You do like to ask good questions. I just wish I had good answers. Gregory Nowak <greg at romuald.net.eu.org> writes: > What I'd like to know is simply if I understand the following > correctly, or not. If you build speakup as modules, you can switch > from synth to synth, but you lose the ability to have speakup come up > as soon as possible, before the file systems get mounted, (this > assumes you aren't using an initrd). If you build > spkglue, and speakup into the kernel, you'll get speech as soon as > possible, and you can still switch to another synth by loading its > module later on, but the only way to switch back to a built-in synth > as of now is to reboot the machine. Do I have this right, and if so, > am I correct also in assuming that this gives more weight to the > argument for using an initrd, as opposed to not using one? You've evaluated the loading correctly. I don't think using initrd really has anything to do with it other than determining how early a root file system gets loaded. We probably should consider having speakup fall back to a built-in synth in the case of a module being removed but there is so much other work needing done I doubt it'll happen. > Also, since /proc/speakup/synth_name, or an equivalent doesn't exist > as of now, is the best way to check if a synth is already loaded or > not in a script by doing a grep against lsmod, or is there a better > way? That works. checking /proc/speakup/version should work also. > Finally, does the version of speakup currently in git support usb to > serial converters, and other nonstandard ports, or is that still to > come? I'm pretty sure the answer is no, but I thought I'd ask Keep dreaming! 'rin' I doubt speakup will work with 2.6.23 currently either as it stands. -- Kirk Reiser The Computer Braille Facility e-mail: kirk at braille.uwo.ca University of Western Ontario phone: (519) 661-3061
