hmmmm
maybe testing the game in an environment that is the closest possible to the actual game would provide opportunities to actually find the bugs where they are. Not meaning map bugs of course. 2 hits, one shot...
Besides you can't honestly separate the two activities in a game development process. I mean it would look 'inefficient' for the least part.
What would be inefficient, would be to impose a task on a team whose expertise is in an entirely different field and for many may not even share the interest of bugtesting i.e.
@Santa_Claws : never said anything like that. Read again. At some point testing what a programmer modified has to go through loading a map and do test. Isn't it obvious?
@IronHorse : after this little trip into docking guts (up to 6 hours); i have close to 500 reasons to doubt (you can check it actually). It made me think of a war zone or somebody who had a HDD crash and copy pasted bit by bit the whole thing.
Refinery? Check mapping section...
It is; after looking into each official map (old and new), legit to think that maybe this boat don't have a captain any more.
WyzcrakPot Pie AficionadoJoin Date: 2002-12-04Member: 10447Forum Moderators, Constellation, NS2 Playtester, Squad Five Blue
@UncleCrunch You say "I see hitreg problems in 279," and then "no video needed -- it's obvious", halting a collaboration opportunity. Then you float defeatist commentary about mapping not having the "captain" you expected, as if justifying someone's decision to quit.
Creation and success are ugly sometimes, and often disturbingly selfless and humble. The "war zone" you see has genuinely entertained hundreds of thousands of people, and it didn't follow habitually idle disparagements of the status quo, as I see you trapping yourself with. My guess, rather, is that it followed a lot of hard, quiet, and very frustrating work that genuinely excited its caretaker(s) and stakeholders, with a mix of open-minded learning and collaboration along the way.
I see you've created a map before, so you're in a better position than most NS2 enthusiasts to create something compelling! I want my server's community to send me a link to /your/ Steam Workshop, so we can be excited about putting /your/ "war zone" on TGNS.
If you put into learning and doing half the energy you seem to be putting into critiquing and justifying personal defeat, I think you could do great things. You don't need a captain -- you ARE the captain. You need to decide you're going to create something memorable and you're going to be successful -- you (and the future consumers of what you'll long be remembered for, as opposed to these forum posts) would be very well served by your starting now to create and collaborate in that spirit /exclusively/.
No more excuses. You can do it, and there are lots of people here who will help.
@UncleCrunch You say "I see hitreg problems in 279," and then "no video needed -- it's obvious", halting a collaboration opportunity. Then you float defeatist commentary about mapping not having the "captain" you expected, as if justifying someone's decision to quit.
Creation and success are ugly sometimes, and often disturbingly selfless and humble. The "war zone" you see has genuinely entertained hundreds of thousands of people, and it didn't follow habitually idle disparagements of the status quo, as I see you trapping yourself with. My guess, rather, is that it followed a lot of hard, quiet, and very frustrating work that genuinely excited its caretaker(s) and stakeholders, with a mix of open-minded learning and collaboration along the way.
I see you've created a map before, so you're in a better position than most NS2 enthusiasts to create something compelling! I want my server's community to send me a link to /your/ Steam Workshop, so we can be excited about putting /your/ "war zone" on TGNS.
If you put into learning and doing half the energy you seem to be putting into critiquing and justifying personal defeat, I think you could do great things. You don't need a captain -- you ARE the captain. You need to decide you're going to create something memorable and you're going to be successful -- you (and the future consumers of what you'll long be remembered for, as opposed to these forum posts) would be very well served by your starting now to create and collaborate in that spirit /exclusively/.
No more excuses. You can do it, and there are lots of people here who will help.
A video content lacks a hundred million times the necessary information to debug this kind of thing and of course the other as well. Ask a programmer. Recording client and / or server data with a debug level setting will ensure better results in no time compared to rudimentary video upload that requires huge online space and leaving a PC uploading it for a full night (i got porn to DL ). I was not talking about 279 specifically. The core dump uploader is still in a bad shape it seems. If anything a developer should gather is critical info. Why is the necessary tools (automated), and corner stone for this, broken ? That would be the first thing to focus on in any development process (game or any other stuff). First thing first.
As for the warzone in Docking; open the editor, see for yourself. It's a war zone. I don't know if Docking was fun to make, the thing is, it doesn't look so if you wander in the editor with that file opened (not the art the technical). Trust me, my hobby production are clean and I have no orphan stuff in there. I wish it was possible to say the same for many official maps that would enable ANY modder (beginner as senior) to edit fast and easy and hopefully improve the whole thing.
I do criticize because as said before by me and others. Each time you can push forward (or have something on the verge of blossoming), we just see ourselves put on the side of the road with nothing but a kick in the balls. What about the Faded, Siege, Last stand (last echo from 2013 it seems) ? They all suffered many trouble not originating from their captain bad bearing skill. So... NO WE CANNOT BE THE CAPTAINS. Something else prevents it.
I'm not talking about rewards or anything related to ego. In fact my friend; being a modder for more than a decade in several games, you would be surprised to see how far beyond I'm am from what you suggest / imagine.
I see many +1/-1 people more than anything. Follower are followers. It looks like K-POP fans groups pushing the disagree button each time someone says something relevant and backed up by facts and a sound logic. And I exclude myself from this last argument. I've seen on this very forum many legit claims and suggestion blown up into pieces by zealots who didn't understood what the author was talking about due to a clear lack of education / thinking. But guess what ?; they whine more, they talk louder and in numbers...
In the end all those people are mouths (as myself you would say), you never see hands at work (not as myself, right?). The most derelict thing is that they could just help with 15 minutes of their time while they pass hours if not days playing games. You just can't ask something to a gamer. It's like asking a kid to tidy the room.
Start to read each of the custom map threads and see for yourself how much feedback they got. You can do it for the modding section as well. I guess +/- button are faster to press.
You want to test on TGNS... great!!! that will make 2 servers up with viable custom maps (maybe 3 - depending on ping/location). I have a list you can run and use in this thread. All reviewed and QC passed by your highness. Trust me when i say 3 of them are far beyond what the official did. I exclude my map from this last one of course.
"No more excuses. You can do it, and there are lots of people here who will help."
That made me laugh for 1/4 hour and made my day... sh|t I'm in a train... They all look at me now... I'm yet to see hands to finally enhance the custom offer in there.
Good one non the less. I like you already just for that one.
WyzcrakPot Pie AficionadoJoin Date: 2002-12-04Member: 10447Forum Moderators, Constellation, NS2 Playtester, Squad Five Blue
I understand where you're coming from. You want a different reality in which to excel, and what you want is reasonable. I'm just certain you're capable of more in the reality available to you, and I hate to see that potential misspent.
A video content lacks a hundred million times the necessary information to debug this kind of thing
No, it's actually the best way to demonstrate reproduction steps and is actually precisely why I requested it.
As for the debug output, I've already explained how to do that and what it looks like in one of the many videos I've made. We need a video AND the debug output.
Why you persist in assuming what we've requested from you as not being sufficient for us to analyze, just makes no sense... we know what data we need, so leave that judgement up to us.
The core dump uploader is still in a bad shape it seems. If anything a developer should gather is critical info. Why is the necessary tools (automated), and corner stone for this, broken ?
I don't get the point of the rest of your post, honestly, and I read it twice.
Just seems like general discontent with no specific examples of what the developers could do to to address it, besides better modding tools?
I don't get the point of the rest of your post, honestly, and I read it twice.
Just seems like general discontent with no specific examples of what the developers could do to to address it, besides better modding tools?
Read the comment he links ('how far beyond)'. It's informative & amusing.
Still: Loads of talented people wasting time here.
( well, /some/ of what you want is reasonable. also, you're hindered by written communication syntax. is English your first language? )
The second on the list. If you find something wrong in terms of grammar or vocabulary (colloquial or not), feel free to help. I'd be glad to get rid of the many mistakes i do / bad habits i must have. Languages change faster than before internet. Especially English it seems.
A video content lacks a hundred million times the necessary information to debug this kind of thing
No, it's actually the best way to demonstrate reproduction steps and is actually precisely why I requested it.
As for the debug output, I've already explained how to do that and what it looks like in one of the many videos I've made. We need a video AND the debug output.
Why you persist in assuming what we've requested from you as not being sufficient for us to analyze, just makes no sense... we know what data we need, so leave that judgement up to us.
There you go, you got told @UncleCrunch
When are you going to post your findings?
The so called "best way" will NOT provide whats needed for debugging such kind of issue (the holy netcode stuff). Ask any programmer.
You need 3 outputs : 2 clients and a server in the same situation I described and that are synchronized with NTP. At least each machine have to know the time diff they have between each other.
Plus (cherry on top) the analysis of network paths (ex: traceroute A to B and B to A) as it can be different but also A to B may vary from time to time... So you have to check & log it too... For each paths. Try Matt's traceroute and play with it and you will see how it can be tricky sometime to just reach google.
Not to mention the DXDIAG, the NS2 settings and all that can be logged should be taken under consideration.
And finally make a cross analysis of the 3 logs step by step (at least the relevant ones). That's why time is important in this matter.
This log factory has to be industrialized / programmed for server side and client side. Also it needs an automated analyzer. Humans are slow readers. It means a long list of things to do.
There's nothing in NS2 right now that can do that to my knowledge. It would give many clues on every other issue related to networking (or discard options) if it is the culprit (or one of them). It may be something else. Doing this debug the proper way may not reveal whats wrong right away or focus on this specific issue.
This is the kind of functionalities that should have been implemented not long after the game was out or before... Why ? Because it also helps to create stats on the long run to see what problem occurs the most. In this case a statistician would be helpful.
BTW I'm not specifically targeting b279. This "thing" has been around like a ghost for some time now and happening in some specific situation. It tends to occur slightly more often (Gut feeling).
First thing first. Any other way is pointless and will make people loose their time.
Ok, @UncleCrunch. You and Ironhorse are not getting anywhere. Although it was humorous, Ironhorse did get a bit childish with his gif.
The toolset and debugging features you describe would be amazing. They would greatly benefit ns2. The sad truth is that no matter how correct you are, we are not going to get those tools. UWE has stated quite clearly they are interested in other things.
I know you don't like it. I don't like my 1988 honda either, but that also will not change anytime soon. There is only one thing left to do. That is work with the tools we do have. If all we have is duct tape and a few sticks then that is what we use. Same with the debugging toolsets we have.
Ironhorse has described to you how the UWE/CDT programmers want data to debug with. You are saying you are able to reproduce issues consistently. You seem to be very good finding bugs no one else can. Would you please share your abilities with the programmers by finding the bugs and sending them the data they want? It may not be all the data they need, but it is all they have to work with right now. Could you be that generous?
Will play again if next patch reads: "Fixed bad old days performance issues"
The same goes to you. No one else is reporting such problems. You are the only one. We can't reproduce your problems without your help. Could you be so kind as to help us find the problem? It would not even take too much of your time.
If you could be so kind to follow these instructions and make a thread in the tech support forums it would be extremely helpful.
Type "p_logall" without quotes in the console (~ key) whenever you want to begin recording in game (don't do this until you at least spawn in a round)
then you can p_endlog /or just wait for the round to end/ or map to change/ or just quit the game and it will end automatically.
Then navigate to your hidden folder
c:/users/YOU/appdata/roaming/natural selection 2/
And you will find your file, something *.plog
Zip it/ compress it.
Upload it to somewhere on the internet
Make a thread in the tech support forums, and put a link to your .plog file there.
This will give me a frame by frame analysis of what is going on, so try to pick a time when you can best reproduce your lowest performance like the 24 fps you said you get.
Don't record for longer than 30 minutes, and don't record for shorter than 5.
Include your techsupport.zip file with this compressed plog.
Even though i'm on all low textures,my game is running SUPER smoothly with the new update! Back in early 2013 though,I used to be able to run it on all high with no problems,guess the performance is wearing down a little
I have no idea how to embed or even if you can, but here's a video i threw together tonight demonstrating a few of the hitreg issues I've been having. I can provide more info if needed (specs, unedited videos, whatever). I had a few others but after reviewing them, it seemed that the counts were right, but the sound was delayed for the first few hits.. similar to the 1st one in the video.
This was all on DMD Dallas 2, which does pop up the warning bout server settings being modified.. However, I can say for sure that this happens on other servers too for me.
I was not running the console command, didn't know about it until I started reading through this thread.. I'll try to use it in the future whenever it's available.
I have absolutely no idea how to reproduce these consistently, but hopefully someone else can pull something from this.
Maybe you could also add keyboard/mouse position to the debug? would help in re-creating the process, maybe even have a record function for this so it can be visualised, that sort of makes sense since input can screw with the way the collision model works too
I have no idea how to embed or even if you can, but here's a video i threw together tonight demonstrating a few of the hitreg issues I've been having. I can provide more info if needed (specs, unedited videos, whatever). I had a few others but after reviewing them, it seemed that the counts were right, but the sound was delayed for the first few hits.. similar to the 1st one in the video.
This was all on DMD Dallas 2, which does pop up the warning bout server settings being modified.. However, I can say for sure that this happens on other servers too for me.
I was not running the console command, didn't know about it until I started reading through this thread.. I'll try to use it in the future whenever it's available.
I have absolutely no idea how to reproduce these consistently, but hopefully someone else can pull something from this.
Replay the ones where you "hit" in slow motion and I think you'll find that you were really missing. If you're using NS2+, turn on "Server confirmed blood" to have 'better' feedback. Personally, I prefer client side feedback even though it shows false positives occassionaly.
Is there a new update ?
Amory menu and kill feed are broken, and the hitreg is bad again (not as bad as last time though). It's on every server, even without mods.
Edit : looks like the problem disappeared, we had this on 3 different servers hmm...
@Pikum
1st clip - you were already dead, looks like a latency issue there
3rd clip - you were miles away when you bit, look at the first frame the bite animation starts
The other two clips look like your bullets practically skimmed the model, even a little network fuckery could cause this. I agree though that client side blood should ideally = server side damage but that may not be 100% possible all the time
Replay the ones where you "hit" in slow motion and I think you'll find that you were really missing. If you're using NS2+, turn on "Server confirmed blood" to have 'better' feedback. Personally, I prefer client side feedback even though it shows false positives occassionaly.
You're right, on the second clip you can also see the bullet hit the ground behind the robotics factory after passing through the fade's shoulder. I've had other people explain this kind of thing with the whole "the end of your gun barrel is behind the alien's model, and you're kind of shooting behind them instead of at/through them" explanation, which is semi-believable with the way rendering things in a 3d space works.
However, that is not the case with the 1st, 3rd, and 4th clips.
I would try turning on "server confirmed blood" in NS2+, but wouldn't that just make it exactly the same as the beeps? Isn't that ignoring the issue? The game is giving the player feedback that they're getting hits when they're not... whether or not the shots are "legit" server-side means one of two things:
1. what the player is seeing is wrong and they aren't able to aim at something appropriately because what is being rendered on their screen is a lie.
2. what the server is receiving from the client is wrong
I get that the client is going to be different because you can't be perfect about the game state across every client every frame/tick, but if the client didn't think i was hitting, the blood wouldn't splatter...
I assume that the game is probably using a different algorithm unrelated to bullet physics for blood splatters, but then why use that at all if it's going to give the user false positives on hit detection?
Maybe this could be an option or a setting? Let people with computers that can handle it use a more taxing/advanced/accurate client-side algorithm for hit-box detection for blood splatters? Or maybe swap them on the fly if the FPS goes below a set number?
Relying on server-side calculation only for player feedback seems about on the same level as saying "I'm not having issues so it must not be happening." Obviously if the server doesn't think they're hits, they're not going to register as hits.
Is there anything else I can do here? I'll go ahead and give the server-validated blood splatters a shot and see how i like the feedback on it... could you explain this though? isn't that just going to only show the splatters if the reg went through? isn't that the same as the audio beep? Won't i still see the bullets hit the model but then just not beep/splatter?
@Pikum
1st clip - you were already dead, looks like a latency issue there
3rd clip - you were miles away when you bit, look at the first frame the bite animation starts
The other two clips look like your bullets practically skimmed the model, even a little network fuckery could cause this. I agree though that client side blood should ideally = server side damage but that may not be 100% possible all the time
1st clip - I think you're right, and I wondered that myself. I had another clip that i omitted before posting the video that was similar in that it was near-death, but I thought the delay on the audio on the first few hits was worth leaving this one in even if I was already dead for the shots at the end of the clip (never have been able to figure out an accurate count on those early hits).
3rd clip - you're right, the first bite after the parasite I was too far, but after lunging forward for the second bite, you hear the sound of the bite connecting with a "munch" sound and see a blood splatter, but there's no "beep" or reg. The 3rd bite did register, and obviously the 4th bite i was already dead even though the "munch" sound played
4th clip - you can clearly see 4 shots cause blood splatter while only one registers... You also cannot see the bullets hitting something behind like you can with what i discussed above in the second clip. saying it's a network issue is essentially proving my point that there are reg issues, unless you mean to say it's an issue with my internet connection and not the game's code, which I'm willing to accept if there's a solid way i can reproduce it consistently while others cannot. (any good way to do this?)
Any other thoughts here? if there's any more info i can provide, let me know... in the meantime I'm going to try to keep putting these together... that video was just one day after i got sick of everyone saying there's no hitreg issues
AsranielJoin Date: 2002-06-03Member: 724Members, Playtest Lead, Forum Moderators, NS2 Playtester, Squad Five Blue, Reinforced - Shadow, WC 2013 - Shadow, Subnautica Playtester, Retired Community Developer
I think there is one thing happening in the video as well. If you shoot something, just before you die, you can get differences between client/server. Simply because for the server, you are already dead, even if on your screen you still life. the time window for this is quite short, but it can easily go as high as 300-400 ms, depending on your and your attakers ping.
There's only two instances of actual hitreg failure in your video, @ 13 sec where you shoot the back portion of the fade moving towards you, and @ 42 sec when shooting the trailing portion of the skulk.
These both demonstrate one of the two known remaining issues, the "trailing model" issue. (see my video that I posted in this thread prior).
The rest are what others have said, due to interpolation, and would happen even with flawless hitreg. If you were running the hitreg tool, the console output would've said "Acceptable Diff found - client already dead on server".
Thank you for making the video tho, and if you ever get the chance for an admin to run the sv_tests command - or you host your own server with a buddy - feel free to turn on hitreg in the console and see if you can find any new issues.
Comments
@IronHorse : after this little trip into docking guts (up to 6 hours); i have close to 500 reasons to doubt (you can check it actually). It made me think of a war zone or somebody who had a HDD crash and copy pasted bit by bit the whole thing.
Refinery? Check mapping section...
It is; after looking into each official map (old and new), legit to think that maybe this boat don't have a captain any more.
Creation and success are ugly sometimes, and often disturbingly selfless and humble. The "war zone" you see has genuinely entertained hundreds of thousands of people, and it didn't follow habitually idle disparagements of the status quo, as I see you trapping yourself with. My guess, rather, is that it followed a lot of hard, quiet, and very frustrating work that genuinely excited its caretaker(s) and stakeholders, with a mix of open-minded learning and collaboration along the way.
I see you've created a map before, so you're in a better position than most NS2 enthusiasts to create something compelling! I want my server's community to send me a link to /your/ Steam Workshop, so we can be excited about putting /your/ "war zone" on TGNS.
If you put into learning and doing half the energy you seem to be putting into critiquing and justifying personal defeat, I think you could do great things. You don't need a captain -- you ARE the captain. You need to decide you're going to create something memorable and you're going to be successful -- you (and the future consumers of what you'll long be remembered for, as opposed to these forum posts) would be very well served by your starting now to create and collaborate in that spirit /exclusively/.
No more excuses. You can do it, and there are lots of people here who will help.
A video content lacks a hundred million times the necessary information to debug this kind of thing and of course the other as well. Ask a programmer. Recording client and / or server data with a debug level setting will ensure better results in no time compared to rudimentary video upload that requires huge online space and leaving a PC uploading it for a full night (i got porn to DL ). I was not talking about 279 specifically. The core dump uploader is still in a bad shape it seems. If anything a developer should gather is critical info. Why is the necessary tools (automated), and corner stone for this, broken ? That would be the first thing to focus on in any development process (game or any other stuff). First thing first.
As for the warzone in Docking; open the editor, see for yourself. It's a war zone. I don't know if Docking was fun to make, the thing is, it doesn't look so if you wander in the editor with that file opened (not the art the technical). Trust me, my hobby production are clean and I have no orphan stuff in there. I wish it was possible to say the same for many official maps that would enable ANY modder (beginner as senior) to edit fast and easy and hopefully improve the whole thing.
I do criticize because as said before by me and others. Each time you can push forward (or have something on the verge of blossoming), we just see ourselves put on the side of the road with nothing but a kick in the balls. What about the Faded, Siege, Last stand (last echo from 2013 it seems) ? They all suffered many trouble not originating from their captain bad bearing skill. So... NO WE CANNOT BE THE CAPTAINS. Something else prevents it.
I'm not talking about rewards or anything related to ego. In fact my friend; being a modder for more than a decade in several games, you would be surprised to see how far beyond I'm am from what you suggest / imagine.
I see many +1/-1 people more than anything. Follower are followers. It looks like K-POP fans groups pushing the disagree button each time someone says something relevant and backed up by facts and a sound logic. And I exclude myself from this last argument. I've seen on this very forum many legit claims and suggestion blown up into pieces by zealots who didn't understood what the author was talking about due to a clear lack of education / thinking. But guess what ?; they whine more, they talk louder and in numbers...
In the end all those people are mouths (as myself you would say), you never see hands at work (not as myself, right?). The most derelict thing is that they could just help with 15 minutes of their time while they pass hours if not days playing games. You just can't ask something to a gamer. It's like asking a kid to tidy the room.
Start to read each of the custom map threads and see for yourself how much feedback they got. You can do it for the modding section as well. I guess +/- button are faster to press.
You want to test on TGNS... great!!! that will make 2 servers up with viable custom maps (maybe 3 - depending on ping/location). I have a list you can run and use in this thread. All reviewed and QC passed by your highness. Trust me when i say 3 of them are far beyond what the official did. I exclude my map from this last one of course.
"No more excuses. You can do it, and there are lots of people here who will help."
That made me laugh for 1/4 hour and made my day... sh|t I'm in a train... They all look at me now... I'm yet to see hands to finally enhance the custom offer in there.
Good one non the less. I like you already just for that one.
As for the debug output, I've already explained how to do that and what it looks like in one of the many videos I've made. We need a video AND the debug output.
Why you persist in assuming what we've requested from you as not being sufficient for us to analyze, just makes no sense... we know what data we need, so leave that judgement up to us.
It's been on the board for a month now.
I don't get the point of the rest of your post, honestly, and I read it twice.
Just seems like general discontent with no specific examples of what the developers could do to to address it, besides better modding tools?
Read the comment he links ('how far beyond)'. It's informative & amusing.
Still: Loads of talented people wasting time here.
The second on the list. If you find something wrong in terms of grammar or vocabulary (colloquial or not), feel free to help. I'd be glad to get rid of the many mistakes i do / bad habits i must have. Languages change faster than before internet. Especially English it seems.
When are you going to post your findings?
You can lead a horse to water...
You need 3 outputs : 2 clients and a server in the same situation I described and that are synchronized with NTP. At least each machine have to know the time diff they have between each other.
Plus (cherry on top) the analysis of network paths (ex: traceroute A to B and B to A) as it can be different but also A to B may vary from time to time... So you have to check & log it too... For each paths. Try Matt's traceroute and play with it and you will see how it can be tricky sometime to just reach google.
Not to mention the DXDIAG, the NS2 settings and all that can be logged should be taken under consideration.
And finally make a cross analysis of the 3 logs step by step (at least the relevant ones). That's why time is important in this matter.
This log factory has to be industrialized / programmed for server side and client side. Also it needs an automated analyzer. Humans are slow readers. It means a long list of things to do.
There's nothing in NS2 right now that can do that to my knowledge. It would give many clues on every other issue related to networking (or discard options) if it is the culprit (or one of them). It may be something else. Doing this debug the proper way may not reveal whats wrong right away or focus on this specific issue.
This is the kind of functionalities that should have been implemented not long after the game was out or before... Why ? Because it also helps to create stats on the long run to see what problem occurs the most. In this case a statistician would be helpful.
BTW I'm not specifically targeting b279. This "thing" has been around like a ghost for some time now and happening in some specific situation. It tends to occur slightly more often (Gut feeling).
First thing first. Any other way is pointless and will make people loose their time.
I don't know if you tried to edit it out, but your link is still there. Still funny.
Will play again if next patch reads: "Fixed bad old days performance issues"
The toolset and debugging features you describe would be amazing. They would greatly benefit ns2. The sad truth is that no matter how correct you are, we are not going to get those tools. UWE has stated quite clearly they are interested in other things.
I know you don't like it. I don't like my 1988 honda either, but that also will not change anytime soon. There is only one thing left to do. That is work with the tools we do have. If all we have is duct tape and a few sticks then that is what we use. Same with the debugging toolsets we have.
Ironhorse has described to you how the UWE/CDT programmers want data to debug with. You are saying you are able to reproduce issues consistently. You seem to be very good finding bugs no one else can. Would you please share your abilities with the programmers by finding the bugs and sending them the data they want? It may not be all the data they need, but it is all they have to work with right now. Could you be that generous?
The same goes to you. No one else is reporting such problems. You are the only one. We can't reproduce your problems without your help. Could you be so kind as to help us find the problem? It would not even take too much of your time.
If you could be so kind to follow these instructions and make a thread in the tech support forums it would be extremely helpful.
then you can p_endlog /or just wait for the round to end/ or map to change/ or just quit the game and it will end automatically.
Then navigate to your hidden folder
c:/users/YOU/appdata/roaming/natural selection 2/
And you will find your file, something *.plog
Zip it/ compress it.
Upload it to somewhere on the internet
Make a thread in the tech support forums, and put a link to your .plog file there.
This will give me a frame by frame analysis of what is going on, so try to pick a time when you can best reproduce your lowest performance like the 24 fps you said you get.
Don't record for longer than 30 minutes, and don't record for shorter than 5.
Include your techsupport.zip file with this compressed plog.
This was all on DMD Dallas 2, which does pop up the warning bout server settings being modified.. However, I can say for sure that this happens on other servers too for me.
I was not running the console command, didn't know about it until I started reading through this thread.. I'll try to use it in the future whenever it's available.
I have absolutely no idea how to reproduce these consistently, but hopefully someone else can pull something from this.
Replay the ones where you "hit" in slow motion and I think you'll find that you were really missing. If you're using NS2+, turn on "Server confirmed blood" to have 'better' feedback. Personally, I prefer client side feedback even though it shows false positives occassionaly.
Amory menu and kill feed are broken, and the hitreg is bad again (not as bad as last time though). It's on every server, even without mods.
Edit : looks like the problem disappeared, we had this on 3 different servers hmm...
I was gonna say it was a glitch where blood was showing when it shouldn't,but your bullets (and teeth) were hitting them,I dunno what to say
1st clip - you were already dead, looks like a latency issue there
3rd clip - you were miles away when you bit, look at the first frame the bite animation starts
The other two clips look like your bullets practically skimmed the model, even a little network fuckery could cause this. I agree though that client side blood should ideally = server side damage but that may not be 100% possible all the time
You're right, on the second clip you can also see the bullet hit the ground behind the robotics factory after passing through the fade's shoulder. I've had other people explain this kind of thing with the whole "the end of your gun barrel is behind the alien's model, and you're kind of shooting behind them instead of at/through them" explanation, which is semi-believable with the way rendering things in a 3d space works.
However, that is not the case with the 1st, 3rd, and 4th clips.
I would try turning on "server confirmed blood" in NS2+, but wouldn't that just make it exactly the same as the beeps? Isn't that ignoring the issue? The game is giving the player feedback that they're getting hits when they're not... whether or not the shots are "legit" server-side means one of two things:
1. what the player is seeing is wrong and they aren't able to aim at something appropriately because what is being rendered on their screen is a lie.
2. what the server is receiving from the client is wrong
I get that the client is going to be different because you can't be perfect about the game state across every client every frame/tick, but if the client didn't think i was hitting, the blood wouldn't splatter...
I assume that the game is probably using a different algorithm unrelated to bullet physics for blood splatters, but then why use that at all if it's going to give the user false positives on hit detection?
Maybe this could be an option or a setting? Let people with computers that can handle it use a more taxing/advanced/accurate client-side algorithm for hit-box detection for blood splatters? Or maybe swap them on the fly if the FPS goes below a set number?
Relying on server-side calculation only for player feedback seems about on the same level as saying "I'm not having issues so it must not be happening." Obviously if the server doesn't think they're hits, they're not going to register as hits.
Is there anything else I can do here? I'll go ahead and give the server-validated blood splatters a shot and see how i like the feedback on it... could you explain this though? isn't that just going to only show the splatters if the reg went through? isn't that the same as the audio beep? Won't i still see the bullets hit the model but then just not beep/splatter?
1st clip - I think you're right, and I wondered that myself. I had another clip that i omitted before posting the video that was similar in that it was near-death, but I thought the delay on the audio on the first few hits was worth leaving this one in even if I was already dead for the shots at the end of the clip (never have been able to figure out an accurate count on those early hits).
3rd clip - you're right, the first bite after the parasite I was too far, but after lunging forward for the second bite, you hear the sound of the bite connecting with a "munch" sound and see a blood splatter, but there's no "beep" or reg. The 3rd bite did register, and obviously the 4th bite i was already dead even though the "munch" sound played
4th clip - you can clearly see 4 shots cause blood splatter while only one registers... You also cannot see the bullets hitting something behind like you can with what i discussed above in the second clip. saying it's a network issue is essentially proving my point that there are reg issues, unless you mean to say it's an issue with my internet connection and not the game's code, which I'm willing to accept if there's a solid way i can reproduce it consistently while others cannot. (any good way to do this?)
Any other thoughts here? if there's any more info i can provide, let me know... in the meantime I'm going to try to keep putting these together... that video was just one day after i got sick of everyone saying there's no hitreg issues
thanks for the comments so far!
Thanks for the vid.
There's only two instances of actual hitreg failure in your video, @ 13 sec where you shoot the back portion of the fade moving towards you, and @ 42 sec when shooting the trailing portion of the skulk.
These both demonstrate one of the two known remaining issues, the "trailing model" issue. (see my video that I posted in this thread prior).
The rest are what others have said, due to interpolation, and would happen even with flawless hitreg. If you were running the hitreg tool, the console output would've said "Acceptable Diff found - client already dead on server".
Thank you for making the video tho, and if you ever get the chance for an admin to run the sv_tests command - or you host your own server with a buddy - feel free to turn on hitreg in the console and see if you can find any new issues.