Mailing-List: contact dev-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Received-SPF: pass (athena.apache.org: domain of mccloud35@gmail.com
 designates 209.85.200.173 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=message-id:date:from:to:subject:in-reply-to:mime-version
         :content-type:references;
        b=MtkAUNcoIrI3lcNFJR7nOG59kUnb/GZ9D2Te6klDdOmnUpb17YMvQAyMMLXMMbzLTM
         sAwofIR8B75FDVTRyD7YDMBPixtjtkbjwdLO0rMXhTRWkqwTI3LIQdFuWCqMEqMrUaGH
         AjI+yVZAerjcBXG5HrY3y+dAWqdY+UpuX0x3g=
Message-ID: <61d2e9b20808140612p6c017172s60fa47f902e33b9d@mail.gmail.com>
Date: Thu, 14 Aug 2008 18:42:48 +0530
From: "Tharindu Mathew" <mccloud35@gmail.com>
To: dev@harmony.apache.org
Subject: Re: possible bug in com.ibm.icu.text.Bidi
In-Reply-To: <3b3f27c60808131935ob250dc7oa736010a84a6173@mail.gmail.com>
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_Part_58436_12606793.1218719568462"
References: <61d2e9b20808120646v5b2688d6tda4b23f481990365@mail.gmail.com>
	 <3b3f27c60808121730t30e11fa3x4a8cad9dc9c55ca6@mail.gmail.com>
	 <61d2e9b20808130346h6d63583asea68f895d6a56de2@mail.gmail.com>
	 <3b3f27c60808131935ob250dc7oa736010a84a6173@mail.gmail.com>

------=_Part_58436_12606793.1218719568462
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Thanks Nathan. I think the default value not being set is the issue. The
test doesn't set any default direction so this should be the problem. Anyway
I put in a work around to solve this problem for now. This checks for mixed
and RighttoLeft for RighttoLeft languages. This thread will be helpful if
anyone else sees this same problem.

Regards,

Tharindu


On Thu, Aug 14, 2008 at 8:05 AM, Nathan Beyer <ndbeyer@apache.org> wrote:

> I did this with ICU 4.0
>
>    com.ibm.icu.text.Bidi b = new
> com.ibm.icu.text.Bidi("\u05DC\u05DD",
> com.ibm.icu.text.Bidi.DIRECTION_DEFAULT_RIGHT_TO_LEFT);
>    System.err.println("Text - " + b.getTextAsString());
>    System.err.println("RTL - " + b.isRightToLeft());
>    System.err.println("LTF - " + b.isLeftToRight());
>    System.err.println("Mixed - " + b.isMixed());
>
>
> It printed -
>
> RTL - true
> LTF - false
> Mixed - false
>
> What flag values are being passed? I know these are RTL code points,
> but I think the algorithm needs to be told what the assumed direction
> is. If you use either the '_DEFAULT_' values you should get the
> correct thing because these are strong characters. If you set to RTL
> you're saying it's RTL. If you set to LTR (the value 0), the it's
> assumed to be LTR with some RTL characters, thus the mixed.
>
> I don't have a lot of experience with BIDI, so bear with me on this.
>
> -Nathan
>
> On Wed, Aug 13, 2008 at 5:46 AM, Tharindu Mathew <mccloud35@gmail.com>
> wrote:
> > is RighttoLeft returns true, isLefttoRight return false.
> >
> > The text used was  - public static final String RTL = "\u05DC\u05DD";
> >
> >
> > Regards,
> >
> > Tharindu
> >
> >
> > On Wed, Aug 13, 2008 at 6:00 AM, Nathan Beyer <ndbeyer@apache.org>
> wrote:
> >
> >> Can you post the example in the post?
> >>
> >> What about the 'isLeftToRight' and 'isRightToLeft' methods? What are
> they
> >> returning?
> >>
> >> On Tue, Aug 12, 2008 at 8:46 AM, Tharindu Mathew <mccloud35@gmail.com
> >> >wrote:
> >>
> >> > I was testing the the failure in Right to Left text updates in
> >> > PlainViewI18N_LineView_UpdateTest.
> >> >
> >> > I used both Hebrew and Arabic as the right to left text and I have
> >> > discovered that bidi.isMixed[1] returns true. This should return false
> >> for
> >> > this.
> >> >
> >> > Can this be verified and reported to them?
> >> >
> >> > [1]
> >> >
> >> > public boolean isMixed() {
> >> >        return icuBidi.isMixed();
> >> >    }
> >> >
> >> > Regards,
> >> >
> >> > Tharindu
> >> >
> >>
> >
>

------=_Part_58436_12606793.1218719568462--