#bitfighter IRC Log

Index Search ←Prev date Next date→

IRC Log for 2012-09-23

Timestamps are in GMT/BST.

00:02:15Watusimotoand I can't compike at all, though it is not related to your issue!
00:04:18raptorso sam686, does that mean i see you only if you're inside one of your spybugs?
00:04:58sam686right now, it looks like I can see any enemy when any enemy is near any spybug (not only my team spybug)
00:05:20raptorbut i can't see you unless you're near one of yours...
00:09:40raptorok recompiled
00:10:22sam686on isVisibleToPlayer, need to change mIsOwnedByLocalPlayer with something that was in SpyBug::packUpdate
00:12:46raptoris that server-side?
00:13:05sam686yes, i think.. Will check if server side only
00:13:29raptoroops
00:13:52sam686clientGame::renderCommander calls isVisibleToPlayer client side
00:14:19sam686but, gameType::performScopeQuery calls that too, server side
00:14:47sam686Probably need to have different arguments for server side, an argument for clientInfo
00:14:53raptorok, i initialized mIsOwnedByLocalPlayer in the constructor
00:15:27sam686It will still be wrong, unless you check of getOwner() == clientInfo server side
00:15:54raptorok, i think i'm not getting this...
00:16:06raptor(i'm also being distracted by family stuff)
00:16:19raptorif you know the fix already, feel free to do it...
00:16:27sam686I can try fixing myself..
00:51:27Watusimoto Quit (Ping timeout: 240 seconds)
00:53:41BFLogBot - Commit 25e278a1c693 | Author: sam8641 | Log: Fixed Spybug Scope problems.
01:04:21BlackBird_ has joined
01:04:26BlackBird_hi
01:04:35BlackBird_i think i might have a problem..
01:05:35BlackBird_the master server apparently is rejecting my username because i sent a duplicate player id
01:05:39BlackBird_help?
01:05:50raptorhi
01:06:26BlackBird_hi
01:06:33sam686hi
01:06:33BlackBird_do you know whats wrong?
01:06:36raptormaster restarted
01:06:40raptorand no, i know nothing
01:06:59BlackBird_is is in the process of restarting? or it just finished restarting?
01:07:07BlackBird_it*
01:07:09raptorwait 30 sec
01:08:02sam686restart your client if you get duplicated ID problem..
01:08:31raptori'm not really here
01:09:03BlackBird_its been at least 30 seconds, right?
01:09:19BlackBird_im getting the error again
01:09:28BlackBird_whaattss ggooiinngg oonn
01:09:36BlackBird_did someone steal my password or something??
01:09:59BlackBird_aaaand its crashing for the 3rd time
01:10:12BlackBird_*crashes*
01:11:04BlackBird_ok whats going oonnnnn
01:11:26sam686what happens when starting bitfighter?
01:11:46sam686is your bitfighter game running 017a or 017b ? (forgot which version is it)
01:12:23BlackBird_it should be the latest..
01:12:27BlackBird_imma reinstall
01:15:51sam686duplicate ID should be nothing to do with password / username (probably not, i think)
01:28:06raptor Quit (Ping timeout: 246 seconds)
01:52:41BlackBird_ Quit (Quit: Page closed)
02:04:23raptor has joined
02:04:24ChanServ sets mode +o raptor
02:42:31raptor Quit ()
02:51:08raptor has joined
02:51:08ChanServ sets mode +o raptor
04:31:12sam686 has left
04:39:52sam686 has joined
04:39:52ChanServ sets mode +v sam686
05:18:10raptor Quit ()
06:27:31sam686 Quit (Ping timeout: 245 seconds)
09:40:46LordDVG has joined
10:31:23Watusimoto has joined
12:04:47Watusimoto Quit (Ping timeout: 264 seconds)
13:14:38Watusimoto has joined
14:10:43LordDVG Quit (Remote host closed the connection)
15:00:05LordDVG has joined
15:01:08Watusimoto Quit (Ping timeout: 268 seconds)
15:08:27Watusimoto has joined
15:48:24sam686 has joined
15:48:24ChanServ sets mode +v sam686
16:00:30Watusimotohi sam686
16:00:44Watusimotodo you know anything about Vector<bool>?
16:06:44sam686the std::vector<bool> is different from std::vector<char> or any other types..
16:07:15sam686since Vector changed to std::vector, it might have some problems specific to Vector<bool>
16:11:50WatusimotoVector<bool> xxx
16:11:57Watusimotoxxx[0] = true;
16:12:01Watusimotodoesn't work
16:12:45Watusimotobut xxx.getStlVector[0] = true does
16:13:08Watusimotoand passing Vector<bool> via s2cxxxx doesn't work
16:13:15Watusimotosadly
16:28:15Watusimotoback in a bit
16:32:51Watusimoto Quit (Ping timeout: 246 seconds)
16:45:11raptor has joined
16:45:11ChanServ sets mode +o raptor
16:47:36raptorDONT USE VECTOR<BOOL>
16:47:43raptoroh wait
16:47:52raptori think TNL::Vector is OK
16:48:20sam686TNL::Vector aren't really ok sincer the change to std::vector
17:02:40raptorhi
17:02:42raptorok
17:06:08Watusimoto has joined
17:06:43raptorsam686: i thought you fixed Vector<bool> in 1897:78011beaf8ef and 1898:0bb5b32e7bf6
17:06:47raptor(long time ago)
17:07:31sam686partly, except can't get address of bool[]
17:08:44sam686I think i am coming up with a different hack to fixing it, by doing struct struct1{T v;}; and std::vector<struct1> innerVector;
17:09:05raptormaybe we can define BOOL int
17:09:20raptorand TRUE = 1; FALSE = 1
17:09:23raptorerr 0
17:10:14raptorwhat were the c++ guys smoking when they decided to implement vector<bool> the way they did..
17:11:45sam686by doing struct struct1{T v;};, i think I can static cast (struct1*) to (T*).. even if (T == bool)
17:12:37raptorwould Vector<bool> still be backed by std::vector?
17:13:04sam686yes, std::vector<struct1> innerVector;
17:13:31sam686because we are inserting our struct1 containing only T std::vector won't see a bool
17:14:07raptorahhh
17:14:19raptorwould that slow down Vector much?
17:15:19sam686I don't know, I think no since it is still only one..
17:15:29raptorwas vector<bool> fixed for c++11?
17:16:29sam686not sure..
17:20:11raptormaybe jsut use std::deque?
17:20:41raptoror can you somehow get the bit it represents on the fly, and return it as a bool?
17:21:30raptorcan you wrap just Vector<bool> in the struct? I think I'd rather leave all other types alone, without the extra wrapper
17:23:03sam686http://sam6.25u.com/upload/tnlVector.h there.. but not limited to just bool, but it appears to work..
17:23:36raptorcan it be limited to just bool?
17:24:15sam686It could, but I would have to copy the entire struct, i think...
17:25:47sam686more like, I would have to copy everything on tnlVector just for bool...
17:26:31raptoreven if you used the previous signatures for Vector<bool>?
17:26:55sam686can you check if there is performance problem with my changes, compared without my changes?
17:27:26sam686the problem is, there is about 15 functions I had to change to make it work with my "Struct1"
17:28:30raptori can't find my old Vector performance testing program... :(
17:30:22raptorwell, there is an extra method call for every function since it has to access the inner member
17:32:41sam686if the compiler optimizer is smart, it can inline, taking away an extra call to inner member (since tnlVector.h is visible to compiler)
17:33:00raptori just found you can define STD_VECTOR_BOOL_NOT_SPECIAL
17:33:11raptorto make it act normally
17:33:59raptorbut i bet that that isn't in the mac gcc for old ppc builds
17:34:55sam686problem with that define, it isn't on visual C++ 2008, and probably not on visual C++ 2010
17:35:01raptorrats
17:35:58raptorhere's another solution: http://stackoverflow.com/questions/8399417/why-vectorboolreference-doesnt-return-reference-to-bool
17:36:42sam686looks like my debug build skips struct1 constructor, and goes straight to "T" constructor when T == string..
17:37:16sam686and since struct1.v tells the compiler it is just a offset of 0, it gets optimized out..
17:38:18sam686problem, is sizeof(char) == sizeof(bool)?
17:38:25Watusimotois it worth the hassle? we could just use the crackhead STL version, no?
17:38:32Watusimotoand take the memory hit
17:38:47raptoryou can't get a reference to the bool with the stl version
17:38:55raptoroh, and hi Watusimoto
17:39:08sam686std::vector<bool> have a speed hit beacuse of packing 8 bit into one byte..
17:39:17Watusimotoonly here for a minute... thanking the gods we don't use 3-d models in bitfighter
17:39:37Watusimotoah, thought it was representing one bit as a byte
17:39:51raptorno, that's what we want
17:40:05raptorbecause bool is 1 byte everywhere but in std::vector
17:40:06sam686i don't think all compilers have sizeof(bool) == sizeof(char)
17:49:59Watusimotoback in a few minutes
17:50:05sam686Another possibility is to maybe hack std vector.h and add in #ifndef STD_VECTOR_BOOL_NOT_SPECIAL
17:51:30sam686but then there will be many complaining it won't compile because of missing #ifndef...
17:54:51raptori'm re-writing my testing class...
17:55:58raptoroh wait! i found it!
17:56:41raptorlet me get your class set into it..
17:58:10sam686can you send your test source code? that will make sure I have it if you, in a future, lost it again...
17:58:15raptorok
17:58:24raptorone moment, i have to clean it up a but
17:58:26raptor*bit
18:00:42sam686will be back in about 10 minutes..
18:10:56raptorwhoa, you're domain is gone
18:11:26sam686I had zoomber "CNAME" that to http://sam6.25u.com/
18:11:41sam686but zoomber seem to not be coming back anytime soon to see that problem
18:12:04raptorhere is my eclipse project for testing: http://sam6.25u.com//upload/testing.tar.gz
18:12:09raptorhttp://sam6.25u.com/upload/testing.tar.gz
18:13:20raptori put your new code in namespace of SAM::
18:14:11sam686My 7-zip can see username and group names inside a ".tar", i see it as dbuck : users
18:15:42raptoryep, tar preserves permissions
18:23:18raptorlooks like my timing code might be broken
18:30:58sam686there is one problem, you are measuring the time of cout or outputtint to console..
18:31:35raptorah yes
18:31:37raptorgood point
18:32:27raptori wrote this along time ago...
18:35:19raptori shoudl rewrite...
18:43:36raptoryuk, lots of bad code - i'm rewriting anyways
18:57:47sam686'type' : symbol cannot be used in a using-declaration when trying "struct struct1{using type = T;};" when you say to try http://stackoverflow.com/questions/8399417/why-vectorboolreference-doesnt-return-reference-to-bool
19:09:00Watusimotohi again
19:10:27raptorhi
19:14:23Watusimotoso what should I do about my vector<bool?
19:14:38raptorwell...
19:14:46WatusimotoI have things kind of working internally, by using the stlcontainer directly
19:14:48raptoruse deque?
19:14:51Watusimotowhich is not really what you want
19:15:14Watusimotobut I want to pass a vector<bool> through a s2c function call
19:15:15raptorif we want it to work with TNL::Vector, sam wrote a solution to wrap any T in a struct first
19:15:45WatusimotoI'm thinking I need to write a custom packer to pass the bits
19:15:47raptorbut it seems to me we should specialize Vector<bool> and wrap it just in the specialization
19:16:03Watusimotoor perhaps pass a number then a number of flags
19:16:06Watusimotoor something
19:16:11sam686it appears to be a little slower...
19:16:20raptoryeah, slightly slower
19:17:00sam686but I might be coming up with maybe another solution to that problem which hopefully won't be so slow, i hope..
19:17:43raptorsam686: can we keep your previous template<> Vector<bool> specialization and somehow wrap just wrap the bool for that case?
19:17:59sam686I think so..
19:23:44raptorWatusimoto: maybe just typedef BOOL int? :)
19:24:10Watusimotowhy not just use the inefficient stl implemenation?
19:24:14Watusimotois it really that bad
19:24:16Watusimoto?
19:24:43raptoryou cannot get back a bool by doing bool something = vec[0];
19:25:53WatusimotoAs an STL container, there are really only two things wrong with vector< bool >. First, it's not an STL container. Second, it doesn't hold bools. Other than that, there's not much to object to.
19:26:02raptorhaha
19:26:05raptorso true
19:26:19Watusimotoyou can't??
19:26:30raptornope
19:28:07Watusimotocan we specialize Vector<bool> to use deque (transparently) instead?
19:28:22raptorsam686 is working on the specialization right now
19:28:35raptornot sure what he has in mind...
19:31:14raptori wonder if it can be hijacked as a char somehow
19:32:59Watusimotohttp://stackoverflow.com/questions/670308/alternative-to-vectorbool
19:34:04sam686can you get a array of pointers from a deque? (i think no)
19:34:09raptorno
19:35:14WatusimotoA vector< char > or similar will use less space, though it also has the potential to take a (very small) speed hit in some circumstances, because characters are less than the machine word size. This is, I believe, the main reason that bools are implemented using ints instead of chars
19:36:01Watusimotoanother solution:
19:36:01WatusimotoIf you really want clean semantics, I also like the suggestion of making your own boolean class -- looks like a bool, acts like a bool, but fools the template specialization.
19:36:21Watusimotosimilar to your previous suggestion?
19:39:24raptoryes
19:41:44Watusimotothat's what I'm doing for the moment, actually
19:42:28sam686i have an idea: std::vector<char[sizeof(bool)]> innerVector; or similar...
19:43:34raptormaybe we should just remove the bool specialization and use a typedef
19:44:27WatusimotoI'm porting my code over to use S32s instead of bools; since we can;t pass a vector of bools through a s2c, that takes away a lot of my desire to use Vector<bool>
19:44:51Watusimotoluckily, in my case, it works pretty well to pack all my data into a U16
19:45:04raptoryou mean using it as a mask?
19:45:08Watusimotoyup
19:45:15raptorsounds good to me
19:45:30Watusimoto99% of the time, I could get away with a U8, but it's one bit per team
19:46:26Watusimotobut a U16 is not bad -- if passing a vec<bool>, we'd need to use 8 bits for the count, and then probably 2-4 more for the data (depending on team count), and so U16 isn't much worse
19:46:43Watusimotoand in the worst case (10 teams), it;s better
19:46:48Watusimotobut that case is very rare
19:47:35raptorbetter be safe than sorry, as the saying goes
19:56:44Watusimotoyes. so I'd say forget about vector<bool> and do something else.
19:57:11raptori wonder if we could still compile by ripping out that specialization..
19:59:19raptorsam686: thanks for fixing that spybug problem..
20:00:24Watusimotomaybe we should TNLAssert it to remind us not to use it
20:00:42raptorand the answer is YES!
20:00:45Watusimotobecause I lost a lot of time to forgetting about that issue
20:01:04raptorgood idea
20:01:19Watusimotoand assert saying ("Don;t do this, C++ vector<bool> is broke; see xxx for why") would be helpful
20:01:34raptorwhere should that go? in each of the 4 constructors?
20:01:41Watusimotonot sure where it would go...
20:01:46Watusimotoah, yes!
20:02:07raptorok, i'll do it - sam686, it looks like we're choosing to never use Vector<bool>
20:02:55Watusimotook, that would be wonderful
20:03:01Watusimotosam686: do you concur?
20:04:58sam686well, i sort of goit it on the right direction, but maybe missing operators..
20:05:38sam686http://sam6.25u.com/upload/text1209/120923_15-09-51.txt is what I have so far
20:06:02sam686doesn't quite compile yet though..
20:06:33sam686oh, and you might need to change namespace SAM back to TNL..
20:15:22BFLogBot - Commit 2d4ed5f57560 | Author: watusim...@bitfighter.org | Log: Comment
20:15:23BFLogBot - Commit 54f2a3eb534f | Author: watusim...@bitfighter.org | Log: Whole lotta nothing
20:15:25BFLogBot - Commit 70c86a004775 | Author: watusim...@bitfighter.org | Log: Comment
20:15:26BFLogBot - Commit 99023a617a6a | Author: watusim...@bitfighter.org | Log: In ZoneControl games, can always see which team has the flag, regardless of whether carrier is in or out of scope. This was a ridiculuous amount of work, and probably is more complex than it needs to be.
20:15:28BFLogBot - Commit c9977a6f6d5a | Author: watusim...@bitfighter.org | Log: Merge
20:22:17Little_Apple has joined
20:22:23Little_Applehelloo
20:31:59raptorWatusimoto: where did you get your quote about vector<bool>?
20:32:08raptori'm looking for a good link to provide the assert with
20:32:16raptorhi Little_Apple
20:32:40Watusimotounless sam fixed it?
20:33:02raptori thought i was gutting the bool stuff and throwing in asserts...
20:34:25Watusimotolink to this
20:34:26Watusimotohttp://www.youtube.com/watch?v=ibEdgQJEdTA&feature=related
20:35:04koda has joined
20:35:35raptorwhat
20:37:07raptori like the canteloupe
20:38:59Watusimotoif sam has made Vector<bool> work, we don;t need to assert. If not, we can link to this http://stackoverflow.com/questions/670308/alternative-to-vectorbool
20:42:41raptoryou stopped, right sam686?
20:44:52Little_Appleraptor:
20:45:02Little_Applewould you know if/when _k would be back?
20:45:04sam686almost have it done..
20:46:27raptorwell sam686 if you don't want to finish, i can just put this in there: http://pastie.org/4787399
20:46:51raptorLittle_Apple: I don't think any of us know anything about _k
20:47:19Little_Appleoh ok.
20:49:01sam686this appears to compile, http://sam6.25u.com/upload/text1209/120923_15-09-28.txt but will run it and see...
20:50:00Little_Appleim going to look for some things that need to be edited on the wiki. hopefully i wont mess things up too badly
20:50:58raptorLittle_Apple: you guys waiting patiently for 018?
20:52:57Little_Appleyup d:
20:52:57Little_Appleraptor: what was that game you were talking about that was similar to bitfighter that you had to compile yourself?
20:53:26raptorxpilot
20:53:42Little_Appleoh ok. i knew it had an x in it...
20:53:44Little_Applethanks
21:13:48Little_Appleraaaaaptor
21:14:02Little_Applei need to bug you some
21:14:06raptorhi
21:14:42Little_Applecould you help me with compiling?
21:14:58raptornot today, sorry
21:15:10Little_Appleaw
21:16:25raptormy brain is a bit fried..
21:17:03Little_Appleokokok one quick question
21:17:09raptorok
21:17:20Little_Applei got mercurial and got the source using the command on the wiki
21:17:27Little_Applebut where the heck is it?
21:17:29raptorgreat!
21:17:36Watusimotowow
21:17:39Little_Applei cant seem to find it .___.
21:17:53raptorit's in the directory in which you ran the command
21:17:59Watusimotowhat he said
21:18:02Little_Appleok...
21:18:34Little_Applelol fail
21:18:47Little_Applei had it set for a minecraft server...
21:18:54Little_Applebut i found it. thanks.
21:19:26raptorsure :)
21:19:51Little_Applei need to change that on terminal..
21:20:11Little_Appleaaaaand its giving me this.
21:20:12Little_AppleThe scheme 'Bitfighter' contains no buildables that can be built for the SDKs supported by the run destination My Mac 64-bit. Make sure your targets all specify SDKs that are supported by this version of Xcode.
21:20:29raptorso that means
21:20:36raptoryou're using a newer Xcode?
21:20:53Little_Appleyep. the older version wont work on mountain lion
21:21:00Little_Applemountain lion messed up a ton of stuff
21:21:19raptorhmmm
21:21:21LordDVG Quit (Remote host closed the connection)
21:21:59raptorwell, there are two solutions: 1. go back to snow leapard 2. wait for someone to create an xcode4 project
21:22:09raptor#1 probably isn't going to happen
21:22:13Little_Appleyea...
21:22:26Little_Appleespescially since i share this computer… which i hate
21:22:48Little_Appleif it was mine i would go back to lion in a heartbeat
21:23:27raptor#2 I am not eager to do because I am not so fond of Mac and I already have a snow leopard solution that works :)
21:24:45Little_Applelolz
21:24:57Little_Applei could try going back to lion...
21:25:07Little_Applethings worked with that...
21:25:23raptori don't think i would
21:25:35Little_Applehmm
21:25:49raptoris there no way to 'import' an old project?
21:26:00Little_Applei dont think so
21:26:05raptoror, you can try to creat your own :)
21:26:15Little_Applehow?
21:26:40Little_Appleid just like to get it working
21:26:57raptoruhh, i don't know.. . 'create new project'? etc...
21:27:23Little_Appleand another reason i hate sharing a computer, i am being kicked off. be back later. thanks for the help
21:27:34Little_Apple Quit (Quit: Page closed)
21:34:31raptorwell, i'v edone nothing bitfighter-y all day...
21:44:38raptor Quit ()
22:35:46BFLogBot - Commit 3853d5271131 | Author: sam8641 | Log: Changed Vector to no longer pack 8 bool into one byte, enable getting array address of it, and is faster.
22:37:11sam686hi Watusimoto, I just got Vector<bool> working much better (internally does std::vector<char>), and look at the faster (lower is faster) speed: http://sam6.25u.com/upload/text1209/120923_17-09-06.txt
22:43:07Watusimotogreat!
22:45:54raptor has joined
22:45:54ChanServ sets mode +o raptor
22:46:04raptoroooo sam, you got it to work?
22:47:10raptorsam686: i'm getting error:
22:47:30raptortnlVector.h:133:4: error: ‘innerVector’ was not declared in this scope
22:50:50BFLogBot - Commit 21c865636e31 | Author: watusim...@bitfighter.org | Log: Update CTF and Rabbit to use new team flag tracking system
22:50:51BFLogBot - Commit 3c5c33b77552 | Author: watusim...@bitfighter.org | Log: Merge
22:50:59Watusimotosleeeeeeeepy
22:51:07raptorsleeeeeeeepy
22:51:09Watusimotogood night
22:51:11raptornight
23:09:12raptorsam686: still around?
23:09:17sam686hi
23:09:28raptori thin i've solved the error: http://stackoverflow.com/questions/7076169/not-declared-in-this-scope-error-with-templates-and-inheritance
23:09:43raptorbasicall make all internal references of innerVector be this->innerVector
23:10:06raptorbut i'm curious - how come you added (T*) to a lot of the returns?
23:10:34sam686its the type casting from bool to char..
23:10:42raptorah ok
23:10:46sam686can you just do #define innnerVector this->innerVector
23:10:46raptorso i'll leave it alone
23:10:58raptorthat's a good idea, le tme try..
23:11:30Watusimoto Quit (Ping timeout: 246 seconds)
23:15:07raptoralmost works... fails with the p.innerVector
23:15:45sam686so define does p.this->innerVector? wrong
23:15:57raptoryep
23:16:31raptorwas there a reason why you commented out getStlVector()?
23:16:43sam686it won't work on bool..
23:17:40raptorah ok, well it's not used anyways..
23:18:45raptori'll just add 'this' everywhere..
23:22:05raptorwell, it compiles...
23:22:11raptorlet's see if it runs...
23:30:44raptori got Assert: index out of range in ../tnl/tnlVector.h line 366
23:30:57sam686got upload_.level file now? (should be zone control)
23:31:36raptorhttp://pastie.org/4788202
23:31:45raptorcan't change to it...
23:32:08sam686I am getting the same problem too
23:32:18raptorok, so not a Vector problem..
23:33:32sam686teamIndex = -1?
23:33:41raptori guess?
23:34:01sam686neutral zones have team -1
23:35:58BFLogBot - Commit 93be5db4f03f | Author: buckyballreaction | Log: Fix compiling in Linux with new Vector code
23:38:55raptori have to go in 5 min..
23:39:04raptor be back after 3 hours or so..
23:39:15sam686ok..
23:39:47sam686I will eventually have a zone control renderInterfaceOverlay fix
23:43:08raptorok
23:43:11raptorbye
23:43:13raptor Quit ()
23:58:58koda Quit (Quit: you can't say 'hello' without saying 'hell')

Index Search ←Prev date Next date→

These logs were automatically created by BFLogBot on irc.freenode.net.