Hi James,
As noted here before, using a multi-threaded activity interferes with power > management. Read Etexts will have functioning power management if > speech-dispatcher is not installed, but will not if it is installed, even if > you never use the Speech feature. I know Hemant Goyal was working on that > problem and I don't know where he left it. I did explore a few solutions to the problem and make the speechd API use gobjects to make the communication with the speech server asynchronous. I believe a certain bugfix in python 2.6 will resolve the problems we are facing, hence for the moment it was decided to postpone the modification of the speechd API (also modification of the API implementation was not so straightforward). > He was also working on packaging RPMs for speech-dispatcher on Fedora. He > has some that work but have many dependencies that chew up scare disk space. > I know he's working on an alternate set of RPMs with fewer dependencies for > OLPC. I haven't tried them. The latest package (OLPC-2 Release 18) can be retrieved from http://koji.fedoraproject.org/koji/buildinfo?buildID=57201 I have removed the pulseaudio and nas dependencies. You do not have to be root to be able to run the speech-dispatcher package on XO. To install speech-dispatcher please uninstall the previous package that you installed using rpm -e speech-dispatcher and then install the latest pacakge. To run speech-dispatcher execute: /usr/bin/speech-dispatcher -C ~/.speechd/ -P /tmp/speechd.pid In case you run into any trouble please let me know. The instructions in the wiki for installing speech-dispatcher are incorrect > and out of date. I'm hoping Hemant will update these as well when he is > ready. Again, until things are not finalized (they seem to be changing so much) I dont want to update the documentation. Cheers! Hemant
_______________________________________________ Devel mailing list [email protected] http://lists.laptop.org/listinfo/devel
