[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFS: cellwriter



I modified the source a bit to make it compile in Etch and built with
cowbuilder for this upload (1.0.2-1).

On Fri, 2007-09-07 at 22:43 -0300, Nelson A. de Oliveira wrote:
> # Add here any variable or target overrides you need.
> 
> As a suggestion, remove this line.

Done.

> It's not necessary to install README (this info is already on the
> package description and on the manpage).
> AUTHORS is also included on debian/copyright, so I don't see a need to
> install it too.

It's a GNU standard to have it, so why not? I've actually just added
some information to the README which isn't mentioned anywhere else.

> I strongly recommend to you create a debian/menu file.

Done.

> All the files inside your tarball have the execution bit set. It's not
> need to have all the files executable (just a suggestion for your next
> release).

An unfortunate side-effect of compiling on a FAT partition (I dual
boot). I've recompiled on an EXT3 and hopefully set all the permissions
correctly this time.

> And I have tested cellwriter and it's not working :-(
> While training its input, the letters don't get darker with training.
> Any ideas why it's happening?

It would be unusual for there to be a bug there. The letters get darker
gradually, it's not very noticeable. When you draw a letter, you must
move the mouse out of the box so that the ink disappears.

If you train the program this way, click Finish, write something and
your input isn't recognized, it still doesn't mean that the training is
broken. If you draw an input character that is very different from the
training samples, CellWriter will reject it. Most often this happens
when someone trains a character by drawing it twice without moving the
pointer out of the box (so the program expects you to draw that
character twice this way every time!) or if you put a dot where there
shouldn't be one on accident.

-- 
Michael Levin <risujin@risujin.org>



Reply to: