This debate regarding scripting emacspeak installation comes up regularly. The reason I believe things don't change much is because the effort of maintaining all the different packaging systems way exceeds the benefit because emacspeak is actually a fast evolving system. Keeping the packages up to date becomes prohibitive. Adding in the need to do both 32 and 64 bit makes it worse.
In the overwhelming number of cases where I've assisted someone to get emacspeak working, the issue has boiled down to one simple problem, failing to ensure the speech server is actually working prior to trying to run emacspeak. You cannot just download the sources, run make config, then run make emacsepak and then fire up emacs and expect it all to work. We might like it to be that way, but that just isn't how it works.
When doing an emacspeak install, before trying to run emacspeak, it is CRITICAL that you first run the speech server script i.e. outloud or espeak or whatever speech server you plan to use. If this doesn't work, then you will almost always get an error message. The point is, if you haven't done this before running emacspeak, all you will get is the dreaded "Process speaker not running" error. This error doesn't give you enough information to diagnose the problem. On the other hand, if you run the speech server script from the terminal, you will get a much more meaningful error message. It could tell you that your missing the Tcl or Tclx libraries or it could tell you that it cannot find one of the viavoice files or the espeak libs etc or it might tell you that one of the library files is of the wrong format i.e. it is finding 64 bit libs, but your running viavoice and need the 32 bit versions etc. Even if you don''t know what the error message means, it is very likely that posting that error to this list will result in answers which will address the specific problem.
The other thing I would recommend is not running make install. You do not need to install emacspeak system wide and can easily run it from within your home directory. You will need to add a line into your emacs init file to tell emacs where to find emacspeak, but once you have that working, your set. Doing it this way also makes updates really easy. I just ahve the emacspeak svn repo checked out into my home directory and point emacs at that directory. When I want to do an update of emacspeak, all I need to do is a svn checkout, make clean, make config and make emacspeak. If it doesn't build for some reason, which is vary rare, I just revert the svn to the previous known good revision and do make clean make config and make emacspeak.
Getting the speech server to run initially can be difficult, especially if your running viavoice and your on a 64 bit system. However, nearly all the linux distros now support running 32 bit programs on 64 bit systems in a straight-forward manner. Possibly the hardest things is getting the right tcl and tclx setup. The key here is not to go with the latest version of Tcl/TclX. However, provided you run the speech server scripts directly in the terminal, you will get the info you need to diagnose the issue. Once you have it setup once, you almost never need to do it again until the next major upgrade of your linux distro and even then, it will probably be OK.
Tim
If you have questions about this archive or had problems using it,
please contact us.