You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
1477 lines
45 KiB
1477 lines
45 KiB
/****************************************************************************
|
|
** $Id: KoComplexText.cpp 504496 2006-02-01 11:16:49Z dfaure $
|
|
**
|
|
** Implementation of some internal classes
|
|
**
|
|
** Created :
|
|
**
|
|
** Copyright (C) 2001 Trolltech AS. All rights reserved.
|
|
**
|
|
** This file is part of the kernel module of the TQt GUI Toolkit.
|
|
**
|
|
** This file may be distributed under the terms of the Q Public License
|
|
** as defined by Trolltech AS of Norway and appearing in the file
|
|
** LICENSE.TQPL included in the packaging of this file.
|
|
**
|
|
** This file may be distributed and/or modified under the terms of the
|
|
** GNU General Public License version 2 as published by the Free Software
|
|
** Foundation and appearing in the file LICENSE.GPL included in the
|
|
** packaging of this file.
|
|
**
|
|
** Licensees holding valid TQt Enterprise Edition or TQt Professional Edition
|
|
** licenses may use this file in accordance with the TQt Commercial License
|
|
** Agreement provided with the Software.
|
|
**
|
|
** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE
|
|
** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
|
|
**
|
|
** See http://www.trolltech.com/pricing.html or email sales@trolltech.com for
|
|
** information about TQt Commercial License Agreements.
|
|
** See http://www.trolltech.com/qpl/ for TQPL licensing information.
|
|
** See http://www.trolltech.com/gpl/ for GPL licensing information.
|
|
**
|
|
** Contact info@trolltech.com if any conditions of this licensing are
|
|
** not clear to you.
|
|
**
|
|
**********************************************************************/
|
|
|
|
|
|
#ifndef TQT_NO_COMPLEXTEXT
|
|
#include "KoRichText.h"
|
|
//#include "tqfontdata_p.h"
|
|
#include "tqfontmetrics.h"
|
|
#include "tqrect.h"
|
|
|
|
#include <stdlib.h>
|
|
|
|
// -----------------------------------------------------
|
|
|
|
/* a small helper class used internally to resolve Bidi embedding levels.
|
|
Each line of text caches the embedding level at the start of the line for faster
|
|
relayouting
|
|
*/
|
|
KoBidiContext::KoBidiContext( uchar l, TQChar::Direction e, KoBidiContext *p, bool o )
|
|
: level(l) , override(o), dir(e)
|
|
{
|
|
if ( p )
|
|
p->ref();
|
|
parent = p;
|
|
count = 0;
|
|
}
|
|
|
|
KoBidiContext::~KoBidiContext()
|
|
{
|
|
if( parent && parent->deref() )
|
|
delete parent;
|
|
}
|
|
|
|
static TQChar *tqshapeBuffer = 0;
|
|
static int tqshapeBufSize = 0;
|
|
|
|
/*
|
|
Arabic shaping obeys a number of rules according to the joining classes (see Unicode book, section on
|
|
arabic).
|
|
|
|
Each tqunicode char has a joining class (right, dual (left&right), center (joincausing) or transparent).
|
|
transparent joining is not encoded in TQChar::joining(), but applies to all combining marks and format marks.
|
|
|
|
Right join-causing: dual + center
|
|
Left join-causing: dual + right + center
|
|
|
|
Rules are as follows (for a string already in visual order, as we have it here):
|
|
|
|
R1 Transparent characters do not affect joining behaviour.
|
|
R2 A right joining character, that has a right join-causing char on the right will get form XRight
|
|
(R3 A left joining character, that has a left join-causing char on the left will get form XLeft)
|
|
Note: the above rule is meaningless, as there are no pure left joining characters defined in Unicode
|
|
R4 A dual joining character, that has a left join-causing char on the left and a right join-causing char on
|
|
the right will get form XMedial
|
|
R5 A dual joining character, that has a right join causing char on the right, and no left join causing char on the left
|
|
will get form XRight
|
|
R6 A dual joining character, that has a left join causing char on the left, and no right join causing char on the right
|
|
will get form XLeft
|
|
R7 Otherwise the character will get form XIsolated
|
|
|
|
Additionally we have to do the minimal ligature support for lam-alef ligatures:
|
|
|
|
L1 Transparent characters do not affect ligature behaviour.
|
|
L2 Any sequence of Alef(XRight) + Lam(XMedial) will form the ligature Alef.Lam(XLeft)
|
|
L3 Any sequence of Alef(XRight) + Lam(XLeft) will form the ligature Alef.Lam(XIsolated)
|
|
|
|
The two functions defined in this class do shaping in visual and logical order. For logical order just replace right with
|
|
previous and left with next in the above rules ;-)
|
|
*/
|
|
|
|
/*
|
|
Two small helper functions for arabic shaping. They get the next tqshape causing character on either
|
|
side of the char in question. Implements rule R1.
|
|
|
|
leftChar() returns true if the char to the left is a left join-causing char
|
|
rightChar() returns true if the char to the right is a right join-causing char
|
|
*/
|
|
static inline const TQChar *prevChar( const TQString &str, int pos )
|
|
{
|
|
//kdDebug() << "leftChar: pos=" << pos << endl;
|
|
pos--;
|
|
const TQChar *ch = str.tqunicode() + pos;
|
|
while( pos > -1 ) {
|
|
if( !ch->isMark() )
|
|
return ch;
|
|
pos--;
|
|
ch--;
|
|
}
|
|
return &TQChar::replacement;
|
|
}
|
|
|
|
static inline const TQChar *nextChar( const TQString &str, int pos)
|
|
{
|
|
pos++;
|
|
int len = str.length();
|
|
const TQChar *ch = str.tqunicode() + pos;
|
|
while( pos < len ) {
|
|
//kdDebug() << "rightChar: " << pos << " isLetter=" << ch.isLetter() << ", joining=" << ch.joining() << endl;
|
|
if( !ch->isMark() )
|
|
return ch;
|
|
// assume it's a transparent char, this might not be 100% correct
|
|
pos++;
|
|
ch++;
|
|
}
|
|
return &TQChar::replacement;
|
|
}
|
|
|
|
static inline bool prevVisualCharJoins( const TQString &str, int pos)
|
|
{
|
|
return ( prevChar( str, pos )->joining() != TQChar::OtherJoining );
|
|
}
|
|
|
|
static inline bool nextVisualCharJoins( const TQString &str, int pos)
|
|
{
|
|
TQChar::Joining join = nextChar( str, pos )->joining();
|
|
return ( join == TQChar::Dual || join == TQChar::Center );
|
|
}
|
|
|
|
|
|
KoComplexText::Shape KoComplexText::glyphVariant( const TQString &str, int pos)
|
|
{
|
|
// ignores L1 - L3, done in the codec
|
|
TQChar::Joining joining = str[pos].joining();
|
|
//kdDebug() << "checking " << str[pos].tqunicode() << ", joining=" << joining << endl;
|
|
switch ( joining ) {
|
|
case TQChar::OtherJoining:
|
|
case TQChar::Center:
|
|
// these don't change tqshape
|
|
return XIsolated;
|
|
case TQChar::Right:
|
|
// only rule R2 applies
|
|
if( nextVisualCharJoins( str, pos ) )
|
|
return XFinal;
|
|
return XIsolated;
|
|
case TQChar::Dual:
|
|
bool right = nextVisualCharJoins( str, pos );
|
|
bool left = prevVisualCharJoins( str, pos );
|
|
//kdDebug() << "dual: right=" << right << ", left=" << left << endl;
|
|
if( right && left )
|
|
return XMedial;
|
|
else if ( right )
|
|
return XFinal;
|
|
else if ( left )
|
|
return XInitial;
|
|
else
|
|
return XIsolated;
|
|
}
|
|
return XIsolated;
|
|
}
|
|
|
|
/* and the same thing for logical ordering :)
|
|
*/
|
|
static inline bool prevLogicalCharJoins( const TQString &str, int pos)
|
|
{
|
|
return ( nextChar( str, pos )->joining() != TQChar::OtherJoining );
|
|
}
|
|
|
|
static inline bool nextLogicalCharJoins( const TQString &str, int pos)
|
|
{
|
|
TQChar::Joining join = prevChar( str, pos )->joining();
|
|
return ( join == TQChar::Dual || join == TQChar::Center );
|
|
}
|
|
|
|
|
|
KoComplexText::Shape KoComplexText::glyphVariantLogical( const TQString &str, int pos)
|
|
{
|
|
// ignores L1 - L3, ligatures are job of the codec
|
|
TQChar::Joining joining = str[pos].joining();
|
|
//kdDebug() << "checking " << str[pos].tqunicode() << ", joining=" << joining << endl;
|
|
switch ( joining ) {
|
|
case TQChar::OtherJoining:
|
|
case TQChar::Center:
|
|
// these don't change tqshape
|
|
return XIsolated;
|
|
case TQChar::Right:
|
|
// only rule R2 applies
|
|
if( nextLogicalCharJoins( str, pos ) )
|
|
return XFinal;
|
|
return XIsolated;
|
|
case TQChar::Dual:
|
|
bool right = nextLogicalCharJoins( str, pos );
|
|
bool left = prevLogicalCharJoins( str, pos );
|
|
//kdDebug() << "dual: right=" << right << ", left=" << left << endl;
|
|
if( right && left )
|
|
return XMedial;
|
|
else if ( right )
|
|
return XFinal;
|
|
else if ( left )
|
|
return XInitial;
|
|
else
|
|
return XIsolated;
|
|
}
|
|
return XIsolated;
|
|
}
|
|
|
|
// -------------------------------------------------------------
|
|
|
|
// The tqunicode to tqunicode shaping codec.
|
|
// does only presentation forms B at the moment, but that should be enough for
|
|
// simple display
|
|
static const ushort arabicUnicodeMapping[256][2] = {
|
|
// base of tqshaped forms, and number-1 of them ( 0 for non shaping,
|
|
// 1 for right binding and 3 for dual binding
|
|
{ 0x0600, 0 }, // 0x600
|
|
{ 0x0601, 0 }, // 0x601
|
|
{ 0x0602, 0 }, // 0x602
|
|
{ 0x0603, 0 }, // 0x603
|
|
{ 0x0604, 0 }, // 0x604
|
|
{ 0x0605, 0 }, // 0x605
|
|
{ 0x0606, 0 }, // 0x606
|
|
{ 0x0607, 0 }, // 0x607
|
|
{ 0x0608, 0 }, // 0x608
|
|
{ 0x0609, 0 }, // 0x609
|
|
{ 0x060a, 0 }, // 0x60a
|
|
{ 0x060b, 0 }, // 0x60b
|
|
{ 0x060c, 0 }, // 0x60c Arabic comma
|
|
{ 0x060d, 0 }, // 0x60d
|
|
{ 0x060e, 0 }, // 0x60e
|
|
{ 0x060f, 0 }, // 0x60f
|
|
|
|
{ 0x0610, 0 }, // 0x610
|
|
{ 0x0611, 0 }, // 0x611
|
|
{ 0x0612, 0 }, // 0x612
|
|
{ 0x0613, 0 }, // 0x613
|
|
{ 0x0614, 0 }, // 0x614
|
|
{ 0x0615, 0 }, // 0x615
|
|
{ 0x0616, 0 }, // 0x616
|
|
{ 0x0617, 0 }, // 0x617
|
|
{ 0x0618, 0 }, // 0x618
|
|
{ 0x0619, 0 }, // 0x619
|
|
{ 0x061a, 0 }, // 0x61a
|
|
{ 0x061b, 0 }, // 0x61b Arabic semicolon
|
|
{ 0x061c, 0 }, // 0x61c
|
|
{ 0x061d, 0 }, // 0x61d
|
|
{ 0x061e, 0 }, // 0x61e
|
|
{ 0x061f, 0 }, // 0x61f Arabic question mark
|
|
|
|
{ 0x0620, 0 }, // 0x620
|
|
{ 0xfe80, 0 }, // 0x621 Hamza
|
|
{ 0xfe81, 1 }, // 0x622 R Alef with Madda above
|
|
{ 0xfe83, 1 }, // 0x623 R Alef with Hamza above
|
|
{ 0xfe85, 1 }, // 0x624 R Waw with Hamza above
|
|
{ 0xfe87, 1 }, // 0x625 R Alef with Hamza below
|
|
{ 0xfe89, 3 }, // 0x626 D Yeh with Hamza above
|
|
{ 0xfe8d, 1 }, // 0x627 R Alef
|
|
{ 0xfe8f, 3 }, // 0x628 D Beh
|
|
{ 0xfe93, 1 }, // 0x629 R Teh Marbuta
|
|
{ 0xfe95, 3 }, // 0x62a D Teh
|
|
{ 0xfe99, 3 }, // 0x62b D Theh
|
|
{ 0xfe9d, 3 }, // 0x62c D Jeem
|
|
{ 0xfea1, 3 }, // 0x62d D Hah
|
|
{ 0xfea5, 3 }, // 0x62e D Khah
|
|
{ 0xfea9, 1 }, // 0x62f R Dal
|
|
|
|
{ 0xfeab, 1 }, // 0x630 R Thal
|
|
{ 0xfead, 1 }, // 0x631 R Reh
|
|
{ 0xfeaf, 1 }, // 0x632 R Zain
|
|
{ 0xfeb1, 1 }, // 0x633 D Seen
|
|
{ 0xfeb5, 3 }, // 0x634 D Sheen
|
|
{ 0xfeb9, 3 }, // 0x635 D Sad
|
|
{ 0xfebd, 3 }, // 0x636 D Dad
|
|
{ 0xfec1, 3 }, // 0x637 D Tah
|
|
{ 0xfec5, 3 }, // 0x638 D Zah
|
|
{ 0xfec9, 3 }, // 0x639 D Ain
|
|
{ 0xfecd, 3 }, // 0x63a D Ghain
|
|
{ 0x063b, 0 }, // 0x63b
|
|
{ 0x063c, 0 }, // 0x63c
|
|
{ 0x063d, 0 }, // 0x63d
|
|
{ 0x063e, 0 }, // 0x63e
|
|
{ 0x063f, 0 }, // 0x63f
|
|
|
|
{ 0x0640, 0 }, // 0x640 C Tatweel
|
|
{ 0xfed1, 3 }, // 0x641 D Feh
|
|
{ 0xfed5, 3 }, // 0x642 D Qaf
|
|
{ 0xfed9, 3 }, // 0x643 D Kaf
|
|
{ 0xfedd, 3 }, // 0x644 D Lam
|
|
{ 0xfee1, 3 }, // 0x645 D Meem
|
|
{ 0xfee5, 3 }, // 0x646 D Noon
|
|
{ 0xfee9, 3 }, // 0x647 D Heh
|
|
{ 0xfeed, 1 }, // 0x648 R Waw
|
|
{ 0xfeef, 1 }, // 0x649 R Alef Maksura // ### Dual according to newest arabicshaping.txt
|
|
{ 0xfef1, 3 }, // 0x64a D Yeh
|
|
{ 0x064b, 0 }, // 0x64b Mark Fathatan
|
|
{ 0x064c, 0 }, // 0x64c Mark Dammatan
|
|
{ 0x064d, 0 }, // 0x64d Mark Kasratan
|
|
{ 0x064e, 0 }, // 0x64e Mark Fatha
|
|
{ 0x064f, 0 }, // 0x64f Mark Damma
|
|
|
|
{ 0x0650, 0 }, // 0x650 Mark Kasra
|
|
{ 0x0651, 0 }, // 0x651 Mark Shadda
|
|
{ 0x0652, 0 }, // 0x652 Mark Sukan
|
|
// these do not exist in latin6 anymore:
|
|
{ 0x0653, 0 }, // 0x653 Mark Maddah above
|
|
{ 0x0654, 0 }, // 0x654 Mark Hamza above
|
|
{ 0x0655, 0 }, // 0x655 Mark Hamza below
|
|
{ 0x0656, 0 }, // 0x656
|
|
{ 0x0657, 0 }, // 0x657
|
|
{ 0x0658, 0 }, // 0x658
|
|
{ 0x0659, 0 }, // 0x659
|
|
{ 0x065a, 0 }, // 0x65a
|
|
{ 0x065b, 0 }, // 0x65b
|
|
{ 0x065c, 0 }, // 0x65c
|
|
{ 0x065d, 0 }, // 0x65d
|
|
{ 0x065e, 0 }, // 0x65e
|
|
{ 0x065f, 0 }, // 0x65f
|
|
|
|
{ 0x0660, 0 }, // 0x660 Arabic 0
|
|
{ 0x0661, 0 }, // 0x661 Arabic 1
|
|
{ 0x0662, 0 }, // 0x662 Arabic 2
|
|
{ 0x0663, 0 }, // 0x663 Arabic 3
|
|
{ 0x0664, 0 }, // 0x664 Arabic 4
|
|
{ 0x0665, 0 }, // 0x665 Arabic 5
|
|
{ 0x0666, 0 }, // 0x666 Arabic 6
|
|
{ 0x0667, 0 }, // 0x667 Arabic 7
|
|
{ 0x0668, 0 }, // 0x668 Arabic 8
|
|
{ 0x0669, 0 }, // 0x669 Arabic 9
|
|
{ 0x066a, 0 }, // 0x66a Arabic % sign
|
|
{ 0x066b, 0 }, // 0x66b Arabic decimal separator
|
|
{ 0x066c, 0 }, // 0x66c Arabic thousands separator
|
|
{ 0x066d, 0 }, // 0x66d Arabic five pointed star
|
|
{ 0x066e, 0 }, // 0x66e
|
|
{ 0x066f, 0 }, // 0x66f
|
|
|
|
// ### some glyphs do not have tqshaped mappings in the presentation forms A.
|
|
// these have the shaping set to 0 for the moment. Will have to find out better mappings for them.
|
|
{ 0x0670, 0 }, // 0x670
|
|
{ 0xfb50, 1 }, // 0x671 R Alef Wasla
|
|
{ 0x0672, 0 }, // 0x672 R Alef with wavy Hamza above
|
|
{ 0x0673, 0 }, // 0x673 R Alef with wavy Hamza below
|
|
{ 0x0674, 0 }, // 0x674 U High Hamza
|
|
{ 0x0675, 0 }, // 0x675 R High Hamza Alef
|
|
{ 0x0676, 0 }, // 0x676 R High Hamza Wav
|
|
{ 0xfbdd, 0 }, // 0x677 R U with hamza above // ### only isolated form found...
|
|
{ 0x0678, 0 }, // 0x678 D High hamza yeh
|
|
{ 0xfb66, 3 }, // 0x679 D ttheh
|
|
{ 0xfb5e, 3 }, // 0x67a D theheh
|
|
{ 0xfb52, 3 }, // 0x67b D beeh
|
|
{ 0x067c, 0 }, // 0x67cD teh with ring
|
|
{ 0x067d, 0 }, // 0x67d D teh with three dots above downwards
|
|
{ 0xfb56, 3 }, // 0x67e D peh
|
|
{ 0xfb62, 3 }, // 0x67f D teheh
|
|
|
|
{ 0xfb5a, 3 }, // 0x680 D beheh
|
|
{ 0x0681, 0 }, // 0x681 D hah with hamza above
|
|
{ 0x0682, 0 }, // 0x682 D hah with two dots vertical above
|
|
{ 0xfb76, 3 }, // 0x683 D nyeh
|
|
{ 0xfb72, 3 }, // 0x684 D dyeh
|
|
{ 0x0685, 0 }, // 0x685 D hah with three dots above
|
|
{ 0xfb7a, 3 }, // 0x686 D tcheh
|
|
{ 0xfb7e, 3 }, // 0x687 D tcheheh
|
|
{ 0xfb88, 1 }, // 0x688 R ddal
|
|
{ 0x0689, 0 }, // 0x689 R dal with ring
|
|
{ 0x068a, 0 }, // 0x68a R dal with dot
|
|
{ 0x068b, 0 }, // 0x68b R dal with dot below and small tah
|
|
{ 0xfb84, 1 }, // 0x68cR dahal
|
|
{ 0xfb82, 1 }, // 0x68d R ddahal
|
|
{ 0xfb86, 1 }, // 0x68e R dul
|
|
{ 0x068f, 0 }, // 0x68f R dal with three dots above downwards
|
|
|
|
{ 0x0690, 0 }, // 0x690 R dal with four dots above
|
|
{ 0xfb8c, 1 }, // 0x691 R rreh
|
|
{ 0x0692, 0 }, // 0x692 R reh with small v
|
|
{ 0x0693, 0 }, // 0x693 R reh with ring
|
|
{ 0x0694, 0 }, // 0x694 R reh with dot below
|
|
{ 0x0695, 0 }, // 0x695 R reh with small v below
|
|
{ 0x0696, 0 }, // 0x696 R reh with dot below and dot above
|
|
{ 0x0697, 0 }, // 0x697 R reh with two dots above
|
|
{ 0xfb8a, 1 }, // 0x698 R jeh
|
|
{ 0x0699, 0 }, // 0x699 R reh with four dots above
|
|
{ 0x069a, 0 }, // 0x69a D seen with dot below and dot above
|
|
{ 0x069b, 0 }, // 0x69b D seen with three dots below
|
|
{ 0x069c, 0 }, // 0x69cD seen with three dots below and three dots above
|
|
{ 0x069d, 0 }, // 0x69d D sad with two dots below
|
|
{ 0x069e, 0 }, // 0x69e D sad with three dots above
|
|
{ 0x069f, 0 }, // 0x69f D tah with three dots above
|
|
|
|
{ 0x06a0, 0 }, // 0x6a0 D ain with three dots above
|
|
{ 0x06a1, 0 }, // 0x6a1 D dotless feh
|
|
{ 0x06a2, 0 }, // 0x6a2 D feh with dot moved below
|
|
{ 0x06a3, 0 }, // 0x6a3 D feh with dot below
|
|
{ 0xfb6a, 3 }, // 0x6a4 D veh
|
|
{ 0x06a5, 0 }, // 0x6a5 D feh with three dots below
|
|
{ 0xfb6e, 3 }, // 0x6a6 D peheh
|
|
{ 0x06a7, 0 }, // 0x6a7 D qaf with dot above
|
|
{ 0x06a8, 0 }, // 0x6a8 D qaf woith three dots above
|
|
{ 0xfb8e, 3 }, // 0x6a9 D keheh
|
|
{ 0x06aa, 0 }, // 0x6aa D swash kaf
|
|
{ 0x06ab, 0 }, // 0x6ab D kaf with ring
|
|
{ 0x06ac, 0 }, // 0x6acD kaf with dot above
|
|
{ 0xfbd3, 3 }, // 0x6ad D ng
|
|
{ 0x06ae, 0 }, // 0x6ae D kaf with three dots below
|
|
{ 0xfb92, 3 }, // 0x6af D gaf
|
|
|
|
{ 0x06b0, 0 }, // 0x6b0 D gaf with ring
|
|
{ 0xfb9a, 3 }, // 0x6b1 D ngoeh
|
|
{ 0x06b2, 0 }, // 0x6b2 D gaf with two dots below
|
|
{ 0xfb96, 3 }, // 0x6b3 D gueh
|
|
{ 0x06b4, 0 }, // 0x6b4 D gaf with three dots above
|
|
{ 0x06b5, 0 }, // 0x6b5 D lam with small v
|
|
{ 0x06b6, 0 }, // 0x6b6 D lam with dot above
|
|
{ 0x06b7, 0 }, // 0x6b7 D lam with three dots above
|
|
{ 0x06b8, 0 }, // 0x6b8 D lam with three dots below
|
|
{ 0x06b9, 0 }, // 0x6b9 D noon with dot below
|
|
{ 0xfb9e, 1 }, // 0x6ba R noon ghunna
|
|
{ 0xfba0, 3 }, // 0x6bb D rnoon
|
|
{ 0x06bc, 0 }, // 0x6bcD noon with ring
|
|
{ 0x06bd, 0 }, // 0x6bd D noon with three dots above
|
|
{ 0xfbaa, 3 }, // 0x6be D heh doachashmee
|
|
{ 0x06bf, 0 }, // 0x6bf D tcheh with dot above
|
|
|
|
{ 0xfba4, 1 }, // 0x6c0 R heh with yeh above = ligature hamza on hah (06d5 + 0654)
|
|
{ 0xfba6, 3 }, // 0x6c1 D heh goal
|
|
{ 0x06c2, 0 }, // 0x6c2 R heh goal with hamza above (06c1 + 0654)
|
|
{ 0x06c3, 0 }, // 0x6c3 R teh marbuta goal
|
|
{ 0x06c4, 0 }, // 0x6c4 R waw with ring
|
|
{ 0xfbe0, 1 }, // 0x6c5 R kirghiz oe
|
|
{ 0xfbd9, 1 }, // 0x6c6 R oe
|
|
{ 0xfbd7, 1 }, // 0x6c7 R u
|
|
{ 0xfbdb, 1 }, // 0x6c8 R yu
|
|
{ 0xfbe2, 1 }, // 0x6c9 R kirghiz yu
|
|
{ 0x06ca, 0 }, // 0x6ca R waw with teo dots above
|
|
{ 0xfbde, 1 }, // 0x6cb R ve
|
|
{ 0x06cc, 0 }, // 0x6cc D farsi yeh
|
|
{ 0x06cd, 0 }, // 0x6cd R yeh with tail
|
|
{ 0x06ce, 0 }, // 0x6ce D yeh with small v
|
|
{ 0x06cf, 0 }, // 0x6cf R waw with dot above
|
|
|
|
{ 0xfbe4, 3 }, // 0x6d0 D e
|
|
{ 0x06d1, 0 }, // 0x6d1 D yeh with three dots below
|
|
{ 0xfbae, 1 }, // 0x6d2 R yeh barree
|
|
{ 0xfbb0, 1 }, // 0x6d3 R yeh barree with hamza above
|
|
{ 0x06d4, 0 }, // 0x6d4 U full stop
|
|
{ 0x06d5, 0 }, // 0x6d5 D ae
|
|
{ 0x06d6, 0 }, // 0x6d6 koreanic annotaion signs
|
|
{ 0x06d7, 0 }, // 0x6d7 ...
|
|
{ 0x06d8, 0 }, // 0x6d8
|
|
{ 0x06d9, 0 }, // 0x6d9
|
|
{ 0x06da, 0 }, // 0x6da
|
|
{ 0x06db, 0 }, // 0x6db
|
|
{ 0x06dc, 0 }, // 0x6dc
|
|
{ 0x06dd, 0 }, // 0x6dd
|
|
{ 0x06de, 0 }, // 0x6de
|
|
{ 0x06df, 0 }, // 0x6df
|
|
|
|
{ 0x06e0, 0 }, // 0x6e0
|
|
{ 0x06e1, 0 }, // 0x6e1
|
|
{ 0x06e2, 0 }, // 0x6e2
|
|
{ 0x06e3, 0 }, // 0x6e3
|
|
{ 0x06e4, 0 }, // 0x6e4
|
|
{ 0x06e5, 0 }, // 0x6e5
|
|
{ 0x06e6, 0 }, // 0x6e6
|
|
{ 0x06e7, 0 }, // 0x6e7
|
|
{ 0x06e8, 0 }, // 0x6e8
|
|
{ 0x06e9, 0 }, // 0x6e9
|
|
{ 0x06ea, 0 }, // 0x6ea
|
|
{ 0x06eb, 0 }, // 0x6eb
|
|
{ 0x06ec, 0 }, // 0x6ec
|
|
{ 0x06ed, 0 }, // 0x6ed
|
|
{ 0x06ee, 0 }, // 0x6ee
|
|
{ 0x06ef, 0 }, // 0x6ef
|
|
|
|
{ 0x06f0, 0 }, // 0x6f0 Arabic indic digit 0
|
|
{ 0x06f1, 0 }, // 0x6f1
|
|
{ 0x06f2, 0 }, // 0x6f2
|
|
{ 0x06f3, 0 }, // 0x6f3
|
|
{ 0x06f4, 0 }, // 0x6f4
|
|
{ 0x06f5, 0 }, // 0x6f5
|
|
{ 0x06f6, 0 }, // 0x6f6
|
|
{ 0x06f7, 0 }, // 0x6f7
|
|
{ 0x06f8, 0 }, // 0x6f8
|
|
{ 0x06f9, 0 }, // 0x6f9 Arabic indic digit 9
|
|
{ 0x06fa, 0 }, // 0x6fa D Sheen with dot below
|
|
{ 0x06fb, 0 }, // 0x6fb D dad with dot below
|
|
{ 0x06fc, 0 }, // 0x6fc D ghain with dot below
|
|
{ 0x06fd, 0 }, // 0x6fd Sindhi ampersand
|
|
{ 0x06fe, 0 }, // 0x6fe sindhi postposition
|
|
{ 0x06ff, 0 }, // 0x6ff
|
|
|
|
};
|
|
|
|
// this is a bit tricky. Alef always binds to the right, so the second parameter descibing the tqshape
|
|
// of the lam can be either initial of medial. So initial maps to the isolated form of the ligature,
|
|
// medial to the final form
|
|
static const ushort arabicUnicodeLamAlefMapping[6][4] = {
|
|
{ 0xfffd, 0xfffd, 0xfef5, 0xfef6 }, // 0x622 R Alef with Madda above
|
|
{ 0xfffd, 0xfffd, 0xfef7, 0xfef8 }, // 0x623 R Alef with Hamza above
|
|
{ 0xfffd, 0xfffd, 0xfffd, 0xfffd }, // 0x624 R Waw with Hamza above
|
|
{ 0xfffd, 0xfffd, 0xfef9, 0xfefa }, // 0x625 R Alef with Hamza below
|
|
{ 0xfffd, 0xfffd, 0xfffd, 0xfffd }, // 0x626 D Yeh with Hamza above
|
|
{ 0xfffd, 0xfffd, 0xfefb, 0xfefc } // 0x627 R Alef
|
|
};
|
|
|
|
static inline int getShape( const TQChar * /* base */, uchar cell, int tqshape,
|
|
const TQFontMetrics * /* fm */ )
|
|
{
|
|
uint ch = arabicUnicodeMapping[cell][0] + tqshape;
|
|
/*
|
|
// we revert to the untqshaped glyph in case the tqshaped version doesn't exist
|
|
if ( fm && !fm->inFont( ch ) ) {
|
|
switch( tqshape ) {
|
|
case KoComplexText::XIsolated:
|
|
break; // try base form
|
|
case KoComplexText::XFinal:
|
|
ch -= 1; // try isolated form
|
|
break;
|
|
case KoComplexText::XInitial:
|
|
ch += 1; // try medial form
|
|
break;
|
|
case KoComplexText::XMedial:
|
|
ch -= 1; // try initial form
|
|
break;
|
|
}
|
|
if ( !fm->inFont( ch ) )
|
|
ch = *base;
|
|
}
|
|
*/
|
|
return ch;
|
|
}
|
|
|
|
TQString KoComplexText::tqshapedString(const TQString& uc, int from, int len, TQPainter::TextDirection dir, const TQFontMetrics *fm )
|
|
{
|
|
if( len < 0 )
|
|
len = uc.length() - from;
|
|
if( len == 0 ) {
|
|
return TQString();
|
|
}
|
|
|
|
// we have to ignore NSMs at the beginning and add at the end.
|
|
int num = uc.length() - from - len;
|
|
const TQChar *ch = uc.tqunicode() + from + len;
|
|
while ( num > 0 && ch->combiningClass() != 0 ) {
|
|
ch++;
|
|
num--;
|
|
len++;
|
|
}
|
|
ch = uc.tqunicode() + from;
|
|
while ( len > 0 && ch->combiningClass() != 0 ) {
|
|
ch++;
|
|
len--;
|
|
from++;
|
|
}
|
|
if ( len == 0 ) return TQString();
|
|
|
|
if( !tqshapeBuffer || len > tqshapeBufSize ) {
|
|
if( tqshapeBuffer ) free( (void *) tqshapeBuffer );
|
|
tqshapeBuffer = (TQChar *) malloc( len*sizeof( TQChar ) );
|
|
// delete [] tqshapeBuffer;
|
|
// tqshapeBuffer = new TQChar[ len + 1];
|
|
tqshapeBufSize = len;
|
|
}
|
|
|
|
int lenOut = 0;
|
|
TQChar *data = tqshapeBuffer;
|
|
if ( dir == TQPainter::RTL )
|
|
ch += len - 1;
|
|
for ( int i = 0; i < len; i++ ) {
|
|
uchar r = ch->row();
|
|
uchar c = ch->cell();
|
|
if ( r != 0x06 ) {
|
|
if ( dir == TQPainter::RTL && ch->mirrored() )
|
|
*data = ch->mirroredChar();
|
|
else
|
|
*data = *ch;
|
|
data++;
|
|
lenOut++;
|
|
} else {
|
|
int pos = i + from;
|
|
if ( dir == TQPainter::RTL )
|
|
pos = from + len - 1 - i;
|
|
int tqshape = glyphVariantLogical( uc, pos );
|
|
//kdDebug() << "mapping U+" << ch->tqunicode() << " to tqshape " << tqshape << " glyph=0x" << arabicUnicodeMapping[ch->cell()][tqshape] << endl;
|
|
// take care of lam-alef ligatures (lam right of alef)
|
|
ushort map;
|
|
switch ( c ) {
|
|
case 0x44: { // lam
|
|
const TQChar *pch = nextChar( uc, pos );
|
|
if ( pch->row() == 0x06 ) {
|
|
switch ( pch->cell() ) {
|
|
case 0x22:
|
|
case 0x23:
|
|
case 0x25:
|
|
case 0x27:
|
|
//kdDebug() << " lam of lam-alef ligature" << endl;
|
|
map = arabicUnicodeLamAlefMapping[pch->cell() - 0x22][tqshape];
|
|
goto next;
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
break;
|
|
}
|
|
case 0x22: // alef with madda
|
|
case 0x23: // alef with hamza above
|
|
case 0x25: // alef with hamza below
|
|
case 0x27: // alef
|
|
if ( prevChar( uc, pos )->tqunicode() == 0x0644 ) {
|
|
// have a lam alef ligature
|
|
//kdDebug() << " alef of lam-alef ligature" << endl;
|
|
goto skip;
|
|
}
|
|
default:
|
|
break;
|
|
}
|
|
map = getShape( ch, c, tqshape, fm );
|
|
next:
|
|
*data = map;
|
|
data++;
|
|
lenOut++;
|
|
}
|
|
skip:
|
|
if ( dir == TQPainter::RTL )
|
|
ch--;
|
|
else
|
|
ch++;
|
|
}
|
|
|
|
if ( dir == TQPainter::Auto && !uc.simpleText() ) {
|
|
return bidiReorderString( TQConstString( tqshapeBuffer, lenOut ).string() );
|
|
}
|
|
if ( dir == TQPainter::RTL ) {
|
|
// reverses the non spacing marks to be again after the base char
|
|
TQChar *s = tqshapeBuffer;
|
|
int i = 0;
|
|
while ( i < lenOut ) {
|
|
if ( s->combiningClass() != 0 ) {
|
|
// non spacing marks
|
|
int clen = 1;
|
|
TQChar *ch = s;
|
|
do {
|
|
ch++;
|
|
clen++;
|
|
} while ( ch->combiningClass() != 0 );
|
|
|
|
int j = 0;
|
|
TQChar *cp = s;
|
|
while ( j < clen/2 ) {
|
|
TQChar tmp = *cp;
|
|
*cp = *ch;
|
|
*ch = tmp;
|
|
cp++;
|
|
ch--;
|
|
j++;
|
|
}
|
|
s += clen;
|
|
i += clen;
|
|
} else {
|
|
s++;
|
|
i++;
|
|
}
|
|
}
|
|
}
|
|
|
|
return TQConstString( tqshapeBuffer, lenOut ).string();
|
|
}
|
|
|
|
TQChar KoComplexText::tqshapedCharacter( const TQString &str, int pos, const TQFontMetrics *fm )
|
|
{
|
|
const TQChar *ch = str.tqunicode() + pos;
|
|
if ( ch->row() != 0x06 )
|
|
return *ch;
|
|
else {
|
|
int tqshape = glyphVariantLogical( str, pos );
|
|
//kdDebug() << "mapping U+" << ch->tqunicode() << " to tqshape " << tqshape << " glyph=0x" << arabicUnicodeMapping[ch->cell()][tqshape] << endl;
|
|
// lam aleph ligatures
|
|
switch ( ch->cell() ) {
|
|
case 0x44: { // lam
|
|
const TQChar *nch = nextChar( str, pos );
|
|
if ( nch->row() == 0x06 ) {
|
|
switch ( nch->cell() ) {
|
|
case 0x22:
|
|
case 0x23:
|
|
case 0x25:
|
|
case 0x27:
|
|
return TQChar(arabicUnicodeLamAlefMapping[nch->cell() - 0x22][tqshape]);
|
|
default:
|
|
break;
|
|
}
|
|
}
|
|
break;
|
|
}
|
|
case 0x22: // alef with madda
|
|
case 0x23: // alef with hamza above
|
|
case 0x25: // alef with hamza below
|
|
case 0x27: // alef
|
|
if ( prevChar( str, pos )->tqunicode() == 0x0644 )
|
|
// have a lam alef ligature
|
|
return TQChar(0);
|
|
default:
|
|
break;
|
|
}
|
|
return TQChar( getShape( ch, ch->cell(), tqshape, fm ) );
|
|
}
|
|
}
|
|
|
|
// Avoid using TQFontPrivate, to which we don't have access. We don't use positionMarks() anyway
|
|
#if 0
|
|
TQPointArray KoComplexText::positionMarks( TQFontPrivate *f, const TQString &str,
|
|
int pos, TQRect *boundingRect )
|
|
{
|
|
int len = str.length();
|
|
int nmarks = 0;
|
|
while ( pos + nmarks < len && str[pos+nmarks +1].combiningClass() > 0 )
|
|
nmarks++;
|
|
|
|
if ( !nmarks )
|
|
return TQPointArray();
|
|
|
|
TQChar baseChar = KoComplexText::tqshapedCharacter( str, pos );
|
|
TQRect baseRect = f->boundingRect( baseChar );
|
|
int baseOffset = f->textWidth( str, pos, 1 );
|
|
|
|
//kdDebug() << "base char: bounding rect at " << baseRect.x() << "/" << baseRect.y() << " (" << baseRect.width() << "/" << baseRect.height() << ")" << endl;
|
|
int offset = f->actual.pixelSize / 10 + 1;
|
|
//kdDebug() << "offset = " << offset << endl;
|
|
TQPointArray pa( nmarks );
|
|
int i;
|
|
unsigned char lastCmb = 0;
|
|
TQRect attachmentRect;
|
|
if ( boundingRect )
|
|
*boundingRect = baseRect;
|
|
for( i = 0; i < nmarks; i++ ) {
|
|
TQChar mark = str[pos+i+1];
|
|
unsigned char cmb = mark.combiningClass();
|
|
if ( cmb < 200 ) {
|
|
// fixed position classes. We approximate by mapping to one of the others.
|
|
// currently I added only the ones for arabic, hebrew and thai.
|
|
|
|
// ### add a bit more offset to arabic, a bit hacky
|
|
if ( cmb >= 27 && cmb <= 36 )
|
|
offset +=1;
|
|
// below
|
|
if ( (cmb >= 10 && cmb <= 18) ||
|
|
cmb == 20 || cmb == 22 ||
|
|
cmb == 29 || cmb == 32 )
|
|
cmb = TQChar::Combining_Below;
|
|
// above
|
|
else if ( cmb == 23 || cmb == 27 || cmb == 28 ||
|
|
cmb == 30 || cmb == 31 || (cmb >= 33 && cmb <= 36 ) )
|
|
cmb = TQChar::Combining_Above;
|
|
//below-right
|
|
else if ( cmb == 103 )
|
|
cmb = TQChar::Combining_BelowRight;
|
|
// above-right
|
|
else if ( cmb == 24 || cmb == 107 )
|
|
cmb = TQChar::Combining_AboveRight;
|
|
else if ( cmb == 25 )
|
|
cmb = TQChar::Combining_AboveLeft;
|
|
// fixed:
|
|
// 19 21
|
|
|
|
}
|
|
|
|
// combining marks of different class don't interact. Reset the rectangle.
|
|
if ( cmb != lastCmb ) {
|
|
//kdDebug() << "resetting rect" << endl;
|
|
attachmentRect = baseRect;
|
|
}
|
|
|
|
TQPoint p;
|
|
TQRect markRect = f->boundingRect( mark );
|
|
switch( cmb ) {
|
|
case TQChar::Combining_DoubleBelow:
|
|
// ### wrong in rtl context!
|
|
case TQChar::Combining_BelowLeft:
|
|
p += TQPoint( 0, offset );
|
|
case TQChar::Combining_BelowLeftAttached:
|
|
p += attachmentRect.bottomLeft() - markRect.topLeft();
|
|
break;
|
|
case TQChar::Combining_Below:
|
|
p += TQPoint( 0, offset );
|
|
case TQChar::Combining_BelowAttached:
|
|
p += attachmentRect.bottomLeft() - markRect.topLeft();
|
|
p += TQPoint( (attachmentRect.width() - markRect.width())/2 , 0 );
|
|
break;
|
|
case TQChar::Combining_BelowRight:
|
|
p += TQPoint( 0, offset );
|
|
case TQChar::Combining_BelowRightAttached:
|
|
p += attachmentRect.bottomRight() - markRect.topRight();
|
|
break;
|
|
case TQChar::Combining_Left:
|
|
p += TQPoint( -offset, 0 );
|
|
case TQChar::Combining_LeftAttached:
|
|
break;
|
|
case TQChar::Combining_Right:
|
|
p += TQPoint( offset, 0 );
|
|
case TQChar::Combining_RightAttached:
|
|
break;
|
|
case TQChar::Combining_DoubleAbove:
|
|
// ### wrong in RTL context!
|
|
case TQChar::Combining_AboveLeft:
|
|
p += TQPoint( 0, -offset );
|
|
case TQChar::Combining_AboveLeftAttached:
|
|
p += attachmentRect.topLeft() - markRect.bottomLeft();
|
|
break;
|
|
case TQChar::Combining_Above:
|
|
p += TQPoint( 0, -offset );
|
|
case TQChar::Combining_AboveAttached:
|
|
p += attachmentRect.topLeft() - markRect.bottomLeft();
|
|
p += TQPoint( (attachmentRect.width() - markRect.width())/2 , 0 );
|
|
break;
|
|
case TQChar::Combining_AboveRight:
|
|
p += TQPoint( 0, -offset );
|
|
case TQChar::Combining_AboveRightAttached:
|
|
p += attachmentRect.topRight() - markRect.bottomRight();
|
|
break;
|
|
|
|
case TQChar::Combining_IotaSubscript:
|
|
default:
|
|
break;
|
|
}
|
|
//kdDebug() << "char=" << mark.tqunicode() << " combiningClass = " << cmb << " offset=" << p.x() << "/" << p.y() << endl;
|
|
markRect.moveBy( p.x(), p.y() );
|
|
p += TQPoint( -baseOffset, 0 );
|
|
attachmentRect |= markRect;
|
|
if ( boundingRect )
|
|
*boundingRect |= markRect;
|
|
lastCmb = cmb;
|
|
pa.setPoint( i, p );
|
|
}
|
|
return pa;
|
|
}
|
|
#endif
|
|
|
|
//#define BIDI_DEBUG
|
|
#ifdef BIDI_DEBUG
|
|
#include <iostream>
|
|
#endif
|
|
|
|
static TQChar::Direction basicDirection(const TQString &str, int start = 0)
|
|
{
|
|
int len = str.length();
|
|
int pos = start > len ? len -1 : start;
|
|
const TQChar *uc = str.tqunicode() + pos;
|
|
while( pos < len ) {
|
|
switch( uc->direction() )
|
|
{
|
|
case TQChar::DirL:
|
|
case TQChar::DirLRO:
|
|
case TQChar::DirLRE:
|
|
return TQChar::DirL;
|
|
case TQChar::DirR:
|
|
case TQChar::DirAL:
|
|
case TQChar::DirRLO:
|
|
case TQChar::DirRLE:
|
|
return TQChar::DirR;
|
|
default:
|
|
break;
|
|
}
|
|
++pos;
|
|
++uc;
|
|
}
|
|
if ( start != 0 )
|
|
return basicDirection( str );
|
|
return TQChar::DirL;
|
|
}
|
|
|
|
// transforms one line of the paragraph to visual order
|
|
// the caller is responisble to delete the returned list of KoTextRuns.
|
|
TQPtrList<KoTextRun> *KoComplexText::bidiReorderLine( KoBidiControl *control, const TQString &text, int start, int len,
|
|
TQChar::Direction basicDir )
|
|
{
|
|
int last = start + len - 1;
|
|
//printf("doing BiDi reordering from %d to %d!\n", start, last);
|
|
|
|
TQPtrList<KoTextRun> *runs = new TQPtrList<KoTextRun>;
|
|
runs->setAutoDelete(TRUE);
|
|
|
|
KoBidiContext *context = control->context;
|
|
if ( !context ) {
|
|
// first line
|
|
//if( start != 0 )
|
|
// kdDebug() << "bidiReorderLine::internal error" << endl;
|
|
if( basicDir == TQChar::DirR || (basicDir == TQChar::DirON && text.isRightToLeft() ) ) {
|
|
context = new KoBidiContext( 1, TQChar::DirR );
|
|
control->status.last = TQChar::DirR;
|
|
} else {
|
|
context = new KoBidiContext( 0, TQChar::DirL );
|
|
control->status.last = TQChar::DirL;
|
|
}
|
|
}
|
|
|
|
KoBidiStatus status = control->status;
|
|
TQChar::Direction dir = TQChar::DirON;
|
|
|
|
int sor = start;
|
|
int eor = start;
|
|
|
|
int current = start;
|
|
while(current <= last) {
|
|
TQChar::Direction dirCurrent;
|
|
if(current == (int)text.length()) {
|
|
KoBidiContext *c = context;
|
|
while ( c->parent )
|
|
c = c->parent;
|
|
dirCurrent = c->dir;
|
|
} else if ( current == last ) {
|
|
dirCurrent = ( basicDir != TQChar::DirON ? basicDir : basicDirection( text, current ) );
|
|
} else
|
|
dirCurrent = text.at(current).direction();
|
|
|
|
|
|
#ifdef BIDI_DEBUG
|
|
cout << "directions: dir=" << dir << " current=" << dirCurrent << " last=" << status.last << " eor=" << status.eor << " lastStrong=" << status.lastStrong << " embedding=" << context->dir << " level =" << (int)context->level << endl;
|
|
#endif
|
|
|
|
switch(dirCurrent) {
|
|
|
|
// embedding and overrides (X1-X9 in the BiDi specs)
|
|
case TQChar::DirRLE:
|
|
{
|
|
uchar level = context->level;
|
|
if(level%2) // we have an odd level
|
|
level += 2;
|
|
else
|
|
level++;
|
|
if(level < 61) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
context = new KoBidiContext(level, TQChar::DirR, context);
|
|
status.last = TQChar::DirR;
|
|
status.lastStrong = TQChar::DirR;
|
|
}
|
|
break;
|
|
}
|
|
case TQChar::DirLRE:
|
|
{
|
|
uchar level = context->level;
|
|
if(level%2) // we have an odd level
|
|
level++;
|
|
else
|
|
level += 2;
|
|
if(level < 61) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
context = new KoBidiContext(level, TQChar::DirL, context);
|
|
status.last = TQChar::DirL;
|
|
status.lastStrong = TQChar::DirL;
|
|
}
|
|
break;
|
|
}
|
|
case TQChar::DirRLO:
|
|
{
|
|
uchar level = context->level;
|
|
if(level%2) // we have an odd level
|
|
level += 2;
|
|
else
|
|
level++;
|
|
if(level < 61) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
context = new KoBidiContext(level, TQChar::DirR, context, TRUE);
|
|
dir = TQChar::DirR;
|
|
status.last = TQChar::DirR;
|
|
status.lastStrong = TQChar::DirR;
|
|
}
|
|
break;
|
|
}
|
|
case TQChar::DirLRO:
|
|
{
|
|
uchar level = context->level;
|
|
if(level%2) // we have an odd level
|
|
level++;
|
|
else
|
|
level += 2;
|
|
if(level < 61) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
context = new KoBidiContext(level, TQChar::DirL, context, TRUE);
|
|
dir = TQChar::DirL;
|
|
status.last = TQChar::DirL;
|
|
status.lastStrong = TQChar::DirL;
|
|
}
|
|
break;
|
|
}
|
|
case TQChar::DirPDF:
|
|
{
|
|
KoBidiContext *c = context->parent;
|
|
if(c) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
status.last = context->dir;
|
|
if( context->deref() ) delete context;
|
|
context = c;
|
|
if(context->override)
|
|
dir = context->dir;
|
|
else
|
|
dir = TQChar::DirON;
|
|
status.lastStrong = context->dir;
|
|
}
|
|
break;
|
|
}
|
|
|
|
// strong types
|
|
case TQChar::DirL:
|
|
if(dir == TQChar::DirON)
|
|
dir = TQChar::DirL;
|
|
switch(status.last)
|
|
{
|
|
case TQChar::DirL:
|
|
eor = current; status.eor = TQChar::DirL; break;
|
|
case TQChar::DirR:
|
|
case TQChar::DirAL:
|
|
case TQChar::DirEN:
|
|
case TQChar::DirAN:
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
break;
|
|
case TQChar::DirES:
|
|
case TQChar::DirET:
|
|
case TQChar::DirCS:
|
|
case TQChar::DirBN:
|
|
case TQChar::DirB:
|
|
case TQChar::DirS:
|
|
case TQChar::DirWS:
|
|
case TQChar::DirON:
|
|
if(dir != TQChar::DirL) {
|
|
//last stuff takes embedding dir
|
|
if( context->dir == TQChar::DirR ) {
|
|
if(status.eor != TQChar::DirR) {
|
|
// AN or EN
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirR;
|
|
}
|
|
else
|
|
eor = current - 1;
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
} else {
|
|
if(status.eor != TQChar::DirL) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirL;
|
|
} else {
|
|
eor = current; status.eor = TQChar::DirL; break;
|
|
}
|
|
}
|
|
} else {
|
|
eor = current; status.eor = TQChar::DirL;
|
|
}
|
|
default:
|
|
break;
|
|
}
|
|
status.lastStrong = TQChar::DirL;
|
|
break;
|
|
case TQChar::DirAL:
|
|
case TQChar::DirR:
|
|
if(dir == TQChar::DirON) dir = TQChar::DirR;
|
|
switch(status.last)
|
|
{
|
|
case TQChar::DirR:
|
|
case TQChar::DirAL:
|
|
eor = current; status.eor = TQChar::DirR; break;
|
|
case TQChar::DirL:
|
|
case TQChar::DirEN:
|
|
case TQChar::DirAN:
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
break;
|
|
case TQChar::DirES:
|
|
case TQChar::DirET:
|
|
case TQChar::DirCS:
|
|
case TQChar::DirBN:
|
|
case TQChar::DirB:
|
|
case TQChar::DirS:
|
|
case TQChar::DirWS:
|
|
case TQChar::DirON:
|
|
if( status.eor != TQChar::DirR && status.eor != TQChar::DirAL ) {
|
|
//last stuff takes embedding dir
|
|
if(context->dir == TQChar::DirR || status.lastStrong == TQChar::DirR) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirR;
|
|
eor = current;
|
|
} else {
|
|
eor = current - 1;
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirR;
|
|
}
|
|
} else {
|
|
eor = current; status.eor = TQChar::DirR;
|
|
}
|
|
default:
|
|
break;
|
|
}
|
|
status.lastStrong = dirCurrent;
|
|
break;
|
|
|
|
// weak types:
|
|
|
|
case TQChar::DirNSM:
|
|
// ### if @sor, set dir to dirSor
|
|
break;
|
|
case TQChar::DirEN:
|
|
if(status.lastStrong != TQChar::DirAL) {
|
|
// if last strong was AL change EN to AL
|
|
if(dir == TQChar::DirON) {
|
|
if(status.lastStrong == TQChar::DirL)
|
|
dir = TQChar::DirL;
|
|
else
|
|
dir = TQChar::DirAN;
|
|
}
|
|
switch(status.last)
|
|
{
|
|
case TQChar::DirET:
|
|
if ( status.lastStrong == TQChar::DirR || status.lastStrong == TQChar::DirAL ) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirAN;
|
|
}
|
|
// fall through
|
|
case TQChar::DirEN:
|
|
case TQChar::DirL:
|
|
eor = current;
|
|
status.eor = dirCurrent;
|
|
break;
|
|
case TQChar::DirR:
|
|
case TQChar::DirAL:
|
|
case TQChar::DirAN:
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; status.eor = TQChar::DirEN;
|
|
dir = TQChar::DirAN; break;
|
|
case TQChar::DirES:
|
|
case TQChar::DirCS:
|
|
if(status.eor == TQChar::DirEN || dir == TQChar::DirAN) {
|
|
eor = current; break;
|
|
}
|
|
case TQChar::DirBN:
|
|
case TQChar::DirB:
|
|
case TQChar::DirS:
|
|
case TQChar::DirWS:
|
|
case TQChar::DirON:
|
|
if(status.eor == TQChar::DirR) {
|
|
// neutrals go to R
|
|
eor = current - 1;
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirEN;
|
|
dir = TQChar::DirAN;
|
|
}
|
|
else if( status.eor == TQChar::DirL ||
|
|
(status.eor == TQChar::DirEN && status.lastStrong == TQChar::DirL)) {
|
|
eor = current; status.eor = dirCurrent;
|
|
} else {
|
|
// numbers on both sides, neutrals get right to left direction
|
|
if(dir != TQChar::DirL) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
eor = current - 1;
|
|
dir = TQChar::DirR;
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirAN;
|
|
} else {
|
|
eor = current; status.eor = dirCurrent;
|
|
}
|
|
}
|
|
default:
|
|
break;
|
|
}
|
|
break;
|
|
}
|
|
case TQChar::DirAN:
|
|
dirCurrent = TQChar::DirAN;
|
|
if(dir == TQChar::DirON) dir = TQChar::DirAN;
|
|
switch(status.last)
|
|
{
|
|
case TQChar::DirL:
|
|
case TQChar::DirAN:
|
|
eor = current; status.eor = TQChar::DirAN; break;
|
|
case TQChar::DirR:
|
|
case TQChar::DirAL:
|
|
case TQChar::DirEN:
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
break;
|
|
case TQChar::DirCS:
|
|
if(status.eor == TQChar::DirAN) {
|
|
eor = current; status.eor = TQChar::DirR; break;
|
|
}
|
|
case TQChar::DirES:
|
|
case TQChar::DirET:
|
|
case TQChar::DirBN:
|
|
case TQChar::DirB:
|
|
case TQChar::DirS:
|
|
case TQChar::DirWS:
|
|
case TQChar::DirON:
|
|
if(status.eor == TQChar::DirR) {
|
|
// neutrals go to R
|
|
eor = current - 1;
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirAN;
|
|
} else if( status.eor == TQChar::DirL ||
|
|
(status.eor == TQChar::DirEN && status.lastStrong == TQChar::DirL)) {
|
|
eor = current; status.eor = dirCurrent;
|
|
} else {
|
|
// numbers on both sides, neutrals get right to left direction
|
|
if(dir != TQChar::DirL) {
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
eor = current - 1;
|
|
dir = TQChar::DirR;
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
++eor; sor = eor; dir = TQChar::DirON; status.eor = TQChar::DirON;
|
|
dir = TQChar::DirAN;
|
|
} else {
|
|
eor = current; status.eor = dirCurrent;
|
|
}
|
|
}
|
|
default:
|
|
break;
|
|
}
|
|
break;
|
|
case TQChar::DirES:
|
|
case TQChar::DirCS:
|
|
break;
|
|
case TQChar::DirET:
|
|
if(status.last == TQChar::DirEN) {
|
|
dirCurrent = TQChar::DirEN;
|
|
eor = current; status.eor = dirCurrent;
|
|
break;
|
|
}
|
|
break;
|
|
|
|
// boundary neutrals should be ignored
|
|
case TQChar::DirBN:
|
|
break;
|
|
// neutrals
|
|
case TQChar::DirB:
|
|
// ### what do we do with newline and paragraph separators that come to here?
|
|
break;
|
|
case TQChar::DirS:
|
|
// ### implement rule L1
|
|
break;
|
|
case TQChar::DirWS:
|
|
case TQChar::DirON:
|
|
break;
|
|
default:
|
|
break;
|
|
}
|
|
|
|
//cout << " after: dir=" << // dir << " current=" << dirCurrent << " last=" << status.last << " eor=" << status.eor << " lastStrong=" << status.lastStrong << " embedding=" << context->dir << endl;
|
|
|
|
if(current >= (int)text.length()) break;
|
|
|
|
// set status.last as needed.
|
|
switch(dirCurrent)
|
|
{
|
|
case TQChar::DirET:
|
|
case TQChar::DirES:
|
|
case TQChar::DirCS:
|
|
case TQChar::DirS:
|
|
case TQChar::DirWS:
|
|
case TQChar::DirON:
|
|
switch(status.last)
|
|
{
|
|
case TQChar::DirL:
|
|
case TQChar::DirR:
|
|
case TQChar::DirAL:
|
|
case TQChar::DirEN:
|
|
case TQChar::DirAN:
|
|
status.last = dirCurrent;
|
|
break;
|
|
default:
|
|
status.last = TQChar::DirON;
|
|
}
|
|
break;
|
|
case TQChar::DirNSM:
|
|
case TQChar::DirBN:
|
|
// ignore these
|
|
break;
|
|
default:
|
|
status.last = dirCurrent;
|
|
}
|
|
|
|
++current;
|
|
}
|
|
|
|
#ifdef BIDI_DEBUG
|
|
cout << "reached end of line current=" << current << ", eor=" << eor << endl;
|
|
#endif
|
|
eor = current - 1; // remove dummy char
|
|
|
|
if ( sor <= eor )
|
|
runs->append( new KoTextRun(sor, eor, context, dir) );
|
|
|
|
// reorder line according to run structure...
|
|
|
|
// first find highest and lowest levels
|
|
uchar levelLow = 128;
|
|
uchar levelHigh = 0;
|
|
KoTextRun *r = runs->first();
|
|
while ( r ) {
|
|
//printf("level = %d\n", r->level);
|
|
if ( r->level > levelHigh )
|
|
levelHigh = r->level;
|
|
if ( r->level < levelLow )
|
|
levelLow = r->level;
|
|
r = runs->next();
|
|
}
|
|
|
|
// implements reordering of the line (L2 according to BiDi spec):
|
|
// L2. From the highest level found in the text to the lowest odd level on each line,
|
|
// reverse any contiguous sequence of characters that are at that level or higher.
|
|
|
|
// reversing is only done up to the lowest odd level
|
|
if(!(levelLow%2)) levelLow++;
|
|
|
|
#ifdef BIDI_DEBUG
|
|
cout << "reorderLine: lineLow = " << (uint)levelLow << ", lineHigh = " << (uint)levelHigh << endl;
|
|
cout << "logical order is:" << endl;
|
|
TQPtrListIterator<KoTextRun> it2(*runs);
|
|
KoTextRun *r2;
|
|
for ( ; (r2 = it2.current()); ++it2 )
|
|
cout << " " << r2 << " start=" << r2->start << " stop=" << r2->stop << " level=" << (uint)r2->level << endl;
|
|
#endif
|
|
|
|
int count = runs->count() - 1;
|
|
|
|
while(levelHigh >= levelLow)
|
|
{
|
|
int i = 0;
|
|
while ( i < count )
|
|
{
|
|
while(i < count && runs->at(i)->level < levelHigh) i++;
|
|
int start = i;
|
|
while(i <= count && runs->at(i)->level >= levelHigh) i++;
|
|
int end = i-1;
|
|
|
|
if(start != end)
|
|
{
|
|
//cout << "reversing from " << start << " to " << end << endl;
|
|
for(int j = 0; j < (end-start+1)/2; j++)
|
|
{
|
|
KoTextRun *first = runs->take(start+j);
|
|
KoTextRun *last = runs->take(end-j-1);
|
|
runs->insert(start+j, last);
|
|
runs->insert(end-j, first);
|
|
}
|
|
}
|
|
i++;
|
|
if(i >= count) break;
|
|
}
|
|
levelHigh--;
|
|
}
|
|
|
|
#ifdef BIDI_DEBUG
|
|
cout << "visual order is:" << endl;
|
|
TQPtrListIterator<KoTextRun> it3(*runs);
|
|
KoTextRun *r3;
|
|
for ( ; (r3 = it3.current()); ++it3 )
|
|
{
|
|
cout << " " << r3 << endl;
|
|
}
|
|
#endif
|
|
|
|
control->setContext( context );
|
|
control->status = status;
|
|
|
|
return runs;
|
|
}
|
|
|
|
|
|
TQString KoComplexText::bidiReorderString( const TQString &str, TQChar::Direction /*basicDir*/ )
|
|
{
|
|
|
|
// ### fix basic direction
|
|
KoBidiControl control;
|
|
int lineStart = 0;
|
|
int lineEnd = 0;
|
|
int len = str.length();
|
|
TQString visual;
|
|
visual.setUnicode( 0, len );
|
|
TQChar *vch = (TQChar *)visual.tqunicode();
|
|
const TQChar *ch = str.tqunicode();
|
|
while( lineStart < len ) {
|
|
lineEnd = lineStart;
|
|
while( *ch != '\n' && lineEnd < len ) {
|
|
ch++;
|
|
lineEnd++;
|
|
}
|
|
lineEnd++;
|
|
TQPtrList<KoTextRun> *runs = bidiReorderLine( &control, str, lineStart, lineEnd - lineStart );
|
|
|
|
// reorder the content of the line, and output to visual
|
|
KoTextRun *r = runs->first();
|
|
while ( r ) {
|
|
if(r->level %2) {
|
|
// odd level, need to reverse the string
|
|
int pos = r->stop;
|
|
while(pos >= r->start) {
|
|
*vch = str[pos];
|
|
if ( vch->mirrored() )
|
|
*vch = vch->mirroredChar();
|
|
vch++;
|
|
pos--;
|
|
}
|
|
} else {
|
|
int pos = r->start;
|
|
while(pos <= r->stop) {
|
|
*vch = str[pos];
|
|
vch++;
|
|
pos++;
|
|
}
|
|
}
|
|
r = runs->next();
|
|
}
|
|
if ( *ch == '\n' ) {
|
|
*vch = *ch;
|
|
vch++;
|
|
ch++;
|
|
lineEnd++;
|
|
}
|
|
lineStart = lineEnd;
|
|
}
|
|
return visual;
|
|
}
|
|
|
|
KoTextRun::KoTextRun(int _start, int _stop, KoBidiContext *context, TQChar::Direction dir) {
|
|
start = _start;
|
|
stop = _stop;
|
|
if(dir == TQChar::DirON) dir = context->dir;
|
|
|
|
level = context->level;
|
|
|
|
// add level of run (cases I1 & I2)
|
|
if( level % 2 ) {
|
|
if(dir == TQChar::DirL || dir == TQChar::DirAN)
|
|
level++;
|
|
} else {
|
|
if( dir == TQChar::DirR )
|
|
level++;
|
|
else if( dir == TQChar::DirAN )
|
|
level += 2;
|
|
}
|
|
#ifdef BIDI_DEBUG
|
|
printf("new run: dir=%d from %d, to %d level = %d\n", dir, _start, _stop, level);
|
|
#endif
|
|
}
|
|
|
|
#endif //TQT_NO_COMPLEXTEXT
|