Re: Bidirectional Bugs in hebrew

[Chookij Vanatham <chookij vanatham eng sun com>]

Hi Pablo,

] 
] > ] And if we use soft hyphen (0xad, available in iso-8859-6 (Arabic),
] > ] iso-8859-8 (Hebrew), cp1255 (Hebrew + Yiddish), cp1256 
(Arabic+Farsi+Urdu))
] > ] would it produce a rtl string ?
] >
] > I haven't tried this but I think, you should get the same result because
] > "soft hyphen" would not be the R type like hebrew alphabet.
] 
] but shouldn't a neutral char inherit the "default" direction (in this case
] being RTL if you write in Hebrew.
I think that a neutral char should inherit the default direction but
"minus sign" is the week type. Not quite sure if it's supposed to inherit
the default direction. Will need to read/focus more in detail though.

] 
] Well, I'll try it...
] It works!:
That's great. This is interesting. I'll try it sometime.

] 
] chanae:~$ echo AA-123BB | fribidi
] BB-123AA
] chanae:~$ echo AA­123BB | fribidi
] BB123­AA
] 
] Now that should be put somewhere in a FAQ.
] 
] > Because you are using this in unicode locale, so, you can try typing
] > RLO or RLE before typing "minus sign" to force the direction to be Right to
] > left,
] 
] Of course; but that isn't very userfriendly.
I do agree too but I guess that there is still some need that only,
RLO,RLE,LRO,LRE,PDF would be able to help. I don't have these experence on
hand but will try to use more arabic/hebrew...

] Having the keyboard map send a soft hyphen (0x00ad) when in Hebrew/Arabic mode
] instead of a "minus" (0x002d) (the numeric keypad should always send
] a 0x002d however, IMHO).
] That would allow much easier typing.
I would say so for this case. 

] 
] A real life example: "8-bit" in Hebrew: "8-á" gives "á8-" which is wrong;
] and "8­á" gives the right "á­8".
] When people are typing text they want the hyphen to follow the direction
] of writting; so having the kbd send a soft hyphen seems a good solution.
] What do you think about that ?
For this context, I would like to have "hyphen" followed the text (to the
right) then, because "soft-hyphen" it works without using LRO/LRE/RLO/RLE/PDF,
then, I would prefer to use "soft-hypen" but how to input it is another point.
Either mapping it like this, I'm fine too. They might be other way to input
soft-hyphen as well.

] 
] And are there any implications in using 0x00ad instead of 0x002d for 
] "hyphen" meaning in Hebrew/Arabic strings ?
] 
] > I haven't tried it but I think fribidi should work with RLO,RLE,PDF as well.
] > For the keyboard input, not too sure if gtk+1.3 is enabled to allow users
] > to input those RLO, RLE, PDF codes.
] 
] AFAIK Gtk+ doesn't handle the keyboard; X11 does.
Yeap. That's correct.

I have the suggestion that how about if we can control which characters
are in which bidirectional types through other kinds of external configuration
files. Then, let's say, under hebrew locale, if we want "minus sign" to be
always written to the right then we just put "minus sign" codepoint into R type
(right-to-left type). Then, everytime in hebrew locale, "minus sign" will be
written to the right. This is just my opinion and it might need to be done
in fri-bidi.


Chookij V.