| |
Required Hardware (One of...): - 1 : i386 : "old" IBM PC, or "new" IBM PC with old compiler
- 2 : i686 : "new" IBM PC with recent compiler
- 3 : ppc : "new" Macintosh (IS NOT SUPPORTED YET - HELP US)
- 4 : armv4l : Corel Netwinder (but jMax does not run on this)
- 5 : mips : Silicon Graphics, non-PC (only tested without jMax)
Required Software: - 1 : Ruby 1.6 (language interpreter)
- 2 : GNU C Compiler (gcc)
- 3 : GNU Make (gmake)
- 4 : GNU Bash
Very Recommended: - 1 : jMax 2.5
- 2 : jMax 3.0 support coming soon
- 3 : jMax 2.4 support removed
- 4 : X11 Display Server and MIT-Xlib (already in linux)
- 5 : instead of X11 you may use SDL.
Optional: - 1 : HeroineWarrior's libmpeg3.so
- 2 : HeroineWarrior's libquicktime.so
- 3 : Greg Ward's libmpeg.so ("Berkeley/MNI")
- 4 : a digitizer card with a Video4linux 1.x driver (videodev.h)
- 5 : Ruby 1.6 with package "xmlparser" (for editing documentation)
NOTE:
for help on installing and compiling jMax, please see Christian Klippel's
extensive help file at nil
as well as the jMax docs.
Also Useful: - 1 : Pentium-compatible CPU (for profiling)
- 2 : CVS (for live update)
NOTE:
There are several different incompatible libmpeg's from different authors;
I found four of them, some numbered 2 and 3 as if they were version numbers
(but it seems they are not).
NOTE: if one of your libraries are only available in .a format, you will not
be able to compile GridFlow. All libraries GridFlow uses must be in .so format.
|
| |
- 1 : if you don't have jMax, get it at
jMax's site .
If have only the RPM, you will need to install this:
jmax headers
in whichever directory as long as you tell it to GridFlow's
configurator.
- 2 : Install Ruby 1.6.6 or any version released in 2002 or later (the
latest from the 1.7 series work and are faster) You will need the file
libruby.so, which is not always included. If you're using an
installer and you're missing the file, then look for another package of a
similar name that may contain extra files (eg: "ruby-dev"). If you're
installing from source code you have to configure with the
--enable-shared option. If you install into a system directory, you
have to recreate the library index using the ldconfig command too
(auto-installers normally do this for you).
- 3 : Download the latest version of GridFlow
and download the images pack (gridflow-images-0.5.tar.gz).
The download site is artengine.ca.
Unpack Gridflow in /usr/local/jmax/packages (or the path of your choice),
which will create a directory called 'gridflow';
unpack the images pack into that directory, which will create
another one called 'images'.
- 4 : Find your jMax ARCH value. This is the one you use when installing
jMax (or other jMax plugins other than GridFlow) manually. For K6,
PentiumPro and such, on Linux, this is i686-linux. If you are
trying to run it on a Macintosh, this would be ppc-macosx or ppc-linux.
- 5 : Run ./configure --jmax-arch your-ARCH from the gridflow directory.
note: if you didn't install GridFlow in the jmax/packages
directory, then you have to add the option --jmax-dist-dir your-jmax-source-directory-goes-here. Giving /usr/lib/jmax will not do because it wants include and Makefiles subdirectories, which do not get installed with
jMax.
- 6 : ./configure will try to detect some optional components (libmpeg,
pentium clock, etc); error messages may appear in case of a no.
usually a yes will not show anything else. In the end, the files config.make, c/src/Sources and c/src/config.h will be generated.
- 7 : Stay in the same directory and do:
- 1 : make all
- 2 : make install
but if the first one stops because of "Error" you shouldn't try the second
one. - 8 : Load ~/.jmaxrc (text). Go to the when start section.
Add a line that says package require gridflow.
Add this line that says dataDirectory gridflow-directory/images
- 9 : (Re)start jMax.
Note: If there's a problem with libmpeg anywhere in the process,
you can add the option --no-mpeg to the configurator. Note: If you want more speed, you can configure with the --fast
switch, which turns off many error-detection mechanisms; if you want
more speed, you could configure with --no-profiling too, but if you do
that, you can no longer find out which of the objects consume the most
time.
|