[SpeechIO-6] speechd v0.31 (cat CHANGELOG > /dev/speech works)

[Darxus <darxus@Op.Net>]

Can now handle "cat CHANGELOG > /dev/speech".  (which a couple people had
complained about not being able to do)

Previous versions had 1 of 2 problems

1) tts_file or SayText w/ $/ undef'ed: Wrote the whole file to festival at
once.  So festive tried to render the whole document into a single sound
file.  As you might imagine, this takes up a rediculous amount of physical
& virtual ram, & cpu.

2) SayText w/out $/ undef'ed:  Apparently if you open /dev/speech, read
one line, and close it, all the other lines are then lost.


v0.31 does not have $/ undef'ed (so it gets text from /dev/speech 1 line
at a time)

I changed $text = <FIFO> to while ($text = <FIFO>) {}, wrapping the curly
braces around everything short of close FIFO.

And I checked -- you can, in fact, write more stuff to /dev/speech while
festival is reading off a document in this manner, and it will speak it
when it gets done the document.


Next thing along these lines -- somebody suggested breaking on sentences,
instead of newlines.  Anybody know how to do that in Perl ?

__________________________________________________________________
PGP fingerprint = 03 5B 9B A0 16 33 91 2F  A5 77 BC EE 43 71 98 D4
            darxus@op.net / http://www.op.net/~darxus
                         Far Beyond Reason