Re: [hlcoders] Player-specific Hitbox Bug? (Help, Alfred? Help, valve? Reproduction steps included.)
Thanks for the reply, Andrew. I think I understand what you're saying (I did play on and help administrate a few CS servers a while back), but I don't believe I'm talking about the same thing. It's not a constant thing - only the last player in a full server is like this. Basically, what I'm seeing is that a fast-moving normal player can be shot by direct hits in a nearly ideal environment (100Mb/s full-duplex switched network), but the bugged player will have to be lead by a certain amount. This amount seems to be directly related to the velocity with which the bugged player is moving. If all players were acting similarly, I could chalk it up to some kind of erroneous prediction. However, it is only the player in the last slot of the server. (In an n player server, player #n will be bugged.) Basically, I am seeing two seemingly identical players move in exactly (Or as nearly so as is humanly possible) the same way with different results. One player will be directly hittable, the other will have to be lead. The server can be restarted, players switched; it doesn't follow any specific player, only the slot that a given player falls into. Andrew Foss wrote: No, you're seeing this, and you are sane. It's common in Counter-Strike servers to have to lead your targets for high deflection shots. IE: when the player is moving left to right rapidly across the FOV. On Fri, 23 Jul 2004 21:46:34 -0700, Arithon Kelis [EMAIL PROTECTED] wrote: I apologize in advance for the cross-posting. I scanned through my archive of the hlds mailing list, but did not see anything related to this issue. It appears, after careful study, that there exists a discrepancy between a player's appearant position and their real position that depends on the player's ID relative to the server. This has been observed in the base HLDM, DoD, and several third-party mods. CS was not tested. Description: This bug allows a player to be seemingly invulnerable - shots fired directly through their body (as seen by the client) will not affect them at all, but shots fired a certain distance ahead of them (depending on player speed) WILL hit them. This only affects the player in the last slot. Reproduction steps: 1) Start a 3-person dedicated server. 2) Three people need to connect. The player to connect last will be the bugged player. 3) Pick one of the first two players to be a control player, the other will be the designated gunner. 4) Have the control player run around. Gunner should CAREFULLY aim and fire ahead, behind, and directly upon the control player. Only the shots that are aimed directly at the control player should hit. 5) Have the bugged player run around. Gunner should CAREFULLY aim and fire ahead, behind, and directly upon the bugged player. Only the shots that are aimed AHEAD of the bugged player should hit. Notes: - Every person we have introduced to these reproduction steps has been able to verify our findings. - Network latency seems to have no effect; the bugged player acts the same to players with a 2.7ms ping as it does to players with a 120ms ping. - This does not SEEM to be a mod-related problem. Every mod tested, wether official or third-party, has displayed the same behaviour. - It is only the player in the last slot that displays this behavior. I'd appreciate some feedback on this. Am I simply going senile, or am I really seeing this? Thanks in advance, --ari ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
Re: [hlcoders] Player-specific Hitbox Bug? (Help, Alfred? Help, valve? Reproduction steps included.)
I'm here to just put my 2 cents in. I'm actually one of the new coders for NS and have been working at finding a solution for our hitbox issue that's happening with player N of a server with maxplayers set at N. I have been working with Arithon on reproducing it and narrowing down what was happening. At first we thought we just had a player who was somehow flagged invincible, but it only seemed to happen when he was moving. Repeated testing proved that the bugged player's hitboxes were simply projected farther in front of the model, in the same direction of movement. In an effort to help eliminate variables, we tried going back and recreating the bug on our current beta build in Won Half-Life. We were successfully able to reproduce it, using Ari's methods already listed. Then, we went the extra step and tested other Half-Life mods. We were successfully able to reproduce getting a bugged player on Counter-Strike, DoD, and HLDM in Steam. Additionally, we were able to recreate it in Counter-Strike 1.5 and HLDM in Won Half-Life also (didn't try DoD). We do have demos available for just about every test and are willing to produce them if people are having problems recreating a bugged player situation. I don't believe it should be an issue. We also found a strange behavior about how you see the bugged player. If you have a spectator or client running a listen server and they're watching player A shoot a bugged player B, they will see player A's aim and shots landing behind player B. Meanwhile, player A will see his shots hitting player B's model on his client, but no damage being taken. However, if player A shoots in front of the model he sees for B, his shots will register and will look like he's aiming directly at the model that spectators/listenserver sees. This appears to be consistently the same thing happening across mods. The paradox is such: only the shooter sees the problem. To everyone else, he just looks like his aim is off His shots will appear to miss (shots that are lead), but will hit and do damage. Everyone else will see it like a normal shot. His shots that appear to hit will not deal damage. Everyone else will see it like a missed aim behind the player. We really noticed this with projectile weapons, such as acid rocket or gorge spit in NS, but it seems to work with any weapon - hitscan or not. It really looks like this bug of sorts has been around for a while, but doesn't seem to have been found out before. Tom Grim aka Elven Thief --- Arithon Kelis [EMAIL PROTECTED] wrote: I apologize in advance for the cross-posting. I scanned through my archive of the hlds mailing list, but did not see anything related to this issue. It appears, after careful study, that there exists a discrepancy between a player's appearant position and their real position that depends on the player's ID relative to the server. This has been observed in the base HLDM, DoD, and several third-party mods. CS was not tested. Description: This bug allows a player to be seemingly invulnerable - shots fired directly through their body (as seen by the client) will not affect them at all, but shots fired a certain distance ahead of them (depending on player speed) WILL hit them. This only affects the player in the last slot. Reproduction steps: 1) Start a 3-person dedicated server. 2) Three people need to connect. The player to connect last will be the bugged player. 3) Pick one of the first two players to be a control player, the other will be the designated gunner. 4) Have the control player run around. Gunner should CAREFULLY aim and fire ahead, behind, and directly upon the control player. Only the shots that are aimed directly at the control player should hit. 5) Have the bugged player run around. Gunner should CAREFULLY aim and fire ahead, behind, and directly upon the bugged player. Only the shots that are aimed AHEAD of the bugged player should hit. Notes: - Every person we have introduced to these reproduction steps has been able to verify our findings. - Network latency seems to have no effect; the bugged player acts the same to players with a 2.7ms ping as it does to players with a 120ms ping. - This does not SEEM to be a mod-related problem. Every mod tested, wether official or third-party, has displayed the same behaviour. - It is only the player in the last slot that displays this behavior. I'd appreciate some feedback on this. Am I simply going senile, or am I really seeing this? Thanks in advance, --ari ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders __ Do you Yahoo!? Yahoo! Mail - Helps protect you from nasty viruses. http://promotions.yahoo.com/new_mail ___ To unsubscribe,
[hlcoders] Re: hlcoders digest, Vol 1 #1374 - 10 msgs
I started a debug session with hlds. I joined my server, typed reconnect, the server crashed and the debugger pointed me to this line where m_pCompany-m_flAmmoReplication was used without checking that m_pCompany wasn't NULL first. Ooops :( Message: 7 From: voogru [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: RE: [hlcoders] reconnect crash Date: Fri, 23 Jul 2004 16:55:41 -0400 Reply-To: [EMAIL PROTECTED] What did you do? I've experienced this crash as well. My fix was applying the EF_NODRAW flag to the player in the ClientDisconnect (after all the other disconnect functions are called). -Original Message- From: Sylvain Gravejat [mailto:[EMAIL PROTECTED] Sent: Friday, July 23, 2004 10:26 AM To: [EMAIL PROTECTED] Subject: Re: [hlcoders] reconnect crash Ok, I thought retry was just performing reconnect a few times. I debugged a bit and managed to fix the bug anyway. Thanks. Reconnect is more of a full resource propagrtion, much like an connect retry just checks the basics, and the loads you into the game --- Sylvain Gravejat [EMAIL PROTECTED] wrote: What's the difference between the reconnect and retry commands? I'm having a major bug where a player performing reconnect crashes the server. retry however is working normally. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders = hl2hosting.com AIM SN - amaperman __ Do you Yahoo!? Yahoo! Mail Address AutoComplete - You start. We finish. http://promotions.yahoo.com/new_mail ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
Re: [hlcoders] Player-specific Hitbox Bug? (Help, Alfred? Help, valve? Reproduction steps included.)
I think a Valve employee should do Find in Files on gpGlobals-maxClients. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
Re: [hlcoders] Player-specific Hitbox Bug? (Help, Alfred? Help, valve? Reproduction steps included.)
Tom Grim wrote: We were successfully able to reproduce getting a bugged player on Counter-Strike, DoD, and HLDM in Steam. Additionally, we were able to recreate it in Counter-Strike 1.5 and HLDM in Won Half-Life also (didn't try DoD). I know CS uses 9 point blending, I believe DoD does as well. I'm not sure if HLDM uses 9 point blending under Steam (I know it didn't before Steam came along). I remember Alfred and Eric describing a problem with 9 point blending where client hit boxes wouldn't be synched up properly with server hit boxes (where hits are really determined). Also I remember discussions on thewavelength.net about 9 point blending having problems with players that are moving rapidly (that may have been linked to players who rapidly switch from one gaited animation to a different gaited animation). The client side prediction of skeletal bone locations (which determine hit box locations) didn't seem to match up with the server's calculations for any MOD who had very high speed player movement. CS movement is relatively slow (compared to deathmatch), so they descrepency between the client and server wouldn't be as much since the players couldn't move that far in space over a small amount of time. Perhaps there is something about the order of the players in the server that is determining when client/server data is replicated to the other side (i.e. the higher the client index, the more that player gets starved of network replication. Of course, if this were true, you wouldn't see this problem at all with the last player on a 2 player server and you would see it 100% of the time with the last player on a 32 player server. -- Jeffrey botman Broome ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
[hlcoders] hlds ping
can you help me!! ive made a server but at each time someone join the game his ping is really high and he lags how can i fix that!! Thx patrick _ Scan and help eliminate destructive viruses from your inbound and outbound e-mail and attachments. http://join.msn.com/?pgmarket=en-capage=byoa/premxAPID=1994DI=1034SU=http://hotmail.com/encaHL=Market_MSNIS_Taglines Start enjoying all the benefits of MSN® Premium right now and get the first two months FREE*. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
Re: [hlcoders] hlds ping
[ Converted text/html to text/plain ] Three questions: How fast is your computer? If you are below 1 GHz, that could be the problem. How are you connected to the internet? If it is with a 56K modem, your server will lag. You probably shouldn't be running a server without DSL or cable modem. What mod are you running, and does it have bots? Bots can slow down a server. -Scott patrick racicot wrote: can you help me!! ive made a server but at each time someone join the game his ping is really high and he lags how can i fix that!! Thx patrick _ Scan and help eliminate destructive viruses from your inbound and outbound e-mail and attachments. http://join.msn.com/?pgmarket=en-capage=byoa/premxAPID=1994DI=1034SU=http://hotmail.com/encaHL=Market_MSNIS_Taglines[1] Start enjoying all the benefits of MSN® Premium right now and get the first two months FREE*. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders[2] ===References:=== 1. http://join.msn.com/?pgmarket=en-capage=byoa/premxAPID=1994DI=1034SU=http://hotmail.com/encaHL=Market_MSNIS_Taglines 2. http://list.valvesoftware.com/mailman/listinfo/hlcoders ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
Re: [hlcoders] huge debug problem
/NODEFAULTLIB:libcmtd.lib... try 500 unresolved externals. I formatted not long ago, installed vs6 sp 5 downloaded sdk2.3, compiled it and set it up for the debugger put one line of code in to crash it pPlayer=NULL then I deferenced pPlayer and it gives me a mscvr70 error and wont point me to the line. How the hell did it ever work? Nothing has been changed from default settings. So bascily if my game crashes, I'm fucked because I dont know where or why it crashed. I have over 10,000 new lines of code in my game easy, so I really do need the debugger to work. I cant afford microsofts premium prices so ill have to get an illegal version of .net. Its just a pain in the ass for nothing because 6.0 was working fine and I even own a legal copy. Quoting cosh [EMAIL PROTECTED]: so...did you try /NODEFAULTLIB:libcmtd.lib ? :) [EMAIL PROTECTED] wrote: That seems logical but it did not work for me, i guess ill just have to find a copy of .net however i do get this build warning LINK : warning LNK4098: defaultlib LIBCMTD conflicts with use of other libs; use /NODEFAULTLIB:library I think that has something to do with it. I got that warning when i added msvcrt.lib to the link text box Quoting cosh [EMAIL PROTECTED]: hey, i already went through the trouble of finding out why my projects always need that damn msvcr70.dll and not the default msvcrt.dll well the solution is quite simple. the msvcrt.lib that comes with vc++6 references to the *default* msvcrt.dll (every operating system since windows 95 should come with it) but the msvcrt.lib that comes with vc++7 references to the msvcr70.dll. and i heard that visual studio 2003's msvcrt.lib references to msvcr71.dll so if you want to compile projects that just need msvcrt.dll you need to get your hands on the msvcrt.lib that comes with vc++6. you always just link with msvcrt.lib but every version of vc++ references a different .dll file for it. hope that helped :) [EMAIL PROTECTED] wrote: What if i just install .net. Do guys think that might solve the problem? It seems like that 70 dll is from the .net version and somehow vs6 wants it for my hl project. I have looked thu the project settings for hours without any luck. I have changed a lot of the #include statements but mabye i should just update. I guess i have a real excuse now. btw, is it pretty easy to upgrade? Quoting tei [EMAIL PROTECTED]: Its not hard to code a strcat and other small standard functions. Maybe you can code these functions in your source body, or search for a free replacement compatible with the SDK_EULA. IF you can avoid extra downloads, or extra pre-requisites, the installation and management will be much smoother... Shit Happends,.. K. Mike Bradley wrote: If you install the .NET framework I think this dll come in that. Run your browser and select Tools, Windows Update. It will be one of the Windows updates offered. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of [EMAIL PROTECTED] Sent: Sunday, July 18, 2004 5:38 PM To: [EMAIL PROTECTED] Subject: Re: [hlcoders] huge debug problem Well I installed the sdk and compiled that and added a line to make it crash and still got that missing dll thing. I don't know why it wants mcsrv70.dll I only have mcsrv40 and 20 on my pc. Why does it ask for 70?. I just formatted and reinstalled vs 6.0 sp4. What is the deal with the sdk? My other coding projects work fine but this has never happened before with the hl sdk. I have never really edited the debugging libs before so I don't really know how I messed things up. Quoting [EMAIL PROTECTED]: I thought I was using the std libs? The only thing i can think of is perhaps im using the old header files iostream.h not iostream. Like I said beofre. I rebuilt my project from the stock SDK so no project settings are changed. Only the code. Quoting Jeffrey \botman\ Broome [EMAIL PROTECTED]: [EMAIL PROTECTED] wrote: lately i have not been able to debug correctly because whenever I get a crash visual stuio tells me it cant find some stupid dll file called msvcr70.dll. You should use the Visual C++ Debug Runtime libraries... http://msdn.microsoft.com/library/default.asp?url=/library/en- us/vccore98/html/_core_c_run.2d.time_debug_libraries.asp ...then you will have source code in your traceback and can see where you are going wrong. -- Jeffrey botman Broome ___ To unsubscribe, edit your list preferences, or view the list archives,
