[FriBidi] Attributed unicode

[Robin Becker]

Thanks for the reply,

 > Robin,
 >
 > Why don't you track the position of your target (the CAPITAL LETTER I here)
 > via the log2vis mapping that most bidi implementations, including FriBidi,
 > provide?
 >
 >
 > -Behnam

unfortunately I know little about the internals of fribidi. I have been using 
work of others who contributed to reportlab quite a while ago. Although I have 
been interested in printing since the 70's I don't have a lot of knowledge in 
non-latin scripts. Currently reportlab routinely uses a very old version 
pyfribidi/fribidi. Looking in the pyfribidi code I see this


fribidi_log2vis(logical, (const FriBidiStrIndex)length, &base_direction, visual,
                              NULL, NULL, NULL)

which I think is leaving out the mapping stuff entirely

The pyfribidi code seems to be unmaintained for a long time and when I search 
fribidi I find fribidi_log2vis in fribidi-deprecated.c so I was a bit concerned 
I would be doing the wrong thing to try and improve based on the log2vis 
interface in pyfribidi.

Is there a less deprecated interface to do text reordering/shaping or is it 
worth providing for the mapping with an improved pyfribidi log2vis?
-- 
Robin Becker
On 05/02/2018 23:21, Behnam Esfahbod wrote:

> 
> 
> On Mon, Feb 5, 2018 at 3:13 AM, Robin Becker <robin at reportlab.com> wrote:
> 
>> Not sure if this is the right place to ask, but I would like to know if
>> there's a way to attach attributes (eg colour) to the unicodes that enter
>> the fribidi api. In the reportlab case we allow rather arbitrary changes to
>> the unicode strings that represent text eg
>>
>> paragraph('<span color="red" fontSize="14">I</span>n the beginning.......
>>
>> where the initial capital I is larger and a different colour.
>>
>> If this happened in a word that I want to pass to fribidi is there a way
>> to see where the attributes change in the return from fribidi? I imagine
>> this must be an issue in html rendering as well.
>>
>> I looked at various unicode mechanisms (eg variation sequences), but none
>> seem appropriate to changes of font, size, weight, style or colour.
>> --
>> Robin Becker
>> _______________________________________________
>> fribidi mailing list
>> fribidi at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/fribidi
>>
> 
> 
>