[BRLTTY] A bug in revision 3008-09 of version 3.8dev?

[Dave Mielke]

[quoted lines by Nicolas Pitre on 2007/04/30 at 12:00 -0400]

>Better yet, the BRLTTY version could be stored in every dynamically 
>loaded drivers and BRLTTY refuse to use them if the version doesn't 
>match.  This is the best way to avoid mysterious bug reports like this 
>one was.

I like this aproach the best as it's automatic and doesn't require human
guesswork. Perhaps, though, brltty should issue a warning rather than an error
just in case it'll keep on running well enough for the user to read it.  This
should be okay as long as we ensure that the mismatch is logged before any code
within the dynamic object is executed.

The nicest way would be to store a true string, i.e. a pointer to a string of
variable length. This may cause a new core with old drivers (which don't yet
provide whis data) to crash, though, so do you think we should resort to a
fixed-length field to store the version number?

Do you think there's any reason to retain the current driver version numbers? 
They aren't really being maintained properly anyway (finger pointing at
myself). Maybe we should just store brltty's version in that already-existing
field (which is a real string).

-- 
Dave Mielke           | 2213 Fox Crescent | I believe that the Bible is the
Phone: 1-613-726-0014 | Ottawa, Ontario   | Word of God. Please contact me
EMail: dave at mielke.cc | Canada  K2A 1H7   | if you're concerned about Hell.
http://FamilyRadio.com/                   | http://Mielke.cc/bible/