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

Re: New wiki page for QT bugs



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thursday 03 June 2004 02:49, Munzir Taha wrote:

> > > 1. In a QLineEdit: brackets, braces, parenthesis,
> > > greater and less than signs, ... are always reversed. Please check
> > > yourself against the latest snapshots and give feedback.
> >
> > Yes. I do confirm and this is not specific to KDE, it's QT's. However,
> > now that I think about it, I'm not sure this is a bug anymore. I mean an
> > 'opening bracket' is supposed to stay 'opening' even if you change from
> > LTR to RTL, no ? This is the same for the '>' sign and Co. For example,
> > in LTR, you write '2 > 1', if you reverse the direction, you should read
> > '1 < 2'.
> >
> > What do others think ?
>
> Is it LEFT PARENTHESIS or an open parenthesis?
> I think you will be right if Unicode naming is wrong. So it's a bug
> somewhere.

In my memory it is the latter (opening bracket), but Unicode says both ;)

   http://www.unicode.org/charts/PDF/U0000.pdf

(page 3 of that document refers to a discussion on semantics of paired 
bracketing characters which I could not find).

> > > 2. Another point: switching the language to Arabic should change the
> > > shape of the cursor to point to the left (something like `| not |` ).
> > > Now, this happen when one begins typing in Arabic the correct behavior
> > > should be upon switching the language not begining typing with the
> > > language, isn't it?
> >
> > Not sure. When there's no text in a QLineEdit, or only Latin chars, the
> > cursor has no shape (it's a '|' actually). The cursor changes shape only
> > when there are Arabic chars (mixed with Latin or not) and its direction
> > notes the QLineEdit's alignment. This looks very logical to me (well,
> > maybe the cursor should be shapeless too when there are only Arabic chars
> > in the QLineEdit's alignment is RTL).
> >
> > What do others think ?
>
> Don't you like to know which lang is going to be typed from the cursor
> shape before begining to type and discover you have to switch the keyboard
> layout?

How can the application know that ? (to me that will create some 
complications). I think of the following:

- - Change the cursor shape only for texts that contain languages that have 
different directions (when there are only Arabic chars, the cursors should be 
shapeless: '|').

or

- - Make the cursor's shape '|`' even for LTR-only scripts.

> > > 3. The dancing effect(tm): This happens for example when I select an
> > > Arabic text in a message I received in kmail or in konquror. Of course
> > > you need to describe it in a better way. I well remember someone
> > > (nadim?) has said something about this somewhere (putty?) before.
> >
> > I don't understand what you mean. You need to explain it a little.
>
> I am not sure if I can explain it better. Go to any Arabic site like
> aljazeera.net. Select any line of text with your mouse. Keep pressing the
> mouse and unselect it. I mean move your hand as if you are brushing your
> teeth in an unheathy way from left to right and to left again. You will see
> the dancing text. Sigh!

No, I don't confirm... Moving the mouse while selecting the texts works the 
same as for Latin chars for me (tested in konqueror, kmail, (mozilla :) etc 
etc).

Update: After installing MDK 10.0 and thus KDE 3.2 I'm observing this now in 
konqueror and kmail (only with 'static' texts, ie not those that you can 
edit). Seems indeed like a KDE bug to me, to be confirmed as a QT one (seems 
like it wasn't there before).

Anyway, for now it's only you and me discussing these issues and it seems that 
we have different opinions/expriments on the subject. I think we need others' 
opinions (especially the Unicode/Bidi experts on the first two issues ;)

Salam,

- -- 
Youcef R. Rahal
Arabeyes.org
http://www.arabeyes.org/~rahal
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)

iD8DBQFAws98HDRR6Cd0eSYRAtw6AKD8WsnUlcle48zZyjiaLt8YwAjohwCfcFPv
ux9Fh0VMGlz1ik+vHzQI0h0=
=4cFL
-----END PGP SIGNATURE-----