View Full Version : Virtual Jaguar CVS doen not compile
Belegdol
June 22nd, 2004, 12:27
Have a look at this:
[jsikorski@DOMOWY virtualjaguar]$ ./compile
*** Checking compilation environment... OK
*** Compiling Virtual Jaguar for generic Unix/Linux...
mkdir obj
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/anajoy.cpp -o obj/anajoy.o
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/blitter.cpp -o obj/blitter.o
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/cdrom.cpp -o obj/cdrom.o
src/cdrom.cpp:313:59: warning: trigraph ??) ignored
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/cdintf.cpp -o obj/cdintf.o
In file included from src/cdintf.cpp:25:
src/cdintf_linux.cpp: In function `bool CDIntfInit()':
src/cdintf_linux.cpp:19: error: `WriteLog' undeclared (first use this function)
src/cdintf_linux.cpp:19: error: (Each undeclared identifier is reported only once for each function it appears in.)
src/cdintf_linux.cpp: In function `bool CDIntfReadBlock(unsigned int, uint8*)':
src/cdintf_linux.cpp:29: error: `WriteLog' undeclared (first use this function)
src/cdintf_linux.cpp: In function `void CDIntfSelectDrive(unsigned int)':
src/cdintf_linux.cpp:41: error: `WriteLog' undeclared (first use this function)
src/cdintf_linux.cpp: In function `uint32 CDIntfGetCurrentDrive()':
src/cdintf_linux.cpp:46: error: `WriteLog' undeclared (first use this function)
src/cdintf_linux.cpp: In function `uint8* CDIntfGetDriveName(unsigned int)':
src/cdintf_linux.cpp:51: error: new declaration `uint8* CDIntfGetDriveName(unsigned int)'
src/include/cdintf.h:15: error: ambiguates old declaration `const uint8* CDIntfGetDriveName(unsigned int)'
src/cdintf_linux.cpp: In function `uint8* CDIntfGetDriveName(unsigned int)':
src/cdintf_linux.cpp:52: error: `WriteLog' undeclared (first use this function)
src/cdintf_linux.cpp:53: error: `NULL' undeclared (first use this function)
make: *** [obj/cdintf.o] Error 1
[jsikorski@DOMOWY virtualjaguar]$
Any suggestions? I am using Mandrake 10.0
Shamus
June 22nd, 2004, 17:53
Uh, yeah. ;) Seems I forgot to put an #include in the Linux and OSX back ends. Should be fixed in CVS now, so update and compile and all should be well. :thumb:
Belegdol
June 22nd, 2004, 19:57
Still does not work :(
[jsikorski@DOMOWY virtualjaguar]$ ./compile
*** Checking compilation environment... OK
*** Compiling Virtual Jaguar for generic Unix/Linux...
mkdir obj
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/anajoy.cpp -o obj/anajoy.o
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/blitter.cpp -o obj/blitter.o
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/cdrom.cpp -o obj/cdrom.o
src/cdrom.cpp:313:59: warning: trigraph ??) ignored
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -Isrc -Isrc/include -I/usr/local/include -I/usr/include -c src/cdintf.cpp -o obj/cdintf.o
In file included from src/cdintf.cpp:25:
src/cdintf_linux.cpp: In function `uint8* CDIntfGetDriveName(unsigned int)':
src/cdintf_linux.cpp:48: error: new declaration `uint8* CDIntfGetDriveName(unsigned int)'
src/include/cdintf.h:18: error: ambiguates old declaration `const uint8* CDIntfGetDriveName(unsigned int)'
make: *** [obj/cdintf.o] Error 1
[jsikorski@DOMOWY virtualjaguar]$
Shamus
June 22nd, 2004, 20:18
D'oh! ;) I forgot a 'const' declaration in front of those functions... Should be fixed in CVS now. :)
Lather, rinse, repeat... :thumb:
Belegdol
June 23rd, 2004, 08:09
It works now. Good to be helpful.
Shamus
July 6th, 2004, 17:55
So, how is CVS working for you? How would you like to be a part of the build team (at least until we hear from Mr. Green)?
Belegdol
July 7th, 2004, 07:50
It works slow, but quite fine - I've tried one game, AVP. I've noticed one bug. When there is no Rom folder set up in vj.cfg file, emulator segfaults when clicking on Load...
What would I be supposed to do as a member of a build team?
Shamus
July 7th, 2004, 21:07
Basically, whenever we do CVS or official builds, you'd pull down the latest CVS and compile it and either send it to us or post it somewhere where we can get it. Very simple, really. :)
Belegdol
July 8th, 2004, 07:03
OK, why not?
Shamus
July 8th, 2004, 19:38
Great! We're hoping to do a 1.0.7 release soon--just need to get the OK from Niels and I'll go ahead and tag CVS for a 1.0.7 release. :thumb:
Belegdol
May 26th, 2005, 07:43
Yed I've switched to Fedora Core 3 and tried to build latest cvs. Have a look:
gcc -Wall -Wno-switch -Wno-non-virtual-dtor -O2 -D__GCCUNIX__ -ffast-math -fomit-frame-pointer `sdl-config --cflags` -I. -I./src -I/usr/local/include -I/usr/include -c src/dac.cpp -o obj/dac.o
src/dac.cpp: In function `void DACWriteWord(uint32, uint16, uint32)':
src/dac.cpp:246: error: `exit' undeclared (first use this function)
src/dac.cpp:246: error: (Each undeclared identifier is reported only once for each function it appears in.)
make: *** [obj/dac.o] Error 1
I guess some include is missing.
jonnymorris
June 14th, 2005, 12:23
Hi! I am very new to CVS (I don't even know what it stands for, and have never actually used it). I am experiencing difficulty compiling vj 1.0.7 on OS X:
<command line>:6:1: macro names must be identifiers
make: ** [obj/cdintf.o] Error 1
Sorry if this isn't really the place to be asking about cvs, but for future reference how do I use it, and is that the best thing for me to be doing for bug fixes etc?
Belegdol
June 14th, 2005, 14:33
CVS stands for concurrent version system. there shold be some instrucions regarding cvs checkout on virtual jaguar website.
nwagenaar
August 4th, 2005, 11:34
Correct concerning CVS. But also be warned that CVS is blood-edge development and there for not stable or compilable. CVS is a system to update sources during development of our project(s) without the need to e-mail the lot. Because the developers live on different continents (USA and Europe) it's rather simple to retrieve and use the newest sources.
Offcourse, everyone may use the CVS sources and tinker with it. But please keep in mind that we cannot garantee a stable build. A stable build based upon CVS sources are sometimes released as CVS Beta builds and are allready compiled by us.
Regards,
Niels Wagenaar
Shamus
August 6th, 2005, 16:35
Yeah, that missing exit() function bug should be fixed in CVS now. Not sure how that one slipped through... As far as compiling VJ from CVS sources, you did read the INSTALL file, no? ;) If you didn't, you can't run make directly--you have to use the included 'compile' script like so:
./compile
That should also tell you whether or not SDL is properly set up as well as handling the make process.
Shamus
August 6th, 2005, 16:56
I guess I should also mention that part of the reason I've been so quiet lately (besides a total lack of sleep and free time due to my 8 month old son ;)) is that I've been doing a complete rewrite of the blitter. Those of you following along in CVS can see what an awful mess it is. :p Anyway, even in its current incomplete state it works better than the one we inherited from Mr. Raingeard. C'est la vie. ;) Besides that, I've also been doing a little work on improving the GUI to use dirty rectangles so that we don't have to redraw the entire frame just because the mouse moved 2 pixels. :p That and a new renderer that has yet to be fully implemented that should take care of the changing-VMODE-in-the-middle-of-the-screen problem. :thumb:
Be sure to look for all of these improvements to show up soon in those other two closed-source Jaguar emulators... :D
nwagenaar
August 6th, 2005, 18:10
Anyway, even in its current incomplete state it works better than the one we inherited from Mr. Raingeard. C'est la vie. ;) Besides that, I've also been doing a little work on improving the GUI to use dirty rectangles so that we don't have to redraw the entire frame just because the mouse moved 2 pixels. :p That and a new renderer that has yet to be fully implemented that should take care of the changing-VMODE-in-the-middle-of-the-screen problem. :thumb:
Be sure to look for all of these improvements to show up soon in those other two closed-source Jaguar emulators... :D
Well, concerning the last remark. It's ackward that the other emulators are at a deadend when the developers from Virtual Jaguar/SDL take a "small" break ;) And it's indeed true, the current codebase (allthough not clean and full of remarks, unused code, etc) is much better then what we started with.
Currently I'm going through all my Iomega ZIP backups (I got about 54 ZIP disks lying around here) to look for some of the old code, especially Handy/SDL. Currently I found old code from NeoPocott/SDL and BoyCott Advance/SDL so I'm hoping I can find Handy/SDL as well. If I get the code back (which I really hope so) I'm gonna get it in par with v0.90 sources, rewrite everything what's SDL based and release it. This is offcourse if I can find it, otherwise it's back to the developer editor for to rewrite everything from scratch. And I really do hope I get the code back since the soundcode was something really ackward and needed a lot of hacking :hdbash:
Regards,
Niels Wagenaar
PS : Shamus, currently I'm on #ngemu. So if you get the chance, drop by to say hello. We have a lot of catching up-to-do :innocent:
vBulletin® v3.8.7, Copyright ©2000-2013, vBulletin Solutions, Inc.