bitfighter custom server

[ZoombeR]

hey, i was trying to compile a bit-fighter game on my Ubuntu computer, to build a custom server later on, and wanted to start out with making a basic build of bit-fighter.
so i got the svn, did what needed to be done beforehand

and got to cd bitfighter

in Bitfighter, when i typed in Make, i got this

max@ubuntux8:~/bitfighter$ make
make[1]: Entering directory `/home/max/bitfighter/tnl'
g++ -g -DTNL_ENABLE_LOGGING -I../libtomcrypt/src/headers -I../libtomcrypt -c assert.cpp
assert.cpp: In function ‘const char* TNL::avar(const char*, ...)’:
assert.cpp:81: error: ‘vsnprintf’ was not declared in this scope
make[1]: *** [assert.o] Error 1
make[1]: Leaving directory `/home/max/bitfighter/tnl'
make: *** [default] Error 2

do I have to modify any files before I make it? or can i go directly to building the game right after i get the source code? If I don't have to mod any files, why do I get this error?

Sorry if i'm confusing or am just a plain newbie at this. This is my first time compiling something

[watusimoto]

Hmmm.... that's not a 64-bit Ubuntu computer, is it?

[karamazovapy]

Just for fun, I logged on to my VPS, which is running 32-bit CentOS. Obviously, it already has all the preferred updates and packages.

Navigated to root.

svn co https://zap.svn.sourceforge.net/svnroot ... lease-012a bitfighter2

cd bitfighter 2

make dedicated LUALIB=/usr/lib/liblua.a

Fail too long to copy/paste from the java terminal.

Ran again to eliminate the working parts from output.

Here are the second results:
http://bitfighter.pastebin.com/XUtTjxn3

[ZoombeR]

Hmmm.... that's not a 64-bit Ubuntu computer, is it?

heh heh, i guessing thats why im getting the problem?

Just for fun, I logged on to my VPS, which is running 32-bit CentOS. Obviously, it already has all the preferred updates and packages.

Navigated to root.

svn co https://zap.svn.sourceforge.net/svnroot ... lease-012a bitfighter2

cd bitfighter 2

make dedicated LUALIB=/usr/lib/liblua.a

Fail too long to copy/paste from the java terminal.

Ran again to eliminate the working parts from output.

Here are the second results:
http://bitfighter.pastebin.com/XUtTjxn3

So, in simple words, what is the reason it cant build? It looks like the endings of our errors are similar.

[karamazovapy]

64-bit linux builds require a couple little patches to the code for them to work properly. It may be easier (if you have a VPS, anyway) to just go with a 32-bit build instead.

You're crapping out in the middle of the Torque Network Library stuff which I think(?) uses some specific references to 32-bit...well...stuff. Like I said, a couple people have provided 64-bit patches, but I've honestly lost track of them. I'd also have to do some googling to figure out how to properly apply one, although you could probably just paste over the relevant bits of code.

Technically, what I did *shouldn't* have resulted in any errors. I also don't know enough about debugging to figure it out for myself without a significant amount of work.

[watusimoto]

I think it's actually simpler than that. I thought I had it working for both 32 and 64 bit linuxes, but maybe not, or maybe only in the 013 code or maybe...

I need to look into this again now that I have a 64 bit machine.

[ZoombeR]

64-bit linux builds require a couple little patches to the code for them to work properly. It may be easier (if you have a VPS, anyway) to just go with a 32-bit build instead.

You're crapping out in the middle of the Torque Network Library stuff which I think(?) uses some specific references to 32-bit...well...stuff. Like I said, a couple people have provided 64-bit patches, but I've honestly lost track of them. I'd also have to do some googling to figure out how to properly apply one, although you could probably just paste over the relevant bits of code.

Technically, what I did *shouldn't* have resulted in any errors. I also don't know enough about debugging to figure it out for myself without a significant amount of work.

yes, my vps will have 32-bit. I was just tying it out on a copy of ubuntu on my macOSX, which is 64 bit.
I guess this will mean hopefully If I do what ive done before, it *should work then?

Also, I remembered getting an error when trying to install Freeglut, stating, "package not found"
i was able to install the other packages after it though.

[watusimoto]

You don't need freeglut in order to compile a dedicated server -- just build with the dedicated option. See the makefile or the wiki for details.

Note that this may be completely unrelated to the other problem you reported.

[ZoombeR]

i see. I guess I will have to try directly with my vps then.
good thing i didnt purchase the 64 bit though
im guessing thats the problem
as I dont know what else it could be

[watusimoto]

Hey ZoombeR, would you be willing to try building again? This time, though, try building from this version:

https://zap.svn.sourceforge.net/svnroot ... -013-beta8

Thanks!

[ZoombeR]

same thing. the error is in the file assert.cpp, and somewhere at the bottom, this function, it gives me the error "vsnprintf is not declared"

  Code:

const char* avar(const char *message, ...)

{

   static char buffer[4096];

   va_list args;

   va_start(args, message);

   vsnprintf(buffer, sizeof(buffer), message, args);

   return( buffer );

}
 

Full terminal message below:

  Code:

make[1]: Entering directory `/home/max/bitfighter/tnl'
g++ -g  -DTNL_ENABLE_LOGGING -I../libtomcrypt/src/headers -I../libtomcrypt    -c  assert.cpp
assert.cpp: In function ‘const char* TNL::avar(const char*, ...)’:
assert.cpp:81: error: ‘vsnprintf’ was not declared in this scope
make[1]: *** [assert.o] Error 1
make[1]: Leaving directory `/home/max/bitfighter/tnl'
make: *** [default] Error 2
 

UPDATE 1: named "vsnprintf" something else, and got the same error, only saying that the new name wasn't declared. Does this mean there needs to be a different word here to replace "vsnprintf"?

[ZoombeR]

UPDATE 2: I left out that whole line. when i went to make Bitfighter, the file assert.cpp was made successfully, only to leave me with more fail.

  Code:

make[2]: Entering directory `/home/max/b8beta/lua/lua-vec/src'
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lapi.o lapi.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lcode.o lcode.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o ldebug.o ldebug.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o ldo.o ldo.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o ldump.o ldump.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lfunc.o lfunc.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lgc.o lgc.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o llex.o llex.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lmem.o lmem.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lobject.o lobject.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lopcodes.o lopcodes.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lparser.o lparser.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lstate.o lstate.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lstring.o lstring.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o ltable.o ltable.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o ltm.o ltm.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lundump.o lundump.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lvm.o lvm.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lzio.o lzio.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lauxlib.o lauxlib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lbaselib.o lbaselib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o ldblib.o ldblib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o liolib.o liolib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lmathlib.o lmathlib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o loslib.o loslib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o ltablib.o ltablib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lstrlib.o lstrlib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o loadlib.o loadlib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o linit.o linit.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lveclib.o lveclib.c
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lgcveclib.o lgcveclib.c
ar rcu liblua.a lapi.o lcode.o ldebug.o ldo.o ldump.o lfunc.o lgc.o llex.o lmem.o lobject.o lopcodes.o lparser.o lstate.o lstring.o ltable.o ltm.o lundump.o lvm.o lzio.o lauxlib.o lbaselib.o ldblib.o liolib.o lmathlib.o loslib.o ltablib.o lstrlib.o loadlib.o linit.o lveclib.o lgcveclib.o
ranlib liblua.a
gcc -O2 -Wall -DLUA_USE_LINUX   -c -o lua.o lua.c
In file included from lua.h:17,
                 from lua.c:15:
luaconf.h:275:31: error: readline/readline.h: No such file or directory
luaconf.h:276:30: error: readline/history.h: No such file or directory
lua.c: In function ‘pushline’:
lua.c:182: warning: implicit declaration of function ‘readline’
lua.c:182: warning: assignment makes pointer from integer without a cast
lua.c: In function ‘loadline’:
lua.c:210: warning: implicit declaration of function ‘add_history’
make[2]: *** [lua.o] Error 1
make[2]: Leaving directory `/home/max/b8beta/lua/lua-vec/src'
make[1]: *** [linux] Error 2
make[1]: Leaving directory `/home/max/b8beta/lua/lua-vec/src'
make: *** [default] Error 2

 

so, are we missing a file here?

[watusimoto]

You may need to install the readline-dev package.

Removing the line may get you through the compile, but it isn't really the fix. The real problem is has to do with va_start and friends, which behave differently on a 64-bit platform for reasons I can't recall at the moment.

[watusimoto]

I've compiled the trunk under 64-bit Linux. At least I have with the dedicated server.

If you want to try again, you'll want to grab the code from the trunk revision of the repository, and you'll want to install the readline module...

apt-get install libreadline6-dev

Please let me know if this works!

[ZoombeR]

Cool, I'll get to it as soon as I can. Unfortonatly I am away from the Internet most of the time right now so it may be a few before I can reply. If I can install that package, i'll try to do the compile offline.

On a side note, I'm also compiling it with mac too. I'll update the status with that too. : D

z

[watusimoto]

If you get the Mac version to compile, that would be really great. We really need some OS X testing from someone who can compile. Please post your progress here, and I'll let you know how you can help.

[ZoombeR]

Mac:
been keeping a log of fixes on mac, which I will post later when I get back from my vacation.

The first time i opened the SVN, before you posted the new beta link, i found a wrecked train in the mixture. But apparently, over 3,000 build errors are fixable with a simple change in filename, which I have included in my log file. Anyways, for the older svn, have managed to trace the remaining 30-40 errors or so, to a couple of missing files and some more "undeclared scopes", also in my Log. I have yet to compile the newest downloaded SVN.

First update: Using the newest beta-8 SVN, I also got 35 build errors, meaning the fixes i made in the previous SVN, were already fixed in the new download. The build errors were the same ones.

Second update: are there missing files? I don't see a folder called test, and only half of the files supposed to be in the test folder are in the osx folder. Missing files include: Server.cpp, TestGame.cpp, Testgame.h, TestGameRender.cpp

There is also no masterclient folder, and no "main.cpp" file, but I'm guessing that is irrelevant right now.

As for errors and warnings, the files affecting the compile are "testGameRender.cpp", "testGame.cpp", and "MyOpenGlView.mm"

testgamerender and testgame, are the files not found.

MyOpenGlView.mm is there, but it is the file that gives the " 'X' is not declared in the scope" errors, and the expected type-"x" before Y

Update 3: after browsing through the errors, I'm getting a feeling if i can locate testgame.h, most or all of these problems will be fixed

I took a screenshot as well, Ill try to get that up too in a second

[ZoombeR]

Ubuntu:

Installed readline package....
Compiling.....
Error: same

Using Svn Trunk package...
Compiling.....
Error


It feels like Im not doing something right, but the fails are in different places depending on the version im using

[watusimoto]

That testgame stuff is cruft left over from the old zap package, and is unused. You can delete all references to it and be good.

How are you compiling on the Mac? I believe it is packaged to compile with xcode, and all that test stuff should be gone.

[ZoombeR]

I use Xcode 3.1.3

The original project file was compiled using an older version of xcode, and therefore, xcode converts a copy of it and saves it as tnl.xcodeproj.

I don't think there is any issue using it with a newer version.

Going on, I believe the main problem exists with MyOpenGlView.

Deleting the Test references, shortened the compile errors down to 33, and the only way for me to get it to compile, is if I delete MyOpenGlView.mm, which of course, kills my screen.

also, is server.h another one of the files no longer needed?

furthermore, am I able to just delete the whole test folder reference?
^edit: no i cannot cannot^

[watusimoto]

I don't see server.h in the copy of Bitfighter I have checked out, so it's probably safe to delete.

Are you sure you aren't looking at an ancient copy of the code?

[ZoombeR]

this one:
https://zap.svn.sourceforge.net/svnroot ... 013-beta8/

the errors Im getting are "" was not declared in this scope

(being a word inside "")

all the while i just get more and more confused

[watusimoto]

Post the whole error, and let me take a look. Though that was the same version that Ryan compiled for the release, so it absolutely should work!

[watusimoto]

I finally managed a OS X build tonight, on my friend's machine. While I had to make a couple of minor modifications to the project to get it to compile (which I checked into the trunk version of the repository), I did not encounter any of the sorts of problems you are reporting.

I would suggest doing a clean checkout of the trunk code, then loading the main project xcode file. I would expect you to be able to compile right off.

[ZoombeR]

GAHHh so close yet so far.

Managed to clear out spelling error (just assuming there) and lower it to 33 errors, all in one file, MyOpenGlView.mm

Also realized I must use the 10.4 sdk for it to even get that close, and Would me having a newer version of xcode affect it in any way?
Mines 3.1.3, when I open the project i must save a copy of it as .xcodeproj instead of .xcode

update

32 and more thankful all the way

[ZoombeR]

Ok, I thought most of the errors were testgame related stuff, but I could not be sure. Just for a shot, I tried removing the text and sized "myopenglview.mm" to this:
(//&& will represent where i get the error)

  Code:

#include <OpenGL/gl.h>

#import "MyOpenGLView.h"

- (id) initWithFrame: (NSRect) frame
{
//&& error:expected unqualified-id before "-" token
    GLuint attribs[] =
    {
        NSOpenGLPFANoRecovery,
        NSOpenGLPFAWindow,
        NSOpenGLPFAAccelerated,
        NSOpenGLPFADoubleBuffer,
        NSOpenGLPFAColorSize, 24,
        NSOpenGLPFAAlphaSize, 8,
        NSOpenGLPFADepthSize, 24,
        NSOpenGLPFAStencilSize, 8,
        NSOpenGLPFAAccumSize, 0,
        0
    };
   [NSTimer scheduledTimerWithTimeInterval: 0.02 target: self selector:@selector(tick) userInfo:NULL repeats: true];
   
    NSOpenGLPixelFormat* fmt = [[NSOpenGLPixelFormat alloc] initWithAttributes: (NSOpenGLPixelFormatAttribute*) attribs];
   
    if (!fmt)
        NSLog(@"No OpenGL pixel format");

   theLog = new TestGameLog(self);
   [self restartAsClient:self];  
    return self = [super initWithFrame:frame pixelFormat: [fmt autorelease]];
}

@end

Strange, Even with a clean revision of the trunk version, i still get 34 errors or so.

Update 2: I really do not know what I'm doing at this point, now, ive somehow gotten it to compile with the file included, and dont see the window.

[watusimoto]

OK, I'm really confused.

myopenglview.mm looks to be totally related to test code, and is not necessary at all, as far as I can see.

Of course, I really don't know what this file is for, and I've never even seen it before. Does it get loaded by the xcode project?

Also, I don't see your error marker!

[ZoombeR]

Ok, here what it was. MyOpenGlView.mm, MyOpenGlView.h, and Main.mm were all in the "test" folder, meaning I missed that and should delete it all, if I'm right?

I did this before, and it successfully compiled, but when done, I get the blank white tnl-test window. Is this supposed to show up?

CHANGE: maybe I was doing this all wrong? It appears the OSX folder hasnt been modified for a while, but just now, I checked the Zap! folder to find an .xcodeproj file, up to date with my computer, ready to compile, without the test files. For the test files, you were right; they are all gone, Im not sure why there still is an osx .xcode file, as it seems more outdated.

Now, when I went to build it, I was finally almost able to compile the game, except this time, the errors I get are saying I need:

libtomcrypt framework,

And it cant find these paths:

/trunk/zap/../libtomcrypt/build/release
is not found

/trunk/zap/../tnl/build/release
is not found

But overall, im guessing the project was just moved?

[watusimoto]

Not sure what happened, but it seems you were looking at an old project file somehow.

No matter, the situation has vastly improved. TNL and libtomcrypt are two libraries that are included with Bitfighter (for networking and encryption, respectively). The source is all there, you just need to get the project to see those somehow.

When was the last time you updated your code? About a week ago, I made an adjustment to the project to get two libs to link in properly (don't remember if it was these). The problem was the paths were absolute paths, wired to the OS X maintainer's machine. I made them relative, and it compiled like a charm. I checked that fix in, so if you don't have it, refresh and things may just work.

Otherwise, it's probably a path setting somewhere.

[ZoombeR]

Fixed that, and again, you were right, it must have been a path problem.
I couldn't find the actual .framework file so i just recompiled them again and added them in myself.

Unfortunately, yet again I got another error
Now, 20 sfx errors have appeared during the linking phase of the compile

  Code:

"_aiSourcef", refrenced from:
Zap::SFXObject::updateGain()    in sfx.0
Zap::SFXObject::updateGain()    in sfx.0
Zap::SFXObject::PlayonSource()    in sfx.0
Zap::SFXObject::Playonsource()    in sfx.0
Zap::SFXObject::Playonsource()    in sfx.0
 

and more like this

[watusimoto]

Odd... you said you added some stuff yourself because you couldn't find the framework. Was sound one of those things?

There should be an ALUT or OpenAL framework or something like that, and it looks like maybe that's missing?

Also, I saw that Ryan checked in some Mac project stuff last night. You might try checking out the trunk into a new folder and seeing if things build better with the default settings.

Sorry this has been so painful, and that I'm not able to provide better assistance!

[ZoombeR]

Its strange. I downloaded the next revision, and when the frameworks still needed to be built. (must not come precompiled for mac or something? maybe thats why theres .xcodeproj files in tnl and libtomcrypt?

The strange thing was when I built libtomcrypt.framework, xcode didnt recognize it in tnl.xcodeproj (building the tnl framework here)
and the main project file, bitfighter.xcodeproj.

Whats even more bazzar, is when I clicked get info, the path to the framework was correct.

EDIT: I looked more carefully and found that it said /Users/ryan at the beginning, meaning the system couldn't find it, because it was searching the wrong file path completely. Aparently the relative to project part didn't work here?

EDIT 2: for some reason, tnl.framework only compiles if i manually add the framework myself. (not sure why. I chose the path to be default, and when i clicked info, it defaulted to "relative to path". Again, Strange, but at least theres a solution

(on a side note, i got a warning building it stating there was an unused variable, "val" if i need to fix that)

Now onto the main project file, libtomcrypt was found on its own, but tnl, again, wasnt found BUT the path was correct.....and this one was relative to the project, so the full path was correct too.

again, re-adding it seemed to work.

EDIT: Libtomcrypt also had to be re-added...even though it was found before, when building it, it wasnt found. Ssssstrange.

Now im back to my 20 errors again. I tried doing the same steps for ALUT.framework and nothing. It was always there.

FIX!!!!!: I deleted the OpenAL reference, and copied the framework into the openal folder, making it a relative to project path.
For some strange unknown reason, the build succeeded

[ZoombeR]

Time for application launch, and it crashes, status bar stating Bitfighter has exited with status 5.



GDB: Interrupted

EDIT:

oh right

Dyld Error Message:
Library not loaded: @executable_path/../Frameworks/tnl.framework/Versions/A/tnl
Referenced from: /Users/hushahn/bitfighter2/zap/build/Debug/Bitfighter.app/Contents/MacOS/Bitfighter
Reason: image not found

where? did?? it go?

its a full intel architecture'd image, and when i checked inside, the MacOS file WAS there

[ZoombeR]

WHOOOOOOHOOOOOOOOO I FIXED IT UP. IT ALL WORKS NOW.

For a very very interesting strange reason, again, I had to copy tnl.framework, navigate through my new bitfighter.app > Contents > Frameworks, and drop that in.

In there it uses Alut, libtomcrypt, and tnl.

Now it was successfull and I was able to connect and play normal games.

Sshhwew

[watusimoto]

As your reward for perserveering, if you want to try something totally whack, edit the file projectile.cpp, and uncomment all the commented stuff starting at line 270. Recompile, and make a level with a test object in it, and see what you think!

[ZoombeR]

IT SHOOTS ALL FUNNYED

btw, while i was compiling it, I got an error on this line
F32 pull = min(100.0f / dist.len(), 1.0); // Pull == strength of attraction


taking out the f and making it this:

F32 pull = min(100.0 / dist.len(), 1.0); // Pull == strength of attraction

seemed to fix it.
WHOOO

EDIT: oh and something interesting you may like to hear. if i bypass this while line, bitfighter wont pop up but console will! and its reported 1994 reports!

talk about detail

[watusimoto]

EDIT: oh and something interesting you may like to hear. if i bypass this while line, bitfighter wont pop up but console will! and its reported 1994 reports!

I'm not sure I understand what you mean...

[ZoombeR]

Never-mind that, I'm not really sure what happened there, and I was being very unclear in my way of describing what I did. I excluded just one of the lines you showed me,

(the only line that gave me an error before I found the real fix to it)

When I launched it, bitfighter didn't open, and instead, an application called "console" opened, (which is used to display logs).

Console stated1994 messages were reported by bitfighter, meaning to the problem of bit fighter not making logs, something was actually there!

I will take some steps and try to trace into this. If Im crazy enough to spend the night searching, I may find an answer to the blank log problem.

oh, and I think that feature you showed me was super-super. It gave me the idea to open up a whole new game-type, using the engineer module's base structure just to mount resource items on you. You then cannot place any turret or forcefield down, (for i have turned it off/errr dont even know how to turn that part on) but now mounted, a nearby ship's phasers will direct to you. I'm also going to try to implement that into a soccer ball, that will really make things interesting.

[watusimoto]

OK, so you said when you commented out a certain line, the logs began to work. What was that line?

[ZoombeR]

F32 pull = min(100.0 / dist.len(), 1.0); // Pull == strength of attraction

odd

[watusimoto]

Removing that line makes logging work????

Are you sure about that?

[ZoombeR]

Sorry, that was the message count for my whole coputer, not btifgither; I'm not sure why console opened, but I couldn't repeat the steps to make it happen again, meaning something else must have been going on at the same time.

BUT

In console, i did a search of "bitfighter", and i DID find ALOT of messages about a bitfighter crash, or something other, meaning there IS a log going on. I also looked at the sender. About half of the messages were sent by bitfighter, and the other half were sent by crash reporter.

update:
Even bigger, I got a post count of 4000 messages over 4 days of playing bitfighter.

(meaning bypassing that line did nothing, but there were messages already)

I checked through the log, and the server messages like client Xx connected or got ping packet, or whatever else, started showing up at august 19th and earlier. Unfortunately, after that. the only messages were crash reports and exits.

[watusimoto]

So if you would kindly play in debug mode for the next few days, and if it crashes, let me know where. That would be very helpful.

[ZoombeR]

Of course I can do this.

and now the whole svn has stopped working agian.

I think my errors are related to framework paths not being found or something, but all of them are in place.

The errors I get again are the sound reference errors like i did before i had the frameworks all in place, only this time, i have the frameworks in place

Update: ok, I got a whole new svn, with a higher revision = more s
Re-did everything, which wasn't much the third time around
And got the debugger back online.
I guess the god of debugging has smiled upon me, as he has also given me an idea to fix logging.
2 actually

Overall Outstanding problem: I have never gotten the sound to work. I thought It was just because I didn't copy the sfx folder into the debug folder, where bitfighter.app is.

Once I did copy the folder over, bitfighter would crash. only if i deleted the sfx folder again, (or delete the contents of the sfx folder), would bitfighter work again.

I feel this may also relate to the error I had earlier, and can be changed by also changing something with the frameworks as well.

Ill edit this post for updates on this subject:

[watusimoto]

Did you update? Ryan checked in some changes on the xcode project, and so if you updated, his settings may have clobbered yours. We need to figure out what you're changing, and, where appropriate, get those changes into the system so it doesn't happen again.

[ZoombeR]

All I did was simply download the svn again, to a different folder. Once I figured out the steps I needed to take to get the game built, I can compile it pretty fast. The problem is, on both the older version and the newer version (never mixed revisions), both had sfx loading problems on my computer, and although building successfully, crash when I put the sfx folder in.

UPDATE:

I fixed the problem. I realized, I needed OpenAl.framework to also be inside Bitfighter's frameworks folder, during launch. Just for resources, I suppose.

Upon compile, Find the newly built application, Right click on it. Select "show package contents". Navigate through contents and create a new folder called frameworks if there isn't already one.

These frameworks are mandatory, they MUST be in the folder for bitfighter to work:

ALUT.framework
Libtomcrypt.framework
tnl.framework
openal.framework

I'm not sure about the rest. They are used in building the application, but I've been able to run it successfully, without any unusual errors, or many errors at all. I would Include them just in anyways:

Cocoa.framework
Coredata.framework
Foundation.framework
GLUT.framework
IOKit.f
OpenGL.f

I don't know if theres a way to automatically copy the frameworks inside Bitfighter.app, so this may have to be given to anyone who wanted to know how to get their copy to open, if there isn't a way.


If you want, I can post all the steps I did, to get my build to work fully, for reference until the errors/work is fixed/done beforehand.

[watusimoto]

That would be great. Please put the info in the bottom section of this wiki page:

http://bitfighter.org/wiki/index.php?ti ... Bitfighter

[ZoombeR]

alright, cool.
and just a request if its possible, can you or ryan remove tnl.xcode located in the OSX folder, or even the 0SX folder completely, if all the files in there are not needed? I feel this is the most confusing part to new builders, as it would seem kind of obvious to look there first right? I know the xcode project in there hasn't been touched for months, being the old one that has MyOpenGlView still in it and having the executable be TNlTest instead of Bitfighter.

EDIT: I took my osx folder out, and bitfighter seems to compile successfully without it, which means the files in the osx folder, must be only relevent to the ancient tnltest project.

[watusimoto]

Wow, that's a lot of work!

I've done a little formatting to try and break it up a bit, and forwarded your request on to Ryan.

Thanks!

[ZoombeR]

Cool!
Also, I recall you saying you changed some of the framework paths from absolute to relative?. If we could get a revision with all relative paths, then all we would have to do is build the frameworks, not choose their paths, and therefore cut the wiki instructions making it quicker and easier to build bitfighter.