Installation instructions for Speakup Version 1.5
This document assumes you have had some experience with kernel
compilation and installation.  If you have not, I recommend you get
the kernel source and read the README and various documents in the
linux/Documentation directory.  In particular the Changes file to make
sure you have the appropriate utilities needed for installing a 2.2.xx
or 2.4xx kernel.  It isn't as difficult as you might think.  The
kernel README is intimidating the first time but once you get the
steps down, it's really pretty easy.  Getting through the "make
config" is the tedious bit.

The first thing to do is to place a copy of the tarball in the /usr/src
directory which is the directory the linux tree is located in as well.
Note: when unpacking the linux-2.4.20 kernel sources you end up with a
directory called linux-2.4.20 instead of 'linux' as in previous kernel
versions, to fix this you must create a link to the tree by
typing the following.

ln -sf linux-2.4.20 linux

Next untar speakup by typing:

tar zxf speakup-1.5.tar.gz
cd speakup-1.5
./install

Note the dot-slash before the install.  This will copy the speakup
directory to the kernel tree and apply the various patches and
components to the appropriate kernel files.  Depending on how
experienced you are with kernel compiling and hacking will determine
whether you should bother looking at any failed patches.  If this
happens, you should probably write to the speakup mailing list for
help or myself.

If all of the patch hunks apply successfully then just continue with
the standard steps to compile the kernel with:

make mrproper
make config

When you get to the section console speech output, answer 'y' to the
CONFIG_SPEAKUP prompt.  You will be given a submenu with the list of
synthesizers which are currently supported.  You can include as many
synths in the kernel as you wish but remember each one takes up kernel
space.  You can only choose one of the synths as the default or none,
so just type dtlk or whatever is the correct string for the
synthesizer you have.  You will also be asked if you wish to build-in
a speakup key map.  If you do not say 'y' to this option you will need
to load a speakup map at boot time with whichever mechanism your
distribution uses for loading key maps.

We have placed the speakup configuration options in make config just
after the vga console choice.  For the DoubleTalk PC driver included
by Jim Van Zandt.  I recommend you say no to that option.  I have not
tried configuring them both in, but I wouldn't be at all surprised if
it didn't work.

If all goes well up to this point you can continue with the compiling
process by doing:

make dep >dep.file 2>&1 &
make bzImage >cc.file 2>&1 &
make modules >mod.file 2>&1 &

I always redirect output to the files dep.file and cc.file so I can
look over the compilation record to make sure there are no errors and
warnings.

Okay, you are ready to install the newly compiled kernel.  Make sure
you make an linux.old entry in your lilo.conf file so you can recover
if it blows up.  next as root run "make modules_install" to install
whatever modules you compiled and move the bzImage from
/usr/src/linux/arch/i386/boot to wherever your kernel lives.  Also
move the System.map from /usr/src/linux to where your System.map
lives.  On our systems we use debian so we create an vmlinuz-speakup
and System.map-speakup in our /boot directory and set the symbolic
links vmlinuz and System.map in the root (/) directory to point to the
images.  Now type lilo to tell lilo to build the new booter file and
install it.

As of version 0.07, the keymap for speakup is automatically built in
at compile time.  If you have other keymaps installed at boot time,
you might want to consider removing them before you reboot the system.

If everything has gone OK up until now, cross your fingers and type:

shutdown -r now

Your system should start talking to you as soon as it starts booting.
It will talk and talk and ... well, you might want to hit the
keypad-enter key to tell it to shut up.  You should also read the
DefaultKeyAssignments file to learn the various review functions
available.

As of v-0.10 the speakup configuration options are in the
/proc/speakup subtree.  The individual options should be fairly
obvious by their names such as rate, volume, punc_level and so forth.
You can manipulate them by cat'ing or echoing new values to them such
as:

echo 9 >/proc/speakup/rate

You can see what the current values are by cat'ing those files to the console:

cat /proc/speakup/rate

I have probably managed to overlook a whole whack of things because
this is the, enter version number here, draft.  Don't worry we'll get
it right eventually.  If you like the package you really should get on
the mailing list and start participating in it's development.

  Kirk

email: kirk@braille.uwo.ca
phone: (519) 679-6845 (home)