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

Re: my Bidi implementation



I checked my old archives yesterday and I found that I still have
the very first releases of fribidi where I was the sole contributor.
I personally have no problem releasing these under a putty compatible
licence. It would save Ahmed some work if he didn't have to 
reimplement this. I think it would be good to rename the sources
though so that there is no confusion with fribidi.

Do you have anything against it, Behdad?

What it would give you:

   * A complete (but probably somewhat buggy) implementation of 
     the implicit unicode bidi algorithm.

   * No support for explicit overrides (but, hey, would use those
     in a terminal emulator!)

   * Less than optimal speed wise. No sophisticated lookup (neither
     two-level nor nine-level. ;-)

Regards,
Dov


On Sat, Mar 13, 2004 at 10:49:27PM -0500, Behdad Esfahbod wrote:
> 
> Without checking the specs I can tell you it's rule N1.
> 
> behdad
> 
> On Sat, 13 Mar 2004, ahmad khalifa wrote:
> 
> >
> > > > i have a quick question...
> > > > for a string that looks like this:
> > > > asdf ASDF GHJKL OPNM
> > > > the paragraph level is 0...
> > > > all capitals are odd level types (i.e type AL or R)
> > > > is this the correct level array..??
> > > > asdf ASDF GHJK OPQW
> > > > 0000011110111101111
> > >
> > >No, I think it should be like:
> > >   0000011111111111111
> >
> > you are right... but, where does the space get the level 1..?
> > it says at the top of UAX#9 that spaces get the level of the
> > surrounding text... but which rule does that...?
> >
> > in the (Xn) rules, spaces get levels of 0 because there are no
> > explicit embeddings...
> >
> > in Rules (In) it deals with strong characters and numbers only
> > thats why ASDF gets the level of 1, and the space remains
> > at level 0...
> >
> > in rule (L1) it doesnt deal with spaces between strong
> > character types...
> >
> > to conclude, no rule states what to do exactly with WS
> > between strong character types (i.e R, AL, L)...
> > is it left to the implementation? i dont think so...
> >
> > its logical enough to parse any sequences of WS and give
> > them the highest level of the nearby runs, or give them the
> > type of the nearby runs, but is that the standard way..?
> >
> > any ideas what im doing wrong..?
> >
> > ak.
> >
> > _________________________________________________________________
> > STOP MORE SPAM with the new MSN 8 and get 2 months FREE*
> > http://join.msn.com/?page=features/junkmail
> >
> > _______________________________________________
> > Developer mailing list
> > Developer at arabeyes dot org
> > http://lists.arabeyes.org/mailman/listinfo/developer
> >
> >
> 
> --behdad
>   behdad.org
> _______________________________________________
> Developer mailing list
> Developer at arabeyes dot org
> http://lists.arabeyes.org/mailman/listinfo/developer