|
Re: POD-bot back into shape.
Quote:
1) Downloaded the BU-hl-bot-build-example.tar.bz2 from the filebase 2) I used this utility to unzip all files and folders structure to my c:\ drive http://utility.downloadatoz.com/zipzag/ I renamed the base directory to: C:\HLDev 3) I modified the /I directives in the podbot_mm.dsp file to point to the correct locations and the bot built fine. Here is the include line. You can edit the podbot_mm.dsp file directly if you are careful. # ADD CPP /nologo /G5 /MT /W3 /WX /GX /O2 /I "\HLDev\metamod\metamod" /I "\HLDev\hlsdk-2.3\singleplayer\dlls" /I "\HLDev\hlsdk-2.3\singleplayer\common" /I "\HLDev\hlsdk-2.3\singleplayer\engine" /I "\HLDev\hlsdk-2.3\singleplayer\pm_shared" /I /D "NDEBUG" /D "WIN32" /D "_WINDOWS" /Fr /c Bots-United: You don't say what version of metamod is included in BU-hl-bot-build-example.tar.bz2 It would be helpful to know. I am assuming 1.17????? Thanks! >edit< The bug fix was for the bug sPlOrYgOn found with the comamnds. I don't know why it would be 20K bigger. It is a release build. |
Re: POD-bot back into shape.
thanks for the updated menus Austin :) what makes the version you made 20 kb bigger???
what command thing did you fix? did i miss something? also i was making my sf_columns2.remaking sf_columns......as an AS_ and an ES map as one......both teams with a mission like csde or so..... im playing the map with PMB's last update......and 2 times in a row a few rounds into the match i get a HL crash reading this message: sz_getspace: overflow without FSB_ALLOWOVERFLOW set on server reliable datagram what in the world is that? anything to do with my double missions not being compatible or what is this? |
Re: POD-bot back into shape.
I can perhaps help to locate this one.
This message means that the bot tried to send too much of "special" network messages to your client. Special network messages are messages that are not handled in the standard "pack" that makes up the HL network protocol. These special messages can be player chat, or temporary entities which are displayed server-side (not client-side temporary entities like smoke puffs or gunshot decals). An example of these are the beams that are displayed in the waypoint editor. Each beam is a temporary entity with a lifetime of 1 second, and the bot re-creates them every second. Your computer crashed because the bot tried to send too much of them to your client in the same frame. If you were running with the waypoints enabled, try not displaying the waypoints next time. If it still fails, try disabling bot chat. Such crashes can also happen randomly if the program writes at incorrect memory offsets. This happens for example when the programmer attempts to write beyond the limit of an array. There is one thing I can think of that I have changed in the bot source code, that might cause that: I've changed the team number identifiers (they were 0 for T and 1 for CT) to 1 for T and 2 for CT, so as for the bots to use the same numbers when selecting their team in the menu. It's not a big deal, but I thought this was cleaner. I ensured that everytime these team numbers were used to index an array, I was substracting 1 in order to get them back into the 0-1 range, but maybe I've missed one out. It can be true for the radio chat times array, for example. Check it out, coders :) |
Re: POD-bot back into shape.
i seem to get it when someone escapes......a terrorist or the vip.......then crash......
i have botchat already disabled so i could see the stat things from AMX, and i didnt have the waypoint editor on........hmmm thanks for the info though PM :) heres another HOST_ERROR im getting while trying to load de_outrage which has yet no waypoint: PF_precache_model_l: "sprites/top_right3.spr" overflow its only giving me this message on de_outrage......the sprite is part of the HUD display im think.......its not used in the map itself |
Re: POD-bot back into shape.
Quote:
DON"T!!! Muaahahahah! Thanks for the lead I will take a look... |
Re: POD-bot back into shape.
does this .dll work any better?
edit: updated .dll again. edit2: If this works, I wouldn't know why. *slaps head* get some sleep MusicMan! |
Re: POD-bot back into shape.
I would like to take over but I'm not too experience in C++ or hl coding so it'll take a long time before i can get used to this stuff and all.
I also cannot make linux builds because my linux has glibc2.3 and it gives a bad load when trying to load a plugin i compile. Hopefully I can get used to the code and navigate it faster to fix bugs when someone finds a bug. |
Re: POD-bot back into shape.
musicman what did you do to this dll before i install it??????
this is going to get crazy i can see that now......everyone is going to do something to this bot.....separately......and it will eventually be ruined. get yourselves organized on whos gonna do WHAT. PMB made tons of changes......and the dll had the same file size almost that entire time. ive seen 2 new dll's from different people today......one was bigger by 20 kb and the other was over 100 smaller......what are you guys putting in/taking out? im really afraid to touch anything PIEEEEEEEEEEEEEEERE!!!!!!!!! hehe |
Re: POD-bot back into shape.
yea when i compile without editing the source at all i get a 256kB dll... while PMB's is like 244kB...
PMB's gotta tell us how he optimized the compiler and all :D [edit] this is probably when a bots-united cvs repository would come in handy... and I only plan on supporting who ever takes full control over the bot. [/edit] |
Re: POD-bot back into shape.
NEXT SERIOUS PROBLEM TO SOLVE
stop the bots from hitting buttons because of automatically placed buttong use flags. let us control them hitting what buttons with a flag we can place. that automatic button flag ruins a map like de_747. thats the most frustrating shit i ever waypointed(besides cs_bikini and its illusionaries) half the terrorist fall out of the plane because of a button. strangley theres no waypoint around that has a use_button flag on it. so its something else and they should stop that. also stop auto connecting through a model...like we did with the illusionaries. |
Re: POD-bot back into shape.
this going to be hard...
and while looking through the WaypointAdd function i notice that when setting a camp waypoint's camp end angles it doesn't set bPlaceNew to FALSE and that would make it place a new waypoint along with the camp end right? [edit] oh nm.. i should probably look closer first... [/edit] [edit2] k i found where he stops it from adding through func illusionary but I don't know what it returns if it goes through a model.. [/edit2] [edit3] alright now i can compile the podbot.dll to be 232kB in size. goto Projects>Settings. then the C/C++ tab Change the Optimizations to "Minimum Size" [/edit3] [edit4] shouldn't all the TRUE and FALSE be true and false? [/edit4] |
Re: POD-bot back into shape.
Post code, guys, post code, whenever you find anything, post 10 lines of code or so. I can help in anything provided you post your code.
TRUE and FALSE are defined in the HL SDK as 1 and 0, which is fine for use with "bool" since "bool" has in fact a character size (8-bit) on our computers. Use my PMTools plugin to trace lines and hulls through one of these models with different parameters to find out which parameters are blocked by the model. *edit* just a suggestion: SF is right, you'll get yourselves all messed up. As long as nothing like the SourceForge CVS is available on Bots United I suggest you define a "team leader" in charge of maintaining the "official" source tree and validating the bugfixes you send him. PLEASE for him, send him ACCURATE bugfixes (e.g, "change line x to y and insert code z at line l") and not "I thought you could do this and that etc..." with no real, working code to show. |
Re: POD-bot back into shape.
deleted post myself.
MusicMan |
Re: POD-bot back into shape.
Now the only code i can think of that wasn't posted here when the source was edited is probably what austin did to his waypoint menus.
|
Re: POD-bot back into shape.
MusicMan, why the heck do you do this ?
The definition of UTIL_ShowMenu is void UTIL_ShowMenu (edict_t *pEdict, int slots, int displaytime, bool needmore, char *pText) the third parameter is the time the bot menu should be displayed on your screen. -1 means 'always', i.e until a choice is made. Almost ALL the occurences of UTIL_ShowMenu in the code use -1. Why are you passing a character ('\0') instead where an integer duration (number of seconds) is expected ?? |
Re: POD-bot back into shape.
To fix SoUlFaThEr's problem of bots hitting buttons randomly (actually the bots LOVE to hit any button they see), I think all is needed is disabling the "func_button" check in BotFindItem(). The bots should then only use buttons when they are declared in their waypoint through the button flag.
|
Re: POD-bot back into shape.
i dunno. I am an idiot:D
BTW I found in the bot.h file your definition of the teams: #define TEAM_CS_TERRORIST 0 #define TEAM_CS_COUNTER 1 is this possibly what you are talking about? BTW I never read the function for the UTIL_ShowMenu so thats probably why I just fucked and guessed that it might have got something to do with it:D MusicMan |
Re: POD-bot back into shape.
ok.....so whats working and whats not?
im REALLY confused now!!!!! hehe is that posted musicman download working or not? |
Re: POD-bot back into shape.
forget about my .dll for now:)
It does not do any good. MusicMan |
Re: POD-bot back into shape.
then do me a favor and test it thoroughly before you upload...
just uploading it is only going to stop all of us from waypointing if we get something that doesnt work at all:) believe me i am happy that someone will continue.....but continue with proper downloads......test them, get them approved and such.......this way.....when you do upload.....we have the feeling already that it will work because we know HOW you are working :) |
Re: POD-bot back into shape.
yeah, but it does work, and I tried to fix the bug you said you had about the crash, but failed.
And I will do alot of testing before uploading anything new, dont worry;) MusicMan |
Re: POD-bot back into shape.
Quote:
Quote:
Let me insist.... NEVER DO THAT! :| phew... |
Re: POD-bot back into shape.
Aye aye sir! ;)
my bad. It is from the podbot sdk, but its just me who have changed it to 0 and 1 in order to try and fix the bug SF was talking about. but where is the code where you -1 to get the original value, cause I assume I need to remove that when I have changed them to 0 and 1 from 1 and 2:) MusicMan |
Re: POD-bot back into shape.
Phew... no I'm confused ???:(
I appreciate your enthusiasm and your determination to continue this project, but it seems like you're off to quite a chaotic start. Before anybody starts "fixing" or changing anything, maybe we should FIRST determine the most apparent shortcomings, THEN discuss how to fix them and FINALLY touch the code. I'm totally lost now concerning this team definition thing. I'm using an older release because I need a dll that works reliably for my waypointing stuff. Have I missed some disastrous giant monster bug concerning team definitions, a bug that seriously affects gameplay? Has anybody complained that something's terribly wrong about using 1 and 2 instead of 0 and 1 ? I really cannot tell any more... I'm too stupid to even read that code, and I guess I wouldn't know a compiler if I had one on my HD, but I do believe that I know something about playing with bots and about what causes difficulties and stuff. So I might dropy by once and again and add my 20 cents to the discussion, if you can bear with me. 9_9 Right now, my 20 cents consist of a little list of things that might be changed, sorted by priority. 1. bots and buttons. 2. bots and buttons. 3. bots and buttons. . . . 487. bots switching to unwaypointed navigation when they'd better not do that... Ok, and I'm not even talking about code changes right now, but about some general questions: How do we want bots to behave? What part of the code is responsible for their current behaviour? Could it be modified to suit our needs? If so, how? Yes, I feel myself that this is liking taking 3 steps back, but when you start getting lost, going back to where you came from and getting a map is not the worst thing to do - so I suggest focusing first on what stefanhendriks said about his button code, and PMB's suggestion about that (currently sort of useless) button flag. Does that sound ok, or am I just talking nonsense again? :| |
Re: POD-bot back into shape.
my end wish for the button thing:
1. that the bots dont hit buttons at all unless a flagged waypoint is there. 2. that a button flag is no longer an automatic thing. it needs to be added to the menus in the waypoint editor so WE can place a button flag if theres REALLY a button a bot needs to hit... this is a huge problem solve! behavior is not an issue right now. only thing would be that bomb carriers have a default NORMAL behavior pattern, hostage holders have a default AGGRESSIVE behavior and a VIP has a default NORMAL behavior. I think this is possible to implement without any serious problems. but.....me= not a coder lets get the button thing going first and then discuss the next possible steps. ABOUT the organization. Splorygon, Musicman and Austin are the ones who seem interested in continueing. Austin is quite experienced. I personally think he should be(if hed like to) be the HEAD guy. maybe you 2 can send back and forth things you do to Austin for checking.....if he doesnt know.....THEN we ask the master PMB. when 1 thing is done right....it gets posted. :) Sound good? iam also thinking of stopping discussions in this thread......and starting a new one......this was PMB's thread.......and he is finished for the moment. plus this thing has over 50 pages now........im going over to start a new podbot2.6mm thread.....please use that one from now on ...... |
Re: POD-bot back into shape.
That's sounds like a great idea SF!
I wont even think about being the HEAD guy as I am way too inexperinced with Bot coding for that, so it would be nice if Austin would take that job if he wants to;) MusicMan |
Re: POD-bot back into shape.
Pierre-Marie - how long time did you lose to understand all lines of code of PODBOT and BOTS whole? About one year? So - what do You think - after 5 or even 20 your posts for help, the new PODBOT coder can continue Your great job? Sorry. I don't believe in it. I think this new man need about 2 or 3 months (maybe I'm wrong) for this. But You have much more experience - this is not only PODBOT coding but all knowledge about HL SDK and some other tools and HL software environment.
Austin and MusicMan - are you sure, that you can continue this PMB big job after short time of teaching You by PMB? You are propably the best coders after PMB for this code, but if I good understand this discussion above, there is a big distance beetwen you and PMB in this knowledge. I think ,the best way is ask the latest time to Pierre-Marie to finish this project PODBOT 2.6 mm by solving all knows bugs (issues) and only add this latest new function (button's flag in addflag menu - if I good understand you, SoUlFaThEr?). And this should definitively finish project PODBot 2.6mm. Pierre-Marie - How long time do you need to do this (including tests some releases by SoUlFaThEr)? About one week? So - can we ask You - the latest time - we promise - can You do this? Only this one week... Pierre-Marie - I think you will lose much more effort and time to explain many things to this new PODBOT coder than you finish this project (like described above by me) by yourself. Pierre- Marie - what do You think? I'know - you have enough of this project and you are tired by it, but I think You are a little bit of good finish of your great job. This is not good time to give up. There is only last round remaining... And other people - what do you think? Maybe if many of people will ask Pierre-Marie about this latest week of his effort, maybe he will be agree to finish this project? After releasing this bugs-free and SF new button's flag function - Austin or any other coder will continue adding or changing new functions, but not yet right now. It's not good time for this. If you are not agree with me, write this here and write me why. |
Re: POD-bot back into shape.
you more or less speak words directly from my thoughts.
but Pierre is not a machine :) He has his own bot (RACC) which he has been cheating on to do this pod stuff. he is also running a major part of the whole Bots-United project.....which is even more time consuming than coding the podbot because its the site, the filebase, the Wiki, and the "game" we plan to develope sooner or later. he also has a RL job...... lets not forget his life away from the keyboard all that taken into consideration.....there can be no complaining about his NEED to back down. i also think it would be neat for him to add that function...and then leave it.....cuz the rest is just small shit...and if those guys need help.....he can write something about it. |
Re: POD-bot back into shape.
i think if you understand general bot coding, you should be able to code for pod. You should know the right vars though. As i bet most coders already did what you think about now, you just have to find the right information to get on with it.
I think in a few weeks you will see MusicMan and Austin are coding this as if it was their bot from the start. |
Re: POD-bot back into shape.
well sometimes it is easier to starrt form scratch than undartainging somemones code ....
Of cousre "sometimes" is cruicial here gonnatest the latest podbot release this weekend |
Re: POD-bot back into shape.
I know. From my experience restarting is only good when you too many things are too messed up or should be done in a different way.
I think by changing the pod source by starting all over again, will actually make a total new bot. So in that case you should not call it a POD off-spring anymore. |
Re: POD-bot back into shape.
As there is a great interesst in podbot26mm maybe it deserves it's own forum?
|
Re: POD-bot back into shape.
not really because its a metamod plugin in essence :)
|
Re: POD-bot back into shape.
Some other bots are MetaMod plugins too - and they have their own forum ;)
|
Re: POD-bot back into shape.
Just downloaded podbot_sdk, and have few questions:
Code:
if ((pBot->vecGrenade == g_vecZero) || (BotGetSafeTask (pBot)->fTime < gpGlobals->time)Code:
(pEdict->v.weapons & (1 << CS_WEAPON_FLASHBANG) == 0Code:
if ((pBot->vecGrenade == g_vecZero) || (BotGetSafeTask (pBot)->fTime < gpGlobals->time) I have VS PRO 2003 (7.1). All settings left untouched :) |
Re: POD-bot back into shape.
1. The official metamod is cross-compiled by GCC on Linux for Windows, and perhaps it's compiled with Unicode strings instead of plain ASCII, or perhaps it contains debug symbols.
2. Even with the same settings, different compilers produce different sizes for the output... Mine was compiled with VS Pro 6. 3. The correct code is rather && (pEdict->v.weapons & (1 << CS_WEAPON_FLASHBANG) == 0)) You forgot the last parenthesises ; they are in the code. It's not what you say btw ; the priorities are first for the '&', and then for the '==' ; hence it's rather && ((pEdict->v.weapons & (1 << CS_WEAPON_FLASHBANG)) == 0)) or if you want && !(pEdict->v.weapons & (1 << CS_WEAPON_FLASHBANG))) |
Re: POD-bot back into shape.
Quote:
|
Re: POD-bot back into shape.
Quote:
I wish I had some time I would work on it tonight. |
Re: POD-bot back into shape.
yes all we need now is for you or someone experienced like you to add in the code to look for and push a button when reaching a waypoint with that flag. I already made them stop pushing buttons at random and stopped the waypoint editor from adding it automaticly and added it to the change flag menu.
|
Re: POD-bot back into shape.
@Terran: the original metamod Makefiles are still here, if that's what you want to know... but I somehow suspect that Will Day personally uses another versions of these Makefiles, because we tried cross-compiling metamod on the BU server but to no avail.
|
| All times are GMT +2. The time now is 13:32. |
Powered by vBulletin® Version 3.8.2
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.