Hi! Please find attached speechd-0.56-r3.ebuild. This ebuild adds a patch applied to the perl script /usr/bin/speechd that allows to pass commands to the festival server through the /dev/speech fifo. I give you an example: With the previous version it is impossible to change the voice used by festival server because any string submitted to /dev/speech is translated into a festival command like '(SayText "...")'. So, if you write: $ echo "(voice_lp_diphone)" > /dev/speech The command sent to festival become something like: (SayText "(voice_lp_diphone)") The patched version of the script, before generating the festival command, tests if the input string has a command-like syntax (text between parenthesis, only spaces on the left) and, in that case, send it as-is to the festival server. After that you can type: $ echo "(voice_lp_diphone)" > /dev/speech $ echo "Ciao" > /dev/speech And hear what is expected (italian voice in this case)... Let me know if this patch is useful and correct. Regards, Francesco.
Created attachment 81604 [details] app-accessibility/speechd-0.56-r3.ebuild The upgraded ebuild basically adds only this line to the src_unpack() function: + epatch ${FILESDIR}/${P}-festival-command.patch
Created attachment 81605 [details, diff] app-accessibility/speechd/files/speechd-0.56-festival-command.patch This is the patch file to the perl script /usr/bin/speechd Actually, I'm not a perl programmer, hope it is good enough...
This package has been removed from portage; see bug #360067 for information.
