libreoffice/include/vcl/vcllayout.hxx

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

147 lines
5.6 KiB
C++
Raw Permalink Normal View History

/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
/*
* This file is part of the LibreOffice project.
*
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/.
*
* This file incorporates work covered by the following license notice:
*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed
* with this work for additional information regarding copyright
* ownership. The ASF licenses this file to you under the Apache
* License, Version 2.0 (the "License"); you may not use this file
* except in compliance with the License. You may obtain a copy of
* the License at http://www.apache.org/licenses/LICENSE-2.0 .
*/
#pragma once
#include <basegfx/point/b2dpoint.hxx>
#include <basegfx/polygon/b2dpolypolygon.hxx>
#include <basegfx/range/b2drectangle.hxx>
tdf#30731: Improve caret travelling in Writer Previously, when measuring caret position, Writer would measure the width of the substring before the caret (i.e. layout it independent of the text after the caret and measure its width). This is incorrect, though. It assumes cutting the string laying it out would result in the same width as when laid out as part of a bigger string, which is invalid assumption when e.g. cutting inside a ligature or between letters that have different shapes when next to each other, etc. This appears to work when the width of the substring laid out alone is close enough to its width when laid out with the full text. But in cases where is widths are largely different, like the extreme case in the bug report, the caret will be jumping around as it is positioned based on the unligated glyphs not the ligated, rendered glyphs. This change introduces a special mode of measuring text width for caret positioning, that will layout the whole string that return the width of the requested substring. Fields and small caps text are trickier to handle, so old behaviour is retained for them. Now one will probably notice but if they do, it can be dealt with then. This also tries to be conservative and keep other pleases using the existing behaviour which might be desirable (e.g. when measuring text width for line breaking, we want the unligated width), but there might be other places that should use the new behaviour. To handle caret inside ligatures, the grapheme clusters in the ligature are counted and the width of the whole ligature is distributed on them evenly. A further improvement would be using HarfBuzz API to get ligature caret positions for fonts that provide them, which helps when the ligature components have different widths. Change-Id: I02062e2e2e1b1a35c8f84307c0a8f5d743059ab5 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/138889 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
2022-08-26 22:20:55 +02:00
#include <i18nlangtag/languagetag.hxx>
#include <tools/gen.hxx>
#include <tools/degree.hxx>
#include <vcl/glyphitem.hxx>
#include <vcl/dllapi.h>
class LogicalFontInstance;
namespace vcl::text { class ImplLayoutArgs; }
namespace basegfx { class BColor; }
class SalGraphics;
class GlyphItem;
// all positions/widths are in font units
// one exception: drawposition is in pixel units
// Unfortunately there is little documentation to help implementors of
// new classes derived from SalLayout ("layout engines"), and the code
// and data structures are far from obvious.
// For instance, I *think* the important virtual functions in the
// layout engines are called in this order:
// * LayoutText()
// * AdjustLayout(), any number of times (but presumably
// usually not at all or just once)
// * Optionally, DrawText()
// Functions that just return information like GetTexWidth() and
// FillDXArray() are called after LayoutText() and before DrawText().
// Another important questions is which parts of an ImplLayoutArgs can
// be changed by callers between LayoutText() and AdjustLayout()
// calls. It probably makes sense only if one assumes that the "string
// related inputs" part are not changed after LayoutText().
// But why use the same ImplLayoutArgs structure as parameter for both
// LayoutText() and AdjustLayout() in the first place? And why
// duplicate some of the fields in both SalLayout and ImplLayoutArgs
// (mnMinCharPos, mnEndCharPos, mnLayoutFlags==mnFlags,
// mnOrientation)? Lost in history...
class VCL_DLLPUBLIC SalLayout
{
public:
virtual ~SalLayout();
// used by upper layers
basegfx::B2DPoint& DrawBase() { return maDrawBase; }
const basegfx::B2DPoint& DrawBase() const { return maDrawBase; }
basegfx::B2DPoint& DrawOffset() { return maDrawOffset; }
const basegfx::B2DPoint& DrawOffset() const { return maDrawOffset; }
basegfx::B2DPoint GetDrawPosition( const basegfx::B2DPoint& rRelative = basegfx::B2DPoint(0,0) ) const;
virtual bool LayoutText( vcl::text::ImplLayoutArgs&, const SalLayoutGlyphsImpl* ) = 0; // first step of layouting
virtual void AdjustLayout( vcl::text::ImplLayoutArgs& ); // adjusting after fallback etc.
virtual void DrawText( SalGraphics& ) const = 0;
Degree10 GetOrientation() const { return mnOrientation; }
void SetSubpixelPositioning(bool bSubpixelPositioning)
{
mbSubpixelPositioning = bSubpixelPositioning;
}
bool GetSubpixelPositioning() const
{
return mbSubpixelPositioning;
}
// methods using string indexing
virtual sal_Int32 GetTextBreak(double nMaxWidth, double nCharExtra, int nFactor) const = 0;
virtual double FillDXArray( std::vector<double>* pDXArray, const OUString& rStr ) const = 0;
virtual double GetTextWidth() const { return FillDXArray( nullptr, {} ); }
virtual double FillPartialDXArray(std::vector<double>* pDXArray, const OUString& rStr,
sal_Int32 skipStart, sal_Int32 amt) const
= 0;
virtual double GetPartialTextWidth(sal_Int32 skipStart, sal_Int32 amt) const
{
return FillPartialDXArray(nullptr, {}, skipStart, amt);
}
virtual void GetCaretPositions( std::vector<double>& rCaretPositions, const OUString& rStr ) const = 0;
virtual bool HasFontKashidaPositions() const = 0;
virtual bool IsKashidaPosValid ( int /*nCharPos*/, int /*nNextCharPos*/ ) const = 0; // i60594
// methods using glyph indexing
virtual bool GetNextGlyph(const GlyphItem** pGlyph, basegfx::B2DPoint& rPos, int& nStart,
const LogicalFontInstance** ppGlyphFont = nullptr) const = 0;
virtual bool GetOutline(basegfx::B2DPolyPolygonVector&) const;
bool GetBoundRect(basegfx::B2DRectangle&) const;
static tools::Rectangle BoundRect2Rectangle(const basegfx::B2DRectangle&);
virtual SalLayoutGlyphs GetGlyphs() const;
virtual void drawSalLayout(void* /*pSurface*/, const basegfx::BColor& /*rTextColor*/, bool /*bAntiAliased*/) const {}
protected:
// used by layout engines
SalLayout();
private:
SalLayout(const SalLayout&) = delete;
SalLayout& operator=(const SalLayout&) = delete;
protected:
int mnMinCharPos;
int mnEndCharPos;
tdf#30731: Improve caret travelling in Writer Previously, when measuring caret position, Writer would measure the width of the substring before the caret (i.e. layout it independent of the text after the caret and measure its width). This is incorrect, though. It assumes cutting the string laying it out would result in the same width as when laid out as part of a bigger string, which is invalid assumption when e.g. cutting inside a ligature or between letters that have different shapes when next to each other, etc. This appears to work when the width of the substring laid out alone is close enough to its width when laid out with the full text. But in cases where is widths are largely different, like the extreme case in the bug report, the caret will be jumping around as it is positioned based on the unligated glyphs not the ligated, rendered glyphs. This change introduces a special mode of measuring text width for caret positioning, that will layout the whole string that return the width of the requested substring. Fields and small caps text are trickier to handle, so old behaviour is retained for them. Now one will probably notice but if they do, it can be dealt with then. This also tries to be conservative and keep other pleases using the existing behaviour which might be desirable (e.g. when measuring text width for line breaking, we want the unligated width), but there might be other places that should use the new behaviour. To handle caret inside ligatures, the grapheme clusters in the ligature are counted and the width of the whole ligature is distributed on them evenly. A further improvement would be using HarfBuzz API to get ligature caret positions for fonts that provide them, which helps when the ligature components have different widths. Change-Id: I02062e2e2e1b1a35c8f84307c0a8f5d743059ab5 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/138889 Tested-by: Jenkins Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
2022-08-26 22:20:55 +02:00
LanguageTag maLanguageTag;
Degree10 mnOrientation;
basegfx::B2DPoint maDrawOffset;
basegfx::B2DPoint maDrawBase;
bool mbSubpixelPositioning;
};
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */