Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo
I've had success building the project with the following (updated) instructions. Please feel free to use this (or not) as you please!
- Microsoft Visual C++ 2008 Express Edition with SP1 (http://www.microsoft.com/express/vc/)
- Perl (cygwin perl (http://www.cygwin.com/) or activestate perl (http://www.activestate.com/activeperl/) )
- Python (2.5 used: http://www.python.org/)
- Gnuwin32 gawk, gzip, flex, bison (and it's dependancies), sed (and it's dependancies (http://gnuwin32.sourceforge.net/)
- Copy gawk.exe to awk.exe
- Make sure that the gnuwin32 binaries are in a directory path with no spaces, like 'C:\gnuwin32\bin'
- Make sure the environment PATH includes the directory where the gnuwin32 binaries are
- Download/install a command line version of svn. http://www.sliksvn.com/en/download (other versions are fine too).
- Make sure that the directory where the svn.exe file is, is added to the System PATH
- Make sure python, perl are in the environment PATH
- make sure the command prompt is set for compiling with the visual studio compiler (vcvars32.bat)
- open freetype\freetype.sln in Visual C++ 2008 Express Edition
Build configurations 'Release Multithreaded' and 'Debug Multithreaded'
- run 'perl Configure VC-WIN32' in the openssl directory
- run 'ms\do_masm.bat' in the openssl' directory
- run 'nmake -f ms\nt.mak' in the openssl directory
- run 'nmake DEBUG=1 -f ms\nt.mak' in the openssl directory (if you want to build the DEBUG version of vcxsrv)
- goto pthreads directory
- run 'nmake VC-static'
- run 'nmake VC-static-debug'
- goto mesalib\windows\VC8\mesa
- open mesa.sln
- Build configuration 'Release'
- open tools\mhmake\mhmake.sln in Visual C++ 2008 Express Edition
- Build the Debug and Release configurations
- Copy Debug\mhmake_dbg.exe and Release\mhmake.exe to a directory in your environment PATH
Add an environment variable named MHMAKECONF having as value the root directory of the sources.
e.g.: set MHMAKECONF=c:\vcxsrv\trunk
- run 'mhmake MAKESERVER=1' in xorg-server directory
- run 'mhmake DEBUG=1 MAKESERVER=1' in xorg-server directory for a debugable version
To build installer:
- install http://nsis.sourceforge.net
- run 'packageall.bat' in xorg-server\installer
To clean the project tree (uses svn) go to the top directory at a command prompt and type in 'python tools\cleantree.py -r'
On another topic what is the odd comment about GPL3 OpenSSL all about? I assume that your work is GPL3 and all the other software is under it's original license. We are all going to need to go to Law School pretty soon :(
Also, I tried to apply the hardware opengl patches from Xming….but got lost rapidly in gcc/msvc incompatibilities - it will have to wait for a REAL c programmer to look at.
Thanks for the info.
Indeed only the stuff I am writing is GPL3, all other software is under it's original license. I do now know why that person has written that comment.
I am currently also adding the patches for using the native wgl. It is compiling and running, but it still does not use any hardware acceleration. Still investigating …
It seems that in this implementation hardware acceleration is only working in multiwindow mode (-multiwindow) and when doing indirect rendering (export LIBGL_ALWAYS_INDIRECT=1). You also have to give -wgl on the command line.