Create a directory that will contain fpgasm sources and test designs. Mine is called FPGASM.
Download and upack the files fpgasm-code.tar.gz and TEST.tar.gz into FPGASM directory. You will now have fpgasm-code and TEST subdirectories.
To compile,
cd fpgasm-code ./build
The build script will create the executable fpgasm and make a copy of it in your TEST/src directory. Feel free to copy it into a PATH-visible directory; I prefer to include ./ into my PATH.
Now, cd ../TEST/src
Using your favorite editor, look at the various example files. You can invoke fpgasm directly:
./fpgasm test01.ham
which will dump the xdl to stdout, or specify the output file:
./fpgasm test01.ham test01.xdl
In order to actually configure an FPGA, you will have to use proprieatary tools from Xilinx. If you already have a running Xilinx installation, great. If not, you are in for a day of work. Download and Install Xilinx WebPACK(R) pre-V14 version and pay careful attention to the PATH - you may have to source a file to set the PATH. At the end, you should be able to type xdl on the command line and get meaningful output. Don't bother doing anything else until xdl works.
I would suggest makeng sure it all works by compiling a demo that came with your board and configuring your fpga with Impact(R).
Now, examine the 'build' script. Try
./build test01
MIt will run fpasm and invoke a makefile in the ../xdl directory. The makefile in turn will invoke the Xilinx tools and create a bitstream in about 10 seconds, if everything works. You will see a bunch warnings; that is normal. Watch for errors.
./build test01 burn
will rebuild and configure the fpga using Impact.
Xilinx, WebPACK and Impact are registered trademarks of Xilinx Inc.