What is the biggest issue that plagues FAF in your opinion?
-
There was actually a post from you about how the old client was better and more stable than the new one. I was the maintainer of the old Python client and the one who gave it a new lease of life by porting it from Python2/PyQt4 to Python3/PyQt5. I can tell you with 100% certainty, that hacked together shitheap that I tried to keep together was not the least bit more stable that what the java client is right now. Some major improvements like the ICE adapter that made connectivity not shit or random maps or team matchmaking were nigh impossible to do in the old client, too.
Your "shitheap" didnt frequently lockup, require a refaf per hour and had a working close button which didnt require me to pkill -9 the parent process. It also didnt swallow up more resources than the actual game uses.
Personally, I count those things are improvements.
If you think I misinterpreted the communication I've had over the last good while from FAF devs whenever I've run into these issues then I dont really see how you can fail to understand how the inverse might also work.
At the very beginning I tried to play the appeasement game and being ultra polite. That got nowhere. Evidence shows that flat out calling this out in plain language is working much, much better (if engagement in this thread is anything to go by).
You're a developer. You should be aware of the "lol it works on my machine" phenomenon. Random users telling you that does not mean the actual developers are like this.
As someone who has an SOC behind me that refuses to be abused in ways it was never intended for because we cant have devs who can write native code and therefore need docker or whatever the fuck container they currently have a hardon for I am acutely aware of the "it works on my machine" problem.
All I'm saying is that all the noises Ive made in random directions to all manner of people all over the place on various FAF teams were either ignored or were pretty much dismissed out of hand. In the meantime, its not just me with more and more serious issues happening more and more frequently. So before I do find the time to file however many bug reports and people get annoyed at the spam that will cause, I was told it might be a good idea to bring this up on the forum, which I have now done.
-
@anachronism_ said in What is the biggest issue that plagues FAF in your opinion?:
@MazorNoob When he referred to the stability of the old client, I believe he was referring to prior versions of the current client, not the python client.
I was referring to both. Old java client might have been a bit more stable than python when I ran it, hard to say. Windows 8.1 ESR in a VFIO KVM setup is not really the ideal target to develop for, but I was definitely hearing a lot less "its also broken for me". Point being I cant say if python was on top or level with old java client.
Recently though it is no contest that java client has gotten an order of magnitude worse and is now firmly below either old java in a whacked up VM setup or python.
Edit: On your language point, as I mentioned to mazor, it took being inflammatory to get the time of day on the issue. Its not like Ive just started mentioning this issue to people. The bait-language is more of a feature at this point because my patience ran out sometime last year after months of really trying and getting nowhere. Evidence here suggests that has worked.
-
@thecodemander said in What is the biggest issue that plagues FAF in your opinion?:
Your "shitheap" didnt frequently lockup, require a refaf per hour and had a working close button which didnt require me to pkill -9 the parent process. It also didnt swallow up more resources than the actual game uses.
I know the code that I spent a year or so maintaining better than you do. We wouldn't have ICE adapter today if we were forced to hack it into the python client, and that alone made FAF work better on average.
Out of curiosity, what specs are you running the java client on?
At the very beginning I tried to play the appeasement game and being ultra polite. That got nowhere. Evidence shows that flat out calling this out in plain language is working much, much better (if engagement in this thread is anything to go by).
No, it's just dissuading anyone from fixing issues if all they get is flaming that they didn't fix ALL issues. These people are not your subordinates, they won't start working better if you crack the whip.
As someone who has an SOC behind me that refuses to be abused in ways it was never intended for because we cant have devs who can write native code and therefore need docker or whatever the fuck container they currently have a hardon for I am acutely aware of the "it works on my machine" problem.
Expand on this, please.
-
@thecodemander said in What is the biggest issue that plagues FAF in your opinion?:
The bait-language is more of a feature at this point because my patience ran out sometime last year after months of really trying and getting nowhere. Evidence here suggests that has worked.
It didn't. Yes, you got "engagement" in this thread, but you didn't move anything closer to a solution. You didn't provide any useful info to investigate the bugs and I don't think you increased anyone's motivation to try to help you.
-
All I'm saying is that all the noises Ive made in random directions to all manner of people all over the place on various FAF teams were either ignored or were pretty much dismissed out of hand.
Can give us a concrete example of where you did this, or where you encountered it being dismissed*?
-
Im not claiming to know the code better than you, im observing its stability to an end user.
I tried a number of java environments. Based on whats installed and configured right now, I have tried on openJDK 7, 8, 11, 17, 20 as well as the oracle jdk/jre 8 environments with mixed results. Currently on 17 apparently since IIRC the last update refused to launch in 20. This I remember because I had to specifically pay attention to its launch options to get the damn thing launched.
If you want system specs its a first gen zen (170 or 1800?) chip with 32GB of ram which is often used for other things, graphics side its whatever the latest upstream MESA (23?) is on a GCN 5 gpu. Kernel is whatever is pretty close to upstream 6.3 if 6.4 isn't out yet.
No, it's just dissuading anyone from fixing issues if all they get is flaming that they didn't fix ALL issues. These people are not your subordinates, they won't start working better if you crack the whip.
I dont care if the issues are fixed at this point, as long as I dont have to listen to "mhhh yes, new release: there are no issues its great!" talk anymore whenever I make a meme about it. I was told it was unfair to call it out if I hadnt raised it here.
It has now been raised here. So the next person to jump in vc to say I'm being unfair to whoever because no one ever mentioned in writing on the forums that people are having issues is getting clowned on justifiably.
Expand on this, please.
If you want the work rant here it is;
I spend most of my time these days in kernel land because my company wants to deploy things and sell a platform so other people can deploy their things which means I have the glorious task of making things that are too cheap to officially support the kinds of things they want to do, work. When you have to meet certain safety critical applications, its like working with your hands tied behind you back which really eases things.
So when some clown goes "what do you mean I cant spin up 8 dockers in 512M of ram?! What do you mean I have to compile and statically link an application? or better yet, my node.js shit runs like dogshit on this, I really have a hard time taking them seriously.
The number of container platforms I have had pushed at me in the last 4 months touting them as curing anything and everything and how itll be so much easier when we dont have tickets of people not understanding that works on their machine and works on the target machine therefore we neeeeed containers, pretty please fix up these cgroups and fucking platform specific network drivers, pleeeeeeease really pisses me off. And then the ASIL and ISO26262 people show up... Its really fun. /s
It didn't. Yes, you got "engagement" in this thread, but you didn't move anything closer to a solution. You didn't provide any useful info to investigate the bugs and I don't think you increased anyone's motivation to try to help you.
Well if someone wants to say Ive never said there are issues that are not currently concretely reproducible that are happening quite often when I point to this thread, it'll be pretty safe to assume they cant read.
Can give us a concrete example of where you did this, or where you encountered it being dismissed*?
Dude you left the discord at the number of times you were exposed to complaining about these issues. So thats a good example for you I guess?
-
Dude you left the discord at the number of times you were exposed to complaining about these issues. So thats a good example for you I guess?
I left the Discord because I perceived the discussions with people, including the server admin, to be excessively toxic.
But to make sure I understand, you reported issues on your server. But did you also report issues on the official server, using the indicated channels such as the
client-bug-reportingor(game)-bug-reportingchannels? -
I left the Discord because I perceived the discussions with people, including the server admin, to be excessively toxic.
I guess the trash talk channel is the only place youd run into the server admin and it is pretty trash tier.
@jip I swear the list of places that are the correct place to report issues continues to get longer. The number of people that come and go in our discord being what it is that I wouldn't know where to begin with who wouldve been told about what. I've heard it needs to be on github. Then it was not github but zoolup. Then it was forum. Now its official discord.
If Im talking directly to someone inside the FAF organisation, either on VC, or in game, or in DMs or on another discord or via morse code shouldnt really matter.
The whole point of me posting in this thread is that I'm fed up of the runaround of "but you havent posted in place n+1, so youre not allowed to say there are issues". There are issues. Currently they are not in a state where a concise bug report can be filed.
Therefore going forward the attitude of "it doesnt exist" moving into "it wasnt reported" morphing into "it wasnt written in the right places" is not going to fly. I'm posting here because there is a serious problem with the process and communication here.
I'm posting here because the excuse of "we werent notified" is no longer going to fly and I'm fed up playing games with people about posting to the right people in the right place. I think this has had enough visibility at this point, I'm sorry it had to be baity and that people might've felt attacked or hurt but seriously, the fact it had to get to this point to get eyes on the issue is the problem.
Whether I can contribute or whether these issues are worth fixing and spending time on is a separate discussion.
I am still cautiously optimistic that the new community manager person will improve communication going forward because trying to raise a very simple point has been like pulling teeth. And in many cases it should even be necessary in the first place.
-
@thecodemander said in What is the biggest issue that plagues FAF in your opinion?:
Im not claiming to know the code better than you, im observing its stability to an end user.
I tried a number of java environments. Based on whats installed and configured right now, I have tried on openJDK 7, 8, 11, 17, 20 as well as the oracle jdk/jre 8 environments with mixed results. Currently on 17 apparently since IIRC the last update refused to launch in 20. This I remember because I had to specifically pay attention to its launch options to get the damn thing launched.
This is interesting. What distro? I'm on Gentoo with jdk 17 and 32 GiB ram and had no stability or memory leak issues whatsoever. That does explain the
kill -9issue, that sounds like java leaking memory and Linux shitting itself the moment it runs low. For what it's worth, I had other programs that eat a lot of memory do it to my graphical environment. OOM killer just refuses to trigger before everything becomes unresponsive, that's a linux thing. -
Therefore going forward the attitude of "it doesnt exist" moving into "it wasnt reported" morphing into "it wasnt written in the right places" is not going to fly. I'm posting here because there is a serious problem with the process and communication here.
You're putting words in my mouth, which is rude.
My intention is to understand what issues you're talking about. You keep mentioning you reported issues and got a lame response in return. But as far as I can see on the official channels that doesn't happen. And with the interactions I've had on your server (t3 clan?) I didn't quite remember myself responding to you or anyone there that the weather is picture perfect either.
If Im talking directly to someone inside the FAF organisation, either on VC, or in game, or in DMs or on another discord or via morse code shouldnt really matter.
Yes it does
, I'm not going to write down every bug people send my way. There's too many communications on a single day to process all of that. And on top of that I'm also sometimes just chilling and enjoying the game. Or at work when I get pinged. Either get it somewhere on a tracker (like Github, the dedicated channels official Discord, even a damn forum post) or the communication will get lost. It is just how it works -
@thecodemander said in What is the biggest issue that plagues FAF in your opinion?:
@jip I swear the list of places that are the correct place to report issues continues to get longer. The number of people that come and go in our discord being what it is that I wouldn't know where to begin with who wouldve been told about what. I've heard it needs to be on github. Then it was not github but zoolup. Then it was forum. Now its official discord.
And Zulip is not for bug reports. It is an rather isolated platform for productive discussions on problems with other contributors to the project. I've explained this on your server
-
@mazornoob Arch. (Insert meme here)
Im not sure its a memory leak, at least not the old one. There was one that was fixed (with crowdfunding no less IIRC). When I load up the client it knocks back about 980MiB and after while it usually settles down between 2-3GiB. Ive not seen a correlation between memory usage and it deciding to go zombie.
There are bugs I can repro with the subwindows and dialogs but I'm also using i3 as a WM which are not worth fixing since Im about 90% sure its an i3 bug those I really couldnt care less about.
OOM pretty much never runs for me, most of the time I have stacks of ram free. I used to have a chiplet run my background work while playing FAF but not any more since the client sits at 250-450% cpu usage now. Back then yeah OOM wouldve been a possibility but at that time I was in windows with its own ram and there was spare and nothing in dmesg.
Like I said, I have no idea whats going on or how Id go about try to repro. The trajectory in the last while combined with the responses Ive gotten on the topic are what has me concerned. The latter of which I am trying to address by following the instructions of "voice this on the forum".
You're putting words in my mouth, which is rude.
I didn't say who said anything though. If you mean why you left, I'm speculating based on the information I have. Im not going to engage in calling out who I spoke to or what was said, and I think theres plenty of evidence to support my point that the process and communication on this issue is suboptimal anyway.
Yes it does
, I'm not going to write down every bug people send my way.Im not suggesting you should. All Im saying is that if there are people out there saying <x>, then its not going to be received well when sweeping statements come out elsewhere saying the direct opposite. Communication getting lost is not ideal but shit happens. This isnt a one off incident with anyone in particular or at any single time.
Im seeing a pattern, the trajectory concerns me and there was concensus that it should be raised very visibly in the forum. As Ive now stated a few times, I feel this has been accomplished so the next time it comes up again, I can move things forward since the current barrier is "it wasnt raised in whatever way is the way du jour".
-
I dont care if the issues are fixed at this point, as long as I dont have to listen to "mhhh yes, new release: there are no issues its great!" talk anymore whenever I make a meme about it. I was told it was unfair to call it out if I hadnt raised it here.
It has now been raised here. So the next person to jump in vc to say I'm being unfair to whoever because no one ever mentioned in writing on the forums that people are having issues is getting clowned on justifiably.So it became pretty obvious that this whole endeavour was mainly to one-up the guys that you are apparently talking to, so you can now rightfully say that there are issues. In that case, good for you, I guess?
It's not really about coming to exactly the right place, many would work if were interested in working to a solution. You were self-sabotaging these efforts multiple times. You claimed that making too many bug reports would get you banned, you claimed that there is nothing useful in the logs etc. But it makes sense, because the important thing was not to solve the problem, but to gain the right to clown on your clanmates or whoever, that claim that there are no issues apparently. I am starting to get the impression that you are reading things into other people's behaviour because I have a hard time believing that anyone says "mhhh yes, new release: there are no issues its great!" either literally or in spirit, but I digress...
Anyway, I don't think I can be of any assistance here, so I won't try any further.
-
So it became pretty obvious that this whole endeavour was mainly to one-up the guys that you are apparently talking to, so you can now rightfully say that there are issues. In that case, good for you, I guess?
I was specifically, instructed to raise my concerns with the trajectory on this topic on the forum. This is what I was doing.
If everyone is so fixated on a sub-point not related to that point, nothing I can do. If you go out of your way to move the goalposts and avoid the point then theres not much I can do other than point out that you are missing the point.
You claimed that making too many bug reports would get you banned, you claimed that there is nothing useful in the logs etc.
Because the direct extension to "just file a bug report every time something you cant repro comes up" is that you file lots of bogus, unusable bug report which gets you banned from any sane tracker. I dont know what you want me to tell you. The bugs arent the problem, the trajectory of these becoming more common and the communication around bugs that exist but arent yet directly addressable is the issue.
When bringing this up I am running into a barrier. That barrier is that I did not raise it visibly on the forum. As I've written multiple times now, that is the point im posting about not the bugs themselves.
Anyway, I don't think I can be of any assistance here, so I won't try any further.
Youre welcome to discuss further if you have something to add to the actual topic instead of twisting it into what you want it to be.
-
You went on an awful lot of tangents about how the python client was better etc. if the actual point was how to discuss bugs better. So how should we improve communication about unsolvable bugs?
I have the strong impression that you are mistaking the opinions of more or less random dudes that try to steer you in a direction to get help for your problems with official communication by actual developers. And we can't solve that.
-
@TheCodemander just chiming in to say that I have never said the client is bug free and have always asked people to make bug reports on github, unless it was specifically for testing of an alpha release. I have never directed people to voice bugs on the forums. So when you talk about people giving you the run around I am not sure who you are talking to. I believe this is the first time we have ever interacted about any client issues.
Additionally from your comments to MazorNoob you say you are running in Arch Linux. Now it is known that the client has plenty of bugs both known and unknown on linux platforms due to the graphics library interactions. However, unfortunately linux is completely unsupported from an application development standpoint since I do not have the time to test across all the different distributions and setups individuals have. Likely in your case I would imagine there is some odd graphics driver interaction or environment constraints since it sounds like you have a technical background and likely have a more customized setup.
All that being said bug reports are always welcome but wanted to say up front that anything that is likely linux specific will go unaddressed unless some other developer takes up the case.
-
@thecodemander said in What is the biggest issue that plagues FAF in your opinion?:
@mazornoob Arch. (Insert meme here)
Im not sure its a memory leak, at least not the old one. There was one that was fixed (with crowdfunding no less IIRC). When I load up the client it knocks back about 980MiB and after while it usually settles down between 2-3GiB. Ive not seen a correlation between memory usage and it deciding to go zombie.
There are bugs I can repro with the subwindows and dialogs but I'm also using i3 as a WM which are not worth fixing since Im about 90% sure its an i3 bug those I really couldnt care less about.
OOM pretty much never runs for me, most of the time I have stacks of ram free. I used to have a chiplet run my background work while playing FAF but not any more since the client sits at 250-450% cpu usage now. Back then yeah OOM wouldve been a possibility but at that time I was in windows with its own ram and there was spare and nothing in dmesg.
Like I said, I have no idea whats going on or how Id go about try to repro. The trajectory in the last while combined with the responses Ive gotten on the topic are what has me concerned. The latter of which I am trying to address by following the instructions of "voice this on the forum".
I'll probably piss you off, but
arch
Given this setup I have to change my mind, you are being unreasonable. Whatever issue you have is definitely not general to Linux, since most devs work on Linux anyway and even something more esoteric like Gentoo is causing me no issues whatsoever. I can only assume that's JVM related in one way or another, Arch has both the philosophy of being bleeding edge and of making minimal to no changes to upstream for purposes of fixing issues.
At the very least you should have the awareness that you're running an uncommon setup and that having other people tell you that it works for them is no reason to be indignant. Expecting your issue to be fixed after reporting it and doing nothing else is also counterproductive in these circumstances, nobody's going to go out of their way to install Arch and reproduce your setup.
Maybe we should think about producing an AppImage for FAF? That sounds like fun.
-
You went on an awful lot of tangents about how the python client was better etc. if the actual point was how to discuss bugs better. So how should we improve communication about unsolvable bugs?
Because I am aware of how the python client is viewed, and its one of the only reference points to ground the argument. Saying trajectory bad isn't as useful as trajectory bad, heres the points its crossing and why we should care.
I would like to address the process of janky issues which are hard to repro. I havent gotten that far because so far every attempt to talk about them has been "its not on git" "this person hasnt been told" "I havent seen it" "why dont you bring it up on the forum" and now "it needs to be on these channels in official faf" along with some other platforms.
I have the strong impression that you are mistaking the opinions of more or less random dudes that try to steer you in a direction to get help for your problems with official communication by actual developers. And we can't solve that.
If there are client blaming issues there should be a (read: one) process communicated and publicly available which says who to talk to, that handles issues that cant be preprod easily etc.
Additionally from your comments to MazorNoob you say you are running in Arch Linux. Now it is known that the client has plenty of bugs both known and unknown on linux platforms due to the graphics library interactions.
Agree 100%. This is why Im cagey about me being the source of bug reports that I can repro. It isnt relevant to faf or worthy of your time to fix my i3 jank. I will fix that. Issues where I have the issue and other people on supported platforms also confirm they are running into them but also cant repro or arent sure, those are I think valid issues we need to tackle.
Likely in your case I would imagine there is some odd graphics driver interaction or environment constraints since it sounds like you have a technical background and likely have a more customized setup.
I have said to people what is infuriating is that I wouldve liked to test on older client versions which I never remember running into these problems but I cant because often client updates are a mandatory version check.
I'll probably piss you off, but
arch
I find this pretty funny tbh. I am aware and this is why I'm not keen on "I have this specific bug". Where Im annoyed is that Im having bugs that I cant repro that other people are having too and at the moment there is no way to even start talking about these, let alone addressing them because the discourse on the topic is cursed.
I have however found consensus in the notion that once its on the forum, visibly, then a big current excuse goes away. It is now here, on the forum. The relevant people (I hope) have seen it, so we should now be past that step.
At the very least you should have the awareness that you're running an uncommon setup
If you read my previous replies you would be aware that I already acknowledged this and I have already stated that I dont care about my specific bugs.
The process of "this is jank for a good number of people, and there is no way to raise this because the process is dysfunctional with regards to this type of problem", I have tried raising this and the barrier is pretty much "it isnt on the forum, and even if it was no one has seen it".
I set out to post about that specific issue (again, not bugs, my bugs, specific bugs or whatever) so that we can get past this step.
If I have to deal with i3 and linux specific bugs thats fine, I dont care (as much). Where I care is that there are actual client and game breaking issues out there right now that impact a good number of games and hours of lobbysim/queusim which are currently unreportable, undiscussable even. I want that to change and a first step is to raise it here (which it now has been) and to fix the process around that (which I am hopeful for with the new community manager)
-
@sheikah Not to be petty but it is technically not the first time. I dont know if you were memeing last time but you said something to that effect.
I would search replays where we had a small back and forth and you came off being pretty dismissive but all the search functions of the client are completely borked right now.
Edit: Found it by clicking through every replay like an animal. I believe you said and I quote "you are welcome to maintain or create your own client"
-
@thecodemander said in What is the biggest issue that plagues FAF in your opinion?:
If there are client blaming issues there should be a (read: one) process communicated and publicly available which says who to talk to, that handles issues that cant be preprod easily etc.
This is where the breakdown is because unfortunately there is no one process that can handle non-reproducible issues. In the big corporation ideal this would entail having a support engineer sit with the user and walk through their actions taking detailed notes or even having remote logging on their machine to take constant diagnostic to try and hopefully catch the cause.
The reality in FAF is that I just do not have the time to sit with every user who has a very niche problem that only sometimes happen. It would result in days of wasted time. Especially when the end user is not technical minded.
So that is really where the requirement of reproducible bugs and clean reports come from. Because otherwise it is truly impossible to fix without some miracle divine intervention.
