[BUG] speechd-up fails to detect one-character messages

[Alexander E . Patrakov]

2008/1/7, Hynek Hanke <hanke at brailcom.org>:
> the current code used for character detection is not a good
> solution -- it only make guesses as to whether the message
> represents a keypress or not, because Speakups communication
> doesn't distinguish between keypresses and ordinary texts.
> As guesses, they will always fail in certain situations.

Just for your information: this is not only for keypresses (that one
already knows given not-really-fat fingers, and keys are not even
echoed by default in yasr), but also for commands to examine
characters on the screen (that are vital for getting information, and
that's why CHAR may be more appropriate than KEY - is there any method
to distinguish between these two cases?).

> Thanks for the excelent description of the problem. I think
> it is easy to modify the code so that the situation described
> also fits in, so I hope to do and test it this week.

A rough patch (used to plug the bug on the current stable LFS LiveCD)
is available at:

http://wiki.linuxfromscratch.org/livecd/browser/trunk/packages/speechd-up/speechd-up-0.4-onechar-1.patch?format=raw

but it can be probably done better instead of testing only the first
byte of the buffer for the "stop talking" character.

-- 
Alexander E. Patrakov