very distorted
Oh yea, i've noticed that as well
hey guys, we are currently playing Liberation Altis on ArmaOnUnix server. If you want to watch us play
https://www.twitch.tv/eeksus
or join us at 92.52.6.237
count me in the reaction counter. I'm all for port to Linux, still
it's been a long road, but you made excellent progress throughout development of this game for Linux
i am deeply grateful for this
@scenic sleet the port_180 branch is for internal testing only.
@long sphinx we dont actively develop against it, we use radeonsi
@molten pulsar if you find reference to someone having fixed it please let me know 😃
Upgraded to a 1080Ti, still 40 FPS in Arma 3. It's now super stable 40 tho.
arma tends to be cpu limited more than anything even on windows
Yeah, I kinda expected this
It's not using my CPU cores tho
It's sitting on one, the others are lightly loaded
having to marshall all the graphics work into one thread probably does impact it somewhat
id like to be able to say VP will do a vulkan backend but it's a massive investment and im not sure they can justify the cost of it
we'll see
i mentioned Vulkan back when the Metal backend was written, and that MoltenVK existed then, but it was still commercial and our graphics guy didnt want to rely on middleware
oh, cool
there was also a more pressing need for it on the Mac as GL was really really cruddy on it
Still, the 1080 allows me to turn up the graphics stuff
And goddamn this game can actually be kinda pretty
and then you get shot in the head by a sniper miles away, lol
i dont really play the game myself tbh, im crap at it.. 😃
I'm only learning, and I'm also crap
BI have an experienced team of testers which does all the testing of game play in the port
But it's perfect to play while listening to youtube
Because a lot of the game is downtime: you get to places, you plan things, you watch
heh glad youre happy, i usually hear from linux people who make it into a pissing contest with windows
if they get 100fps on windows and 99fps on linux, they say the port is shit and throw a load of abuse
I dunno man, I've grown up playing games on computers that couldn't handle them well
It's not a perfect experience, but it's also not the worst.
And my CPU is 4+ years old.
multimonitor really does bug me but i have spent hours on that trying to get it to behave for linux.. no matter what i test my end, i sent it out into the field and it breaks for someone
it also doesnt help that some window managers appear to IGNORE move messages across monitors
see unlike windows, you can only go fullscreen on a monitor that your window is on, so you have to move it there first
Multi-monitor is extremely fiddly on linux, without bringing games into it.
Linux users should have gotten used to it by now, heh.
and well.. ive observed that my code sends a move command to the window, its pumping the message loop.. and yet the window does.. not.. move.
so either sdl2 has a major bug, or x11/window managers think they know better than an app does
And yeah, it's curious when people use non-standard shit, then complain that other shit breaks
...duh!
thing is i use xfce and ive seen it do it
My setup is really weird
stuff like i3, well.. yeahl. not supporting that
So I just expect shit to break
Hmmm. Have you, by chance, flipped a switch like “ignore WM hints” in your XFCE settings or smth?
Shouldn't be the default
My WM is explicitly set to do that, but I would also not come to you with “Arma 3 won't start on the right monitor” 😄
Cuz I know why
But I also find the entire concept of a “main” monitor to be silly.
My WM allows me to just write rules for windows based on metainfo, so I 'd just write one for the game if I wanted it to always start in a certain state on a certain desktop and/or monitor.
People use magic things that do automagic, hiding stuff from you, and expect things to always magically work right.
Take responsibility for your workspace dammit.
i only use the xfce default settings
but yes i suspect a lot of the "multimonitor doesnt work correctly" is window managers interfering with window placement
SDL doesnt really help either, it has no mechanism of its own to specify this, which is rather a glaring omission
they have monitor enumeration and they have telling a window to go fullscreen - but they have no "go fullscreen on this monitor"
i think if i did the linux backend again id not choose sdl this time.. id probably chose GDK oddly enough
heh, which DOES have gdk_window_fullscreen_on_monitor() 😃
linux dudes are always pissed off regarding linux games
they don't accept that most ot the stuff like SDL and others are really not that AWESOME like they do think
it's life
sdl isnt bad but ive seen people saying its better than directx.. which is just hilarious
theres a reason directx is such a popular api and it is NOT because MS force people to use it
after talking about the linux port for awhile
I finally bought the game
so, keep up the good work lol
heh thanks 😃
btw, how's the multicore usage of this game?
I plan to build an AMD rig next year
maybe an Intel
I just really need a decent cpu + lots of ram lol
i think arma3 is one of those games which isnt extensively multithreaded so it benefits more from 4 powerful cores rather than 8 weaker ones
that's good to know, thanks!
higher the clockspeed the better, some older CPUs actually perform better because of the recent trend for clock speeds to go down as the number of cores increases
Does anyone here know how to disable/modify amdgpu dpm quirks? Apparently my dGPU is downclocked to a mere 600MHz (instead of 1000MHz), and I can't find an online guide for that ((((
jaycee1980: @marceldev89 if you find reference to someone having fixed it please let me know
@smoky crater the only reference I can find is me asking the same question about a year ago in here and someone saying it might happen after ALT+TABing
😦
that bit mentions DPM
I'm not suffering from screen artifacts, it's just the low GPU clockspeed and low fps. But still, thanks tho
Btw I'm running a GCN 1.0 gpu
that setting will force the GPU clocks though
the one setting /sys/class/drm/card0/device/power_dpm_force_performance_level to high
does forcing the system to full performance mode really help with gameplay that much?
it's still clocking in 600MHz, I guess I have to find the "thing" in the precompiled mainline kernel and disable it manually
its a pity theres no way an app can hint to the system that while it's running, the system should be in a high performance mode
For the record, multimonitor works fine for me with i3 and Arma 3.
Not sure what combination of things I did, but it works well.
It didn't always.
yeah thats my problem- for me here on my own setup, no matter how i configure it, it works fine
yet i see lots of people complain about it
i think to 100% solve it i would have to get rid of SDL2 and start hitting X11 myself, work around stupid WM's ignoring window move requests and so on. Or possibly look into replacing it all with GDK or something like that
but i would have to do that on my own time as there's no budget for it
also not sure if GDK would work well with SteamOS
See
If you go around the WM, it starts working weird for people who actually meant the WM to ignore hints.
Basically, those people stop complaining, I come to you with a bugreport
Because an app ignoring my WM is a problem.
Personally, I think that people who disabled WM hints (or never had them in the first place) and then complain multimonitor is “broken” are fucking morons
And I wish companies had the balls to allow their employees to just say that to customers, when they are being fucking morons.
the problem is i can never be sure
and unfortunately Linux people generally have very loud voices when they are pissed about something
actually, something brought up on reddit recently, was when Linus got involved over a Witcher 2 bug
he did so because lots of people were throwing shit at us (VP) about it, even though I fixed the problem. he came along and said it was the Linux kernel's fault for breaking a userspace condition that worked before. That certainly surprised a lot of people.
including me 😃
Linus is always very adamant about not breaking userspace
Unless you absolutely have to
This is very much in line with his usualy policies 😃
Ideally, what should happen with multimonitor is people should realise multimonitor isn't a simple 2d easy task
I've tackled it multiple times in my own setup
And the code dealing with it is complex.
Especially if you want to auto-detect stuff
As much as we two don't see eye to eye a lot, on this, it seems, we do 😃
multimonitor on linux is just a joke tbh
its one thing x11 just does not do well, at all
and i find it disappointing that wayland would rather fight with a major graphics vendor and ignore others
tbh, does ANY game get multimonitor right on Linux?
In my opinion? Both yours and Feral's ports do well by just starting wherever I'm focused
Or, in case of Arma 3, where I goddamn told it to via the rules
As long as they keep doing just that, I'm happy
I don't want it to autodetect anything
I'll tell it where to start
generally we do one of two things.. enumerate monitors and tell the "Windows" side about it, and let the game control which monitor it opens on etc. Generally we do that if the game has config options which allow sensible setting of it
or
we lie to the "Windows" side, tell it theres one monitor
and provide a method to configure that ourself
e.g. --eon_force_display
Sounds reasonable.
In bspwm (I have WM hints disabled explicitly), I've never seen a window try to open on the “wrong” monitor
Either bspwm is brilliant and everyone should steal its code
--eon_force_display=0 should always be whatever xrandr etc says is the "primary monitor"
Or disabling magic autodetection makes things nice and predictable :3
Maybe both
Oh. So you do use xrandr? Good.
A lot of people think they are way too smart and go around XRandr altogether
im not 100% convinced sdl2 uses xrandr fully
That breaks every single setup except for GNOME 3 and Unity
And then, when I go tell the dev about it, they get pissy :<
we also dont use the old XF86 change resolution stuff
XRandr basically is considered the way to deal with X11.
If it can do something and you're going around it, you're doing something wrong
though tbh i dont like how this is handled in linux tbh. what they should have done is make the XF86 API's work, but not change res.. pretty much as os x does now
Yeah, stub and mock, mock and stub
fullscreen in x11 takes far too much handling from the app side
tbh I've not used true fullscreen in ages
as far as my app should be concerned i should just say "i want to go fullscreen on this display, at this res, with this colour depth - do it"
Everything supports fullscreen windowed these days
how it does it should be up to the WM/X11/whatever
@scenic pendant yes but it should be like windows and osx - "fullscreen windowed" should happen transparently to the app
bspwm won't, it just stubs
as far as the app is concerned it thinks its fullscreen. it doesnt care that actually the window is having its contents scaled to fit the native res
Although half the games that still try doing true FS crash under my setup lol
My fault, not theirs, obv.
particularly nasty is how when you crash in "real" fullscreen mode, it leaves the res set
thats just dumb
Yeah
windows hasnt done that in years, os x never has
Happens a lot with older titles and titles under WINE
Although yeah, newer Windows games work well on WINE with windowed fullscreen too
id very much like to tear out the SDL2 layer and start again in eON, but the cost of time spent cant be justified... at least not now most of the publishers have lost interest again
so id have to do it in my own time.. but yeah..
That's one of the basic and nastiest curses of a developer
“Ima just re-write this”
Warp knows I've done this hudreds of times.
And hey, a lot of the time it's at least sort of justified.
Bernstein said you should write a new instrument if the existing ones fail.
I think he wasn't a stupid guy.
well i just dont believe SDL2 offers enough control for what eON needs to do
its also quite old
eg it uses xlib instead of xcb
GDK looks interesting because it actually has the low level stuff we'd need to use, and because GTK sits on top of it it would mean we could integrate GTK interface into eON
also they handle wayland already
and GL contexts. i presume they will also add support for Vulkan to it
we also dont use that much of SDL2.. just the video and input events stuff
sound is all openal, joypads is our own udev code, filesystem is all our own code
id probably rewrite mouse input stuff so it skips xevents too and uses udev directly, at least for the raw mouse/directinput mouse stuff
Well. As a much smaller example, I thought xdg-open was bloated and took control from the user too much, now there's sx-open :D
Really, if someone took the time to write something that could actually compete with DX...
People should reinvent the wheel if the wheel sucks
Ot
the problem with "competing with directx" is directx has an advantage - it's a standard API all the way down to the system level
It's curious that GTK ended up reinventing parts of sdl
Probably for the same reasons that you wanted
And hey, GTK is ubuqutous.
SDL2 or SFML or GLFW all have to sit on top of existing api's
I say go for it if it works
each one of which has their own quirks
Well. I'm not saying “do a 1 to 1 competitor to DX”. It's fine to provide the same opportunities in different ways
GDK had a different target to SDL though, they wanted to focus on user interface stuff
Or maybe other opportunities even
It's just that SDL was written... not with gaming in mind
the problem with using GDK is i know theres a huge bunch of linux users who would simply spit and refuse to use any game requiring it
ive already had people argue with me over this.. because i used GTK2 for a launcher app... and i shouldve used GTK3
O_o
but then ive used GTK3... and apparently.. i shouldve used GTK2
and then i shouldve used Qt
or something else...
or ALL of them!
Ah, those crowds
its bad enough there is no standard UI API on Linux as it is
Don't pay them that much attention. There are purists who want their desktop to use only one toolkit
Which defeats the purpose of having several that compete
But w/e
People want to somehow have the benefits of both a cathedral and a bazaar
Best case scenario you just get a very organized bazaar or a very messy cathedral
It won't work
personally i like gtk, and gtk3 because it's the one theyre maintaining and it's fairly clean
I am firmly of the opinion that bazaars produce way more exciting and beneficial in the long run results
But hey.
At least don't try to mix opposites
glade really should be less shit though... they really need to fix that crash happy thing
ultimately though for UI stuff, the Mac has the best toolkit/UI for that
Probably because it's the only one in the system
well.. that and it is actually very well designed
That and Apple at least used to have people who knew how to do UI... reasonably well?
heh well the UI stuff actually comes from NeXTStep
I mean they've never appealed to me, and I've been almost literally forced into using Apple products.
As a result, I have a very strong bias against them
But still, they at least do competent UIs.
but the one thing it does right is that UI layout and design is NOT driven by code, its driven by template that can be done by a designer.. and hooked up to code easily by a coder
i see so many programs using gtk, and theyre building the entire user interface from code.. just... why.
Because it's coders making UIs for themselves
yeah but its the wrong way to do it
That's one of the main problems of the opensource community overall
It's people with limited skills using hammers on everything
Cuz a hammer is all we know
Took me years to do simple CSS, I didn't bother otherwise.
You can imagine how those web pages looked haha
it works similar on the Mac, you can create your UI from code if you REALLY want, but noone does. The IDE has an interface editor, it outputs a file called a Nib, the Nib contains the definition of the UI.. you create your UI by opening the Nib in your code
so then you can have some designer who doesnt really code but DOES know how to design a good UI, do your UI for you.. and you can drive it easily from your code
Well. It's people making software for free on their own time. They don't choose an optimal tool for the job, they just do things in a way they know how
heh
Hammers and nails
such people tend to loudly broadcast their opinions too
a good comment i read once
about our witcher 2 port
"theyre not proper Linux programmers... they dont use make files, they use an IDE"
Eh. The industry is young. I think these are growing pains
i have had SO much stick because eON is compiled using code::blocks
And hey, there is a kernel of truth to that horribly mangled statement
To play devil's advocate, people who never bothered to learn the lower-level tools are bound to repeat mistakes that are very easy to avoid otherwise.
well sure there are some IDEs who hide the underlying process from the user and teach ignorance. code::blocks isnt one of them. Neither is visual studio tbh
Xcode... well.. yeah
though not all that much
When I scoff at people “using IDEs”, it's not specifically for that. It's when they come to a help channel and ask questions a programmer should already know or at least know how to find an answer to.
IDEs aren't a problem, they are sometimes a symptom, one of many.
Otherwise, how the fuck are you supposed to manage a huge-ass project with vim
IDEs are useful man
tbh i could build eON with a makefile if i wanted. I see no reason to. It would give absolutely zero benefit
similarly i use codeblocks for my microcontroller stuff
i could use makefiles, but i dont, because i cba to maintain them by hand. why should i when an IDE can do it for me ?
The benefit of using a simpler common tool is ease of buid for your users/devs.
Less dependencies — less hassle.
You don't publish eON, so.
Who cares
well this is another problem with Linux really
using compiler tools as an installer tool
well, no its not really Linux exclusively, its any opensource OS
That's not really the way we install programs these days tho.
It's just that yes, traditionally makefiles also handle deployment.
installing everything from ./configure && make && make install does not make you l337 at all
and yes exactly. everyone uses packages of some kind
These days we wrap that in a package manager and voila
I think you misunderstand, because people tend to shout at you instead of explaining, lol.
Yes, Makefiles traditionally handle “where to put files”.
They don't have to.
What we want from you as packagers is a way to say “install the files where appropriate like we're in an actual system”.
It's an API of sorts.
What will actually happen is we would then run your script (be it Makefile or anything else) under fakeroot and package the result
i figured that was how stuff got packaged tbh
The benefits are simple: you now have a program that can easily be packaged AND deployed on a system without a package manager
built under a chrrot or even use --prefix and then archive the result
ive seen some things do very dumb things though such as hardcode paths into themselves at build time
eg if i build with --prefix=/home/peterm/libbuild/foo - it will hardcode that somewhere
Well that's a choice. Runtime configuration adds a lot of code.
You may not want to deal with that
But still allow the user some way of configuring stuff
Also you have to start somewhere. Some starting path, like the default config path or runtime dir is usually hardcoded
You may of course just not provide a way to change that on build time, but you usually have other macros in there anyway
So why not
i usually use --prefix when in building libs or something thats going into eON
we have a build chain of the stuff we use
Honestly, I stopped caring. My service manager just has some very basic defaults baked in
i rely on system installed headers as little as possible
And if you want to “change” them... those are FS objects. There are symlinks and bind mounts.
That's the beauty of “everything's a file”
Also consider that in ye olde times runtime configuration could actually not have been cheap enough resource-wise to implement.
UNIX descendants are knee-deep in tradition
Oh. You know what the linux gaming world needs? Something like Ruby's rvm or Python's virtualenv.
I mean yes, there's LD_LIBRARY_PATH, but I mean as a complex solu-wait a second, I'm describing containers.
@smoky crater so the question is then, have you considered deploying via containers? :3
Although I have no idea how the hell would that even work with Steam
It would require some very specific unsafe configuration on the user's machine, unfortunately.
What would be better is if Steam integrated into the package manager.
Then I could just shim apt and voila
Goddammit it would have been such a simple system too >_>
Honestly, give me a decent UI person and a month of vacation and I'll write my own storefront.
It will be weird and no one will use it, but I will!
It would do things with the system, not against it
honsestly ? no. i think things like flatpak etc are a bad idea. bundling lots of libs is the wrong approach
the proper approach would be for linux to stop making binary and ABI compatibility so frickin hard
and anyway yeah, most of our stuff on Linux gets deployed via Steam
for the odd "standalone" product releases we've done, ive used mojosetup
generally though we keep external dependancies to the minimum.
and anything we can bundle such as openal-soft we do
and again its another case of 3-4 things to do the same thing... flatpak, appimage, snappy...
support one and you get shunned by those who use the others
so instead its simpler to use something that is not as good, but works everywhere
I don't really see containers or flatpak etc. as sustainable solutions either
I should remember not everyone knows me and what's obvious sarcasm if you do >_>
Wow Tanoa's nice
honestly, when those kind of things happens
distros like gentoo will cease to exist
lol
There is no reason Gentoo can't evolve
No specific physical law is preventing that.
The only reason to use gentoo, at least back in the day, was the possibility to have multiple compilers and libraries in place
without much effort to switch between it all
For you maybe
Gentoo provides a number of various flexibilities other distros don't.
And doesn't provide some that others do.
Everyone who uses Gentoo suddenly thinks they are programming gods and when they completely eff up an application in an attempt to "optimise" during compilation they just waste everyone's time filling invalid bug reports and bad mouthing real developers for their "broken applications"
Packagers are bad enough but having everyone build their own applications is a terrible idea. Gentoo needs to die
@scenic sleet LOL
@scenic pendant sorry, just had a lot of bad experiences with Gentoo users in the past. Probably shouldn't have said "everyone" that's an exaggeration
It's not an exaggeration though when I say that 99% of bug reports filed on projects I worked on by Gentoo users ultimately turned out to be the result of users ignoring the recommended flags when compiling and often attempting to compile with contradictory options. So much time was wasted in trying to track down bugs that never existed in the first place that for a long time we automatically rejected all reports from Gentoo users.
@scenic sleet that's how newbies fuck up with everything haha
I prefer to use safe flags
Anyway, I'm feeling eager to try gentoo once again
and use LLVM for everything
Gentoo's global flag strategy is also kind of wrong. Different applications can be compiled with different optimisations and some will benefit greatly from something obscure while others wont. It doesn't rely on the usual chain where the developer and then the maintainer know the best way to compile an app. There are problems with the existing model as you compile for a very wide range and hence don't use the latest enhancements, so I do think something in between might yield a better solution.
Against my better judgement, I've bought Arma 3 for another two of my friends 😄
the server browser bug should be fixed today, just waiting for them to merge it into the steam branches
incidentally, if people file crash reports with our games on gentoo, i ignore them - generally because of what @scenic sleet said
i test on Arch (well, Antergos) and thats bleeding edge enough
even then with arch youre on your own
BTW, regarding the port being behind on version - this currently doesnt seem to be much of an issue for multiplayer. Most of the servers I am seeing in the server browser are still on 1.82
That sounds strange, though servers usually show up and then reject you saying you're on an older version once you connect
Still, great news that the browser is fixed!
dedicated servers
- with version 1.82:
1242 - with version 1.84:
5007
(that's what valve's master servers return)
huh, cool! Are there any statistics about how many of those are populated?
KoTH RHS isnt bad for casual play, but no one runs 1.82
1.82: 5 servers are not empty
1.84: 824 servers are not empty
Looks like arma3-servers can filter by version and connected players https://arma3-servers.net/search/ although it also filters by build number so you can search 1.82.144872 for example but not all 1.82 servers at once; might help to find those servers
Linux/Mac ports 1.82 were updated with fix for the server browser, to show servers 😃
can confirm, server browser works again :)
Great news, time to buy the last DLC im missing :)
@smoky crater the latest patch doesn't have a productVersion fix, does it?
Seems it does
ACE3 works on Linux once again!
Jesus ACE is so much nicer than vanilla
I can actually consistently and reliably kill enemies when I don't fuck up!
Also the medic system is awesome, I had such a great time being medic in my old clan
so you're saying that ACE works fine with default settings?
using the workshop version?
I'd expect that we'd have to use ACE 3.12.2. The latest version claims to require Arma 3 1.84.
afaik 1.84 is not really required. But you might need to edit some pbo's to lower the minimumVersion entry.
11:26:36 CallExtension 'ace_medical' could not be found
11:26:36 CallExtension 'ace_medical' could not be found
etc ...
I guess it doesn't really work
no that's a genuine bug
Even get's called on linux server. It ignores the "don't call this extension" flag
I reported that about a year ago I think
so in the case of the Linux port client it's good?
preferably it would support extensions
well, it seems to work kinda
is it normal when throwing a grenade with G instead of the advanced thingy that it gets deleted on throw?
and text on the interaction thing doesn't work either
throwing a ton of
11:42:31 CallExtension 'ace_break_line' could not be found
11:42:31 CallExtension 'ace_break_line' could not be found
I think you had to change the interaction menu to be a list instead of the default radial menu
that works 😃
now I just have to figure out how to stop gnome from stealing the windows key input
Gnome Tweak Tool should let you change the modifier key, but idk which options it lets you change it to.
don't really want to change it, I just want gnome to piss off when I'm in a fullscreen application ^^
hmm I guess I could switch to the default Super+s keybind for the overview and disable the global Super one
🤔
Once you get a good mod collection going that works for you, I'd advise backing it up
Wholesale
Or can workshop mods be rolled back?
Any mod could update and want a new version at any time
1.82 is a hard depend for the current ACE3
true, but the mods that actually check the game version probably have something up on github
^^
hmm something is off with the medical system though
1 shot with a pistol in the foot -> medic heals the guy -> 1 shot in the other foot -> dead
I keep my vital organs in my feet.
😁
Woo! Thanks for fixing the Server Browser thing 😄
you're welcome
can any of you let me know if you see an issue where the Server Browser list refuses to refresh?
Refresh button seems to work for me, the list is cleared and servers reappear when like when I enter the browser in the first place
ok good that is what it should do. on my machine, hitting Refresh usually leaves me with an empty server list
unless i wait 10 minutes or so and then it works. seems to be a Steam matchmaking API bug on my machine, very odd
That does sound odd
its whats been delaying the server browser fix being released
as ive been trying to debug it
but seems im the only one getting that problem :0
So the opposite of the famous "it works on my machine"
actually now i think about it i havent tried it on my other distro's. might just be a problem with my steam install under xubuntu.. but that would be just as odd
or with a different steam account
😦
The rangecard in ACE 3 doesn't work properly
It says that if my scope is zeroed at 300m, I should correct +3.5 for it lol
That doesn't sound right at all! 😄
Might be a long shot but have you tried another modem/gateway, @smoky crater ?
My cousin's modem used to give me trouble refreshing DayZ servers when it was still just an Arma 2 mod but worked fine at home. Dunno if it's relevant here though
@scenic pendant why does it sound wrong?
Some scopes are zeroed at 300m.
And +3.5 for what range?
When it's zeroed at 100m by default, it still shows corrections for 100m
100m should show 0.0, that's what zeroed means
Not sure if that's enabled by default.
Okey yeah... Sounds wrong 😄 Only with a specific weapon or with all?
I don't know how..... ...... Oh crap. Gotta check something
Jep.. As I thought.
The range card is calculated in the extension. Which you don't have...
It will fallback to script if you disable advanced ballistics...
Are you on linux? I could try to build you a extension
I am on linux, yeah. I did suspect the problem is in the extension
AFAIK there was work being done on linux extensions for ACE3
But there's no one who works on ACE3 and also uses linux, so
Or at least it seems so
https://github.com/acemod/ACE3/pull/5762 Yeah last activity here is in 2017
Well we are at 3.12.3 now.. 3.13 might be the next one and that's the milestone for that pull request soo.. soon™
That pull request has a linux extension inside it.
https://github.com/acemod/ACE3/blob/6a1b801416503da5ada33a6d215f56bda81d733a/ace_advanced_ballistics.so Can you try putting that into your @ornate summit directory? it might work
currently theres no support for extensions in the port.. i would add it but im not aware of anything i could test against
So a catch 22? 😄
@warm dune very long shot - i have Ethereal and know how to use it 😉
oh well
wuuut (╯°□°)╯︵ ┻━┻
But I thought you already had extension support 🤔
And you have the ACE extensions to test. Or just make a minimal sample Extension that just echo's back
https://community.bistudio.com/wiki/Extensions#Visual_C.2B.2B_3
That btw hard-blocks linux players from joining ACE3 servers, as most have advanced ballistics on
And hey, who can blame them, that's the best way to play
I totally cannot play without AB 😄 I helped bringing AB up to full real-life equivalency
And I'm using a real ballistics calculator app for my shots :U
BTW is that extension the only thing blocking advanced ballistics on linux?
I think yeah. The other extensions shouldn't be needed
And I don't see anything else that would be missing.
Unless scripts produce magically different results on linux than on windows.. which they don't 😄
https://8fw.me/hell/e1a9dbc4c4049384c179fa5184cc51c9ab877283.png this is how it looks btw
And yeah, it won't grab the extension
That's the "advanced ballistics disabled" display 🤔
It is. Because they are disabled
Waaait.. You have to zero in negative the higher the range is?
That's completely wrong 😮
Zero should go up with range. Not down
@gritty sage we dont officially support running DLL extensions.. thats a can of worms to allow because we dont know in advance what windows API's the extension would try to use
i would prefer to implement native .so/.dylib support
Yeah. I was talking about .so
ok well the port doesnt support .so
Well that fucks up ACE3 for linux users pretty much indefinitely :<
And makes the work to port them in vain haha
precisely why i am thinking of adding it
When would that happen though, realistically? A year?
we'll see
https://community.bistudio.com/wiki/Extensions#Visual_C.2B.2B_3 in the bottom is GCC crossplatform code for an extension. If you need something to test on
If it's not even planned right now, I'm betting half a year minimum heh.
its probably not that much work so i could probably release it as a hotfix to 1.82
breakfast first then i'll think about it:)
Hi
Not to be a complete dick (yes actually I love being a complete dick), but so far the Arma 3 experience on linux has been way worse than the disclaimers indicate.
And not to be an irredeemable asshole (that's going a bit far), I'll also mention that I've spend a considerable ammount of money buying Arma 3 and all the DLCs for several of my friends, and hope that helps as a drop in the Linux usage statistics.
YEs, I'm just one linux user, but I've bought basically 7 Windows copies, 4 with all DLCs 😄
Sounds a bit counter-intuitive, but Arma 3 is the only game like that in the entirety of human existence, so throwing money at it is the only thing I can do for now.
While also complaining loudly.
As an aside, I don't think the game is overpriced.
It's fine
I also have a friend who went downright suicidal while reading the last SITREP btw
I'm also switching my Laptop over to Linux (Antergos that I heard from jaycee [Got it running in a VM now and I love it already]) so that I can better port TFAR stuff
Same guy btw is an arma modder. Works exclusively on linux
I think your friend should seek professional help
He is.
@gritty sage be careful about your first impressions from “friendly” distros like Antergos btw.
I'd say especially Antergos and the likes of it, as what it does is it hides Arch from you.
But then it's not like Ubuntu or Debian, it's not a complete product that you can just use.
It's a fancy installer for a particular way of setting up Arch + some eyecandy repos
I've been running debian for years now. My biggest gripe with it is the outdated-ness of packages.
It's still Arch, and Arch's not a product, it's a toolbox
Ah. I've assumed a new user.
Still, if you're new to arch-based stuff — all of it is not like Debian, where “sensible defaults” is the word of the day.
nah I can find my way around. I just don't have the time to setup Arch from scratch.
I tried once and failed 😄
In Arch, we take stable upstream versions of software, see if it launches and doesn't break immediately, then push it to the repos
It broke compatability? Ooops.
“Stable” means “released”
Not “tested for 17 years by mute virgin monks”, like in Debian 😄
If the dev says it's stable, then it's stable. Unless it breaks in an obvious and blocking way
Yeah I've lost X a few times
Same here, the graphical setup is a touchy thing, especially if you made it yourself from bits and pieces.
Usually fixed or rolled back fast, so
Usually these things are caught in [testing] tho
Sounds just like the TFAR beta then... 😄
honestly i only advise arch and such for experienced linux users who know how to fix stuff when (not if ) it breaks
which i do, but i cba with it.. so i stick with xubuntu
There's also a massive difference in knowledge of the system even between what you'd probably called experienced
I can build a system with some essential software sources. It will probably even run.
@scenic pendant when you say "Worse than disclaimers indicate", what do you mean?
So anything that happens in Arch is almost always an instant “oh I know where this broke” for me. Not for all.
so far the picture im getting is that Linux users are complaining that there isnt 100% parity on all features with the Windows version
What I mean is that I've not expected the game to be this fiddly to get right to play with Windows users.
And hey, the last example is clearly not entirely the fault of VP. Ok so you didn't think anyone would want .so extensions in the port even tho the Windows version is using those widely. Ok, wat.jpg, but I've done worse mistakes. Hundreds of users had their sites down because of me, this is a minor mistake compared to that.
But what it entailed is me spending literally several hours figuring out what happened and why, and even then I had a dev confirm it completely.
And that has been my experience so far.
It's not that it clearly doesn't work.
It's that it's broken in subtle annoying ways
And yes, that is literally what I'm being paid to do: fiddle with software, but I want to play Arma 3, not tinker with it. Or at least tinker in the editor, not outside the game reading logs and restarting the game 32 times.
And that is not an exaggerated number, I counted. It just happens to be round
-ish
i shall point out that its offered as an experimental port, not a supported product. hence the need to tinker
I guess some of it is my fault for misunderstanding just how experimental this is
i could also point out youd be doing a lot more tinkering if you were trying to get the game to run in WINE
The productVersion problem was the same: annoying to catch, blocked me from playing the best version of this game.
I'm not blaming anyone specifically for anyone here
But this is why the port is not doing well
It's annoying to play
Not super hard, nothing is broken completely, but it's annoying and yes, I can't play with any of my friends, which is my 2nd place problem with the game.
Which also speaks to the quality of my friends, who all just fucked off to linux-incompatible servers after I bought them the game.
But that's really beside the point
And no, it's not enough for a server to run 1.82, unfortunately. A bunch of mods use .so's, and all of those don't work on linux
Which is why it's so surprising to me that the idea has just recently been introduced 😄
can you give me examples of mods that ship with .so's ?
Oops. I mean dlls
no i know some of them have .so's for the Linux dedicated server
extDB3 <--
id like examples of that
ACE3
^ no
Wait does it not have linux .sos for the server?
Only ones I know are extDB3 and Intercept. Though you probably can't really count the latter
if im going to add .so support i might as well make it compatible with the Linux dedicated server
I could have sworn O_o
No. ACE doesn't need extensions on server
Sorry then, I'm confused
But extDB is quite tedious. Requires you to install a special glibc version and it tries to use Arma's allocator. Not the easiest thing to use for testing
@elfin plover honestly btw I'm just venting. If it does push you to implement this [sooner] — sure, but it isn't entirely my intention to apply pressure, mostly I'm just annoyed at myself that I've spent last few hours of my life on stuff that made me ultimately feel worse
It's more my problem than anyone else's
But then again, would have this been pushed up if not for me getting angry today? 😄
@gritty sage extdb sounds like the sort of thing i would say "No support" to tbh
depends how it tries to use arma's allocator
Yeah same.
As long as the normal extension interface works. He can probably fix it up by himself to work.
A few of my friends and I are probably the only ones who were even aware of the range card problem lol.
we dont actualy support different allocators either
Also extDB is intended to run serverside anyway. Not really relevant for clients
@scenic pendant @gritty sage whats this range card thing and why does it need an extension ?
@scenic pendant not anymore. Already forwarded it to ACE team
It's advanced ballistics. It does ballistics calculations inside the extension
external bullet ballistics that is
Wind, bullet drop, coriolis, magnus effect
@gritty sage yeah that's what I'm saying. Maybe I shouldn't feel bad about this, it pushed new info to people who didn't have it
hm ACE3 seems like a good test candidate actually, its open source
And they already have the linux extensions here: https://github.com/acemod/ACE3/pull/5762 you can directly download them there
Just didn't get merged because the filesize is too big
And the team is decent to work with if you ever need to talk to them
Pretty ok people
hmm
i wonder why they built .so if they have nothing to test it against, interesting 😉
I might be a little slow today 😄
well.. i cant promise anything here.. but i will try
Usual disclaimers apply, yeah.
It's a job for you, after all.
You already have the Extension interface for Linux server ready anyway.
All you need to do it pipe it through your wrapper. And I don't know how hard that is
i actually need to alter the Windows side of things so that it can effectively call dlopen, dlsym and so on.. and also hope that native calls will map across ok
Heh. I told people on #gol back in the day that VP's wrapper is not a “simple thing” and “essentially WINE”, because if it was, everyone would already have one
And the more I hear about it, the more I think I was right, heh.
And that it itself obviously introduces its own set of problems
Software ain't easy yo
hm, is anyone here getting the crash when they have a gamepad connected?
hm, reproduced it with a PS4 controller. looks like evdev has gone nuts
Aw, I don't have my controller rn to test
heh no problem. looks like our code thinks 164 controller elements were reported for a device, and then crashes trying to enumerate them all
it is of course incorrect, theres only 27 on a ps4 pad
I can replicate it. Is it worth sending the crash reports with the controllers I do have?
it could be. please make one for each controller, and message me the crash report references
i suspect it will be the same problem..
QRPTIDHID2 - Crash with PS3 controller
1R34FZZCFK - crash with Xbox 360 controller
3K4QD28OHR - crash with Steam Controller
yeah same issue im seeing here. interesting you get it with the steam controller and xbox 360 pad, and i dont... oh well 😃
I'm using SteamOS' version of the xpad module if it makes a difference.
ok this is odd. it is only crashing when launched by steam itself
when launched from a debugger, or from commandline, it doesnt :S
very odd
ah, its something to do with controller mappings that steam is supplying
Yay I've actually finished a mission
My xbox 360 controller is connected but I haven't experienced any crashes
at all btw 👍
yeah seems to depend on exactly which type it is, and what GUID gets through to the game. should have it sorted soon
turns out it was a controller mapping Steam supplies at launch time was the missing bit i needed to get it to crash in my debugger 😃
So um
Ever since the start of this year Arma has been bugging out in how terrain doesn't render
Dunno if anyone has experienced and solved this issue
hopefully we get 1.84 soon.
@oak acorn any more info?
no
oh rip
read 2 messages above you ^^
is it planned to eventually bring it up to date and put it on par with the windows version?
Someone tried DXVK with arma? How not receive ban from battleye if play single player?
With the Windows version? I've heard it works fine, though BattlEye doesn't work in Wine so no multiplayer at all there.
Someone in the armaonunix server got it running successfully. As you mentioned battleye doesn't work in wine so you can't play on MP servers with it enabled
though BattlEye doesn't work in Wine so no multiplayer at all there there are tons of multiplayer servers without battleye. Just no real public and open ones
I didn't use BE for almost 3 years
But single player and MP on servers without BE should be fine
Ok, cool, will check it today
Yeah the specific symptoms make the game like this
Computer runs the game really well, always has. But only starting this year has this issue been popping up
What distro, what version of (i assume) Mesa.. what GPU...
I've done 3 or 4 complete reinstalls and verified the game files each time as well
High Sierra 10.13.4
oh right, the Mac... what GPU is in your Mac?
to be honest if this has "just started happening" my guess is an Apple update has broken their graphics drivers, wouldnt be the first time...
I'm running with an Intel Iris 6100 1536MB GPU and memory is 16GB
i'll ask our D3D layer guy about it tomorrow
welll, intel gpu's are not the best for games in general
apple having buggy drivers makes them even less suitable
drivers that they usually break on a whim and then ignore our bug reports about
you might try removing the eONprecompiledShaders.dat file which will be in <homedir>/Library/Application Support/com.bohemiainteractive.arma3/ somewhere
linux support 👍
If it dont work I'll just wait til I get my desktop oof
No dice, the ground still doesn't render in many situations. I'm just gonna wait until either it's solved or I get my desktop
is there a level/map that reproduces it well?
The screenshots he posted earlier looked like the second and third levels of the prologue (Common Denominator and Diplomatic Relations, I think)
rendering in general looks pretty bad in those screenshots, not just the terrain but everything on screen
Oh yea, the gun is turned inside out
I tested the grid, prologue and an eden editor altis. Grid had an odd bug where the hand that holds the underside of the gun behind the gun and firing removes the rendering of the ground
Altis and the prologue have hills and terrain that just didn't show up other than the now floating trees and rocks
Those screenshots weren't mine but they showed the symptoms I'm experiencing with my terrain. The game hasnt let me screenshot while I play
is it just me (and/or my drivers) or is the lighting from car headlights somewhat weird?
the ground in front of me isn't properly lighted
it fades off in about 2 meters
could be a vanilla thing though, it's been ages since I've played on Windows
That looks pretty much like car headlights work
like how*
They are angled down and fade out pretty quickly
Unless you use the “far lights”.
But if you use those in a city or in a combat zone, you're not very smart.
I'd assume people using cars in a place where you can easily be shot would angle the lights down even more, if anything.
Although properly angled headlights should illuminate the right side buildings somewhat, but not the left side ones.
So yeah, it looks off
https://8fw.me/hell/9a0fbace872eaa5fdfae52138d0bd217f44cff26.png hmm no they're just evenly angled
Yeah that's how normal car lights work
That's even how my bike's headlight works
It somehow illuminates the signs and the sides of the road really well, but won't go that far directly forward.
I'm not actually sure of the optics involved
I'm told it's called brights and lows. Lows work pretty much like that, brights are for driving in safe but unilluminated areas.
Would still be ok to have brights as an option I guess.
well it's true that car lights don't shine far on the ground but with terrain sloping up I'd expect that to be illuminated
it just felt off to me
hmm interesting
it appears that glsl_correct_derivatives_after_discard=true is no longer needed
with the current mesa git
this time for real 😛
flew around on Altis for a bit and changing settings every now and again to see if that affected anything
didn't see anything
besides the occasional screen artifacts I've been having for a while now
doesn't seem to increase FPS though
perhaps they're forcing it somewhere in their code?
it is being turned on by the launch script now
mesa devs have made it clear they "wont fix" the issue, as they consider it caused by "relying on DirectX behaviour"
shrug, if the workaround works and doesnt hit FPS, it's fine
ah, that's why
the FPS hit was kinda high before though
hmm commented out the export in the script and the FPS is the same
I guess they fixed that at some point or something
yep they did, which is why I decided to add it to the launcher script
they'd probably add it to the default drirc eventually
@smoky crater https://github.com/acemod/ACE3/pull/5762#issuecomment-413258392 btw
We look forward to hear official Linux support for Arma 3 as Linux player community
I'm sure most of the community doesn't care
Arma 3 is a pretty niche game as it is
Although I guess it should be appealing to the technical people, being a huge versatile sandbox and all
hmm, if I remove the -mod line from the steam launch params it's still loading the mods
any clue what that might be?
I guess I have accidentally clicked something in the expansions menu thingy
It remembers the mods you loaded after removing the launch option
You could check if -mod="" clears it or disable them in the expansions menu
-mod="" worked but I just removed the modset in the arma3.cfg file
something about ModLauncherList
you probably have still something ticked on in the extensions menue
Wine in steam for linux
Finally volvo got it
inb4 Arma multiplayer wouldn't work because of battle eye
I prefer an official port anyway
Arma 3 MP != BattleEye. A lot of servers don't require it.
Personally, I'm way more interested in playing ACE3 properly with my friends on a private server than in running a proprietary piece of spyware.
I'm very conflicted by this announcement, it's extremely damaging to native porters like Feral and VP. It's also going to result in fewer native ports.
It's equally good and bad for linux as a platform.
I hope that Valve disable it for games that already have a native version.
Can't shake the feeling that this is just the end of linux gaming rather than a shot in the arm. No-one will really appreciate that game breaking bugs are in the wine layer aren't going to get fixed by game's developers. They won't be providing any support.
well maybe the native porters need to put pressure on Valve or setup deal with STEAM, GOG, Origin, Discord w/e else want to distribue native ports
and after license fees are stored we will just see STEAM native port layer, powered by Feral and VP tech. blinking in the console 😉
My hope is that this will encourage windows players to switch over to linux which will increase market share and make ports more viable
Allthough, maybe with enough backing from valve and other companies, perhaps in a few years time we won't need ports for most games because they run well enough in proton
I see no bad news in Valve wanting to abstract Windows.
Whether they succeed or fail — that willl be the news
The idea is obvious and simple: they don't like the tyranny of Microsoft (and stated so multiple times openly) and want a less restrictive platform
So they are pouring money into that.
Valve are a “good corporation”, at least they've been so far.
They've shown a lot of bad sides, but greed has not been one of those.
So you can expect Valve to continue pouring money into this project even if it seemingly fails.
Which gives this a fighting chance
They've basically funded DXVK.
Now we know exactly why
It is already disabled for games with native versions lol
Otherwise we would have a lot of unplayable games still.
I see it as the other way around. Devs can see their Linux numbers when we use proton in the client instead of wine separately. This might influence them to make a native port, especially if they use an engine that supports it already... or at least a Vulkan renderer.
This could also lead to something like a Steam Machine being more viable again.
And all of this is free software, which is good news too!
But, of course, the biggest thing is market share. More games working without any extra steps is good for bringing people to Linux, which is good for Linux gaming as a whole.
fbt well, i think you forgot about the greed part where they originally offered like 30% to modders and then slashed it to like 15 or less %
corporate ways, they not that much better than other ones, they just lucky enough to have practical monopol and lack of serious competing parties
Personally I think this is going to do heavy damage to Feral and VP's business. It was a small market to start with, and now publishers of AAA have absolutely no reason to invest in it, they'll just let Proton do the job
and if it proves to be problematic, they will just say "sorry we dont officially support Proton"
valve may count it as a linux sale, but as far as the publisher is concerned, you stillbought a Windows product
At some point, they need to simply be called insane for counting Linux gamers among Windows customers.
heh well.. yeah i certainly dont agree with them counting these sales as Linux sales - theyre not, because they are not being sold officially as Linux products
its a Windows product that may or may not work with an emulation (meh spare me any lectures) layer
counting it as Linux sales is givin the illusion of support that doesnt exist
I believe Steam is providing support for those games that are officially supported under Steam Play
per their FAQ
Q: I'm a developer; my game got whitelisted in Steam Play; does this mean I have to support an additional platform?
No; if a game was whitelisted as a result of our testing, we've assessed the experience to be identical (save for an expected moderate performance impact). Users playing through Steam Play experiencing Linux-specific issues should be directed to Steam for support. Keep in mind users were most likely already playing your game using Wine; you just have better visibility into it now.
also i don't see the difference between Steam Play and what already was happening between WINE, Lutris, DXVK, etc
they'll soon back out of that tbh
its bad enough emulating quirks of windows behaviour, or rather quirks of how apps use the windows api... and thats when you HAVE the source code to look at. doing it blind with binary only is a nightmare, which is why WINE frequently breaks working stuff
they might be better doing it like playonlinux
where they have numerous builds of Proton packaged, and they keep them separate
then keep the games running with the versions they were certified with
Valve's entire purpose with this is to get as many people as possible to be comfortable with Linux as their gaming operating system. This will increase the market share of Linux to the point where publishers who only care about supporting platforms with marketshare might consider natively supporting Linux with future titles. The support for current and past games will have to be artificial in order for support for future games to be natural.
i dont think the market segment is going to grow in the way people think it will
put it this way mac is a much bigger market segment, and trying to convince publishers to support that is a challenge
linux is tiny, and now there is a "solution" which will probably keep 80% of that tiny market happy
but it's not "linux support"
Surely being able to play half of your steam library rather than a quarter will push some people over the edge
I honestly don't see any good reason why Macintosh even IS a gaming platform. I think that there is a great potential for Linux growth as long as Linux can be made to be user friendly and Microsoft has a major blunder.
@short tulip because people own Macs... and the people who generally own Macs are not technical enough to want to mess about with the likes of Wine
they want to click Purchase in App Store, and play it. and they will pay for that.
The thing is, people who come from Windows to Linux don't care about finding what good games work natively on Linux. They want to play their GTAs, their CoDs, their Warframe, their PUBG, their Blizzard games. And if it's a hassle to play that on Linux, they're going to back to Windows, even if it has it's own problems. If it works generally well out of the box, they won't feel the need to return to Windows.
until that game they cant run.. or that bit of software they cant run.. or they get sick of linux having far too many hidden gotchas - and yes you know what i mean 😃
plus linux really is not that user friendly
Sure, not everyone who switches over will stay, but people will switch over and more people than before will be fine with the downsides