Files @ r28630:fe3c5717ad2b
Branch filter:

Location: cpp/openttd-patchpack/source/src/os/windows/string_uniscribe.cpp

Loïc Guilloux
Fix: [Script] Apply random deviation only at script start (#11944)
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
/*
 * This file is part of OpenTTD.
 * OpenTTD is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, version 2.
 * OpenTTD is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 * See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with OpenTTD. If not, see <http://www.gnu.org/licenses/>.
 */

/** @file string_uniscribe.cpp Functions related to laying out text on Win32. */

#include "../../stdafx.h"
#include "../../debug.h"
#include "string_uniscribe.h"
#include "../../language.h"
#include "../../strings_func.h"
#include "../../string_func.h"
#include "../../table/control_codes.h"
#include "../../zoom_func.h"
#include "win32.h"

#include <windows.h>
#include <usp10.h>

#include "../../safeguards.h"

#ifdef _MSC_VER
#	pragma comment(lib, "usp10")
#endif


/** Uniscribe cache for internal font information, cleared when OTTD changes fonts. */
static SCRIPT_CACHE _script_cache[FS_END];

/**
 * Contains all information about a run of characters. A run are consecutive
 * characters that share a single font and language.
 */
struct UniscribeRun {
	int pos;
	int len;
	Font *font;

	std::vector<GlyphID> ft_glyphs;

	SCRIPT_ANALYSIS sa;
	std::vector<WORD> char_to_glyph;

	std::vector<SCRIPT_VISATTR> vis_attribs;
	std::vector<WORD> glyphs;
	std::vector<int> advances;
	std::vector<GOFFSET> offsets;
	int total_advance;

	UniscribeRun(int pos, int len, Font *font, SCRIPT_ANALYSIS &sa) : pos(pos), len(len), font(font), sa(sa) {}
};

/** Break a string into language formatting ranges. */
static std::vector<SCRIPT_ITEM> UniscribeItemizeString(UniscribeParagraphLayoutFactory::CharType *buff, int32_t length);
/** Generate and place glyphs for a run of characters. */
static bool UniscribeShapeRun(const UniscribeParagraphLayoutFactory::CharType *buff, UniscribeRun &range);

/**
 * Wrapper for doing layouts with Uniscribe.
 */
class UniscribeParagraphLayout : public ParagraphLayouter {
private:
	const UniscribeParagraphLayoutFactory::CharType *text_buffer;

	std::vector<UniscribeRun> ranges; ///< All runs of the text.
	std::vector<UniscribeRun>::iterator cur_range; ///< The next run to be output.
	int cur_range_offset = 0; ///< Offset from the start of the current run from where to output.

public:
	/** Visual run contains data about the bit of text with the same font. */
	class UniscribeVisualRun : public ParagraphLayouter::VisualRun {
	private:
		std::vector<GlyphID> glyphs;
		std::vector<Point> positions;
		std::vector<WORD> char_to_glyph;

		int start_pos;
		int total_advance;
		int num_glyphs;
		Font *font;

		mutable std::vector<int> glyph_to_char;

	public:
		UniscribeVisualRun(const UniscribeRun &range, int x);
		UniscribeVisualRun(UniscribeVisualRun &&other) noexcept;

		const std::vector<GlyphID> &GetGlyphs() const override { return this->glyphs; }
		const std::vector<Point> &GetPositions() const override { return this->positions; }
		const std::vector<int> &GetGlyphToCharMap() const override;

		const Font *GetFont() const override { return this->font;  }
		int GetLeading() const override { return this->font->fc->GetHeight(); }
		int GetGlyphCount() const override { return this->num_glyphs; }
		int GetAdvance() const { return this->total_advance; }
	};

	/** A single line worth of VisualRuns. */
	class UniscribeLine : public std::vector<UniscribeVisualRun>, public ParagraphLayouter::Line {
	public:
		int GetLeading() const override;
		int GetWidth() const override;
		int CountRuns() const override { return (uint)this->size();  }
		const VisualRun &GetVisualRun(int run) const override { return this->at(run);  }

		int GetInternalCharLength(char32_t c) const override
		{
			/* Uniscribe uses UTF-16 internally which means we need to account for surrogate pairs. */
			return c >= 0x010000U ? 2 : 1;
		}
	};

	UniscribeParagraphLayout(std::vector<UniscribeRun> &ranges, const UniscribeParagraphLayoutFactory::CharType *buffer) : text_buffer(buffer), ranges(ranges)
	{
		this->Reflow();
	}

	~UniscribeParagraphLayout() override {}

	void Reflow() override
	{
		this->cur_range = this->ranges.begin();
		this->cur_range_offset = 0;
	}

	std::unique_ptr<const Line> NextLine(int max_width) override;
};

void UniscribeResetScriptCache(FontSize size)
{
	if (_script_cache[size] != nullptr) {
		ScriptFreeCache(&_script_cache[size]);
		_script_cache[size] = nullptr;
	}
}

/** Load the matching native Windows font. */
static HFONT HFontFromFont(Font *font)
{
	if (font->fc->GetOSHandle() != nullptr) return CreateFontIndirect(reinterpret_cast<PLOGFONT>(const_cast<void *>(font->fc->GetOSHandle())));

	LOGFONT logfont;
	ZeroMemory(&logfont, sizeof(LOGFONT));
	logfont.lfHeight = font->fc->GetHeight();
	logfont.lfWeight = FW_NORMAL;
	logfont.lfCharSet = DEFAULT_CHARSET;
	convert_to_fs(font->fc->GetFontName(), logfont.lfFaceName, lengthof(logfont.lfFaceName));

	return CreateFontIndirect(&logfont);
}

/** Determine the glyph positions for a run. */
static bool UniscribeShapeRun(const UniscribeParagraphLayoutFactory::CharType *buff, UniscribeRun &range)
{
	/* Initial size guess for the number of glyphs recommended by Uniscribe. */
	range.glyphs.resize(range.len * 3 / 2 + 16);
	range.vis_attribs.resize(range.glyphs.size());

	/* The char-to-glyph array is the same size as the input. */
	range.char_to_glyph.resize(range.len);

	HDC temp_dc = nullptr;
	HFONT old_font = nullptr;
	HFONT cur_font = nullptr;

	while (true) {
		/* Shape the text run by determining the glyphs needed for display. */
		int glyphs_used = 0;
		HRESULT hr = ScriptShape(temp_dc, &_script_cache[range.font->fc->GetSize()], buff + range.pos, range.len, (int)range.glyphs.size(), &range.sa, &range.glyphs[0], &range.char_to_glyph[0], &range.vis_attribs[0], &glyphs_used);

		if (SUCCEEDED(hr)) {
			range.glyphs.resize(glyphs_used);
			range.vis_attribs.resize(glyphs_used);

			/* Calculate the glyph positions. */
			ABC abc;
			range.advances.resize(range.glyphs.size());
			range.offsets.resize(range.glyphs.size());
			hr = ScriptPlace(temp_dc, &_script_cache[range.font->fc->GetSize()], &range.glyphs[0], (int)range.glyphs.size(), &range.vis_attribs[0], &range.sa, &range.advances[0], &range.offsets[0], &abc);
			if (SUCCEEDED(hr)) {
				/* We map our special sprite chars to values that don't fit into a WORD. Copy the glyphs
				 * into a new vector and query the real glyph to use for these special chars. */
				range.ft_glyphs.resize(range.glyphs.size());
				for (size_t g_id = 0; g_id < range.glyphs.size(); g_id++) {
					range.ft_glyphs[g_id] = range.glyphs[g_id];
				}
				for (int i = 0; i < range.len; i++) {
					if (buff[range.pos + i] >= SCC_SPRITE_START && buff[range.pos + i] <= SCC_SPRITE_END) {
						auto pos = range.char_to_glyph[i];
						if (range.ft_glyphs[pos] == 0) { // Font doesn't have our special glyph, so remap.
							range.ft_glyphs[pos] = range.font->fc->MapCharToGlyph(buff[range.pos + i]);
							range.offsets[pos].dv = (range.font->fc->GetHeight() - ScaleSpriteTrad(FontCache::GetDefaultFontHeight(range.font->fc->GetSize()))) / 2; // Align sprite font to centre
							range.advances[pos] = range.font->fc->GetGlyphWidth(range.ft_glyphs[pos]);
						}
					}
				}

				range.total_advance = 0;
				for (size_t i = 0; i < range.advances.size(); i++) {
#ifdef WITH_FREETYPE
					/* FreeType and GDI/Uniscribe seems to occasionally disagree over the width of a glyph. */
					if (range.advances[i] > 0 && range.ft_glyphs[i] != 0xFFFF) range.advances[i] = range.font->fc->GetGlyphWidth(range.ft_glyphs[i]);
#endif
					range.total_advance += range.advances[i];
				}
				break;
			}
		}

		if (hr == E_OUTOFMEMORY) {
			/* The glyph buffer needs to be larger. Just double it every time. */
			range.glyphs.resize(range.glyphs.size() * 2);
			range.vis_attribs.resize(range.vis_attribs.size() * 2);
		} else if (hr == E_PENDING) {
			/* Glyph data is not in cache, load native font. */
			cur_font = HFontFromFont(range.font);
			if (cur_font == nullptr) return false; // Sorry, no dice.

			temp_dc = CreateCompatibleDC(nullptr);
			SetMapMode(temp_dc, MM_TEXT);
			old_font = (HFONT)SelectObject(temp_dc, cur_font);
		} else if (hr == USP_E_SCRIPT_NOT_IN_FONT && range.sa.eScript != SCRIPT_UNDEFINED) {
			/* Try again with the generic shaping engine. */
			range.sa.eScript = SCRIPT_UNDEFINED;
		} else {
			/* Some unknown other error. */
			if (temp_dc != nullptr) {
				SelectObject(temp_dc, old_font);
				DeleteObject(cur_font);
				ReleaseDC(nullptr, temp_dc);
			}
			return false;
		}
	}

	if (temp_dc != nullptr) {
		SelectObject(temp_dc, old_font);
		DeleteObject(cur_font);
		ReleaseDC(nullptr, temp_dc);
	}

	return true;
}

static std::vector<SCRIPT_ITEM> UniscribeItemizeString(UniscribeParagraphLayoutFactory::CharType *buff, int32_t length)
{
	/* Itemize text. */
	SCRIPT_CONTROL control;
	ZeroMemory(&control, sizeof(SCRIPT_CONTROL));
	control.uDefaultLanguage = _current_language->winlangid;

	SCRIPT_STATE state;
	ZeroMemory(&state, sizeof(SCRIPT_STATE));
	state.uBidiLevel = _current_text_dir == TD_RTL ? 1 : 0;

	std::vector<SCRIPT_ITEM> items(16);
	while (true) {
		/* We subtract one from max_items to work around a buffer overflow on some older versions of Windows. */
		int generated = 0;
		HRESULT hr = ScriptItemize(buff, length, (int)items.size() - 1, &control, &state, &items[0], &generated);

		if (SUCCEEDED(hr)) {
			/* Resize the item buffer. Note that Uniscribe will always add an additional end sentinel item. */
			items.resize(generated + 1);
			break;
		}
		/* Some kind of error except item buffer too small. */
		if (hr != E_OUTOFMEMORY) return std::vector<SCRIPT_ITEM>();

		items.resize(items.size() * 2);
	}

	return items;
}

/* static */ ParagraphLayouter *UniscribeParagraphLayoutFactory::GetParagraphLayout(CharType *buff, CharType *buff_end, FontMap &fontMapping)
{
	int32_t length = buff_end - buff;
	/* Can't layout an empty string. */
	if (length == 0) return nullptr;

	/* Can't layout our in-built sprite fonts. */
	for (auto const &pair : fontMapping) {
		if (pair.second->fc->IsBuiltInFont()) return nullptr;
	}

	/* Itemize text. */
	std::vector<SCRIPT_ITEM> items = UniscribeItemizeString(buff, length);
	if (items.empty()) return nullptr;

	/* Build ranges from the items and the font map. A range is a run of text
	 * that is part of a single item and formatted using a single font style. */
	std::vector<UniscribeRun> ranges;

	int cur_pos = 0;
	std::vector<SCRIPT_ITEM>::iterator cur_item = items.begin();
	for (auto const &i : fontMapping) {
		while (cur_pos < i.first && cur_item != items.end() - 1) {
			/* Add a range that spans the intersection of the remaining item and font run. */
			int stop_pos = std::min(i.first, (cur_item + 1)->iCharPos);
			assert(stop_pos - cur_pos > 0);
			ranges.push_back(UniscribeRun(cur_pos, stop_pos - cur_pos, i.second, cur_item->a));

			/* Shape the range. */
			if (!UniscribeShapeRun(buff, ranges.back())) {
				return nullptr;
			}

			/* If we are at the end of the current item, advance to the next item. */
			if (stop_pos == (cur_item + 1)->iCharPos) cur_item++;
			cur_pos = stop_pos;
		}
	}

	return new UniscribeParagraphLayout(ranges, buff);
}

/* virtual */ std::unique_ptr<const ParagraphLayouter::Line> UniscribeParagraphLayout::NextLine(int max_width)
{
	std::vector<UniscribeRun>::iterator start_run = this->cur_range;
	std::vector<UniscribeRun>::iterator last_run = this->cur_range;

	if (start_run == this->ranges.end()) return nullptr;

	/* Add remaining width of the first run if it is a broken run. */
	int cur_width = 0;
	if (this->cur_range_offset != 0) {
		std::vector<int> dx(start_run->len);
		ScriptGetLogicalWidths(&start_run->sa, start_run->len, (int)start_run->glyphs.size(), &start_run->advances[0], &start_run->char_to_glyph[0], &start_run->vis_attribs[0], &dx[0]);

		for (std::vector<int>::const_iterator c = dx.begin() + this->cur_range_offset; c != dx.end(); c++) {
			cur_width += *c;
		}
		++last_run;
	}

	/* Gather runs until the line is full. */
	while (last_run != this->ranges.end() && cur_width <= max_width) {
		cur_width += last_run->total_advance;
		++last_run;
	}

	/* If the text does not fit into the available width, find a suitable breaking point. */
	int remaining_offset = (last_run - 1)->len + 1;
	int whitespace_count = 0;
	if (cur_width > max_width) {
		std::vector<SCRIPT_LOGATTR> log_attribs;

		/* Get word break information. */
		int width_avail = max_width;
		int num_chars = this->cur_range_offset;
		int start_offs = this->cur_range_offset;
		int last_cluster = this->cur_range_offset + 1;
		for (std::vector<UniscribeRun>::iterator r = start_run; r != last_run; r++) {
			log_attribs.resize(r->pos - start_run->pos + r->len);
			if (FAILED(ScriptBreak(this->text_buffer + r->pos + start_offs, r->len - start_offs, &r->sa, &log_attribs[r->pos - start_run->pos + start_offs]))) return nullptr;

			std::vector<int> dx(r->len);
			ScriptGetLogicalWidths(&r->sa, r->len, (int)r->glyphs.size(), &r->advances[0], &r->char_to_glyph[0], &r->vis_attribs[0], &dx[0]);

			/* Count absolute max character count on the line. */
			for (int c = start_offs; c < r->len && width_avail > 0; c++, num_chars++) {
				if (c > start_offs && log_attribs[num_chars].fCharStop) last_cluster = num_chars;
				width_avail -= dx[c];
			}

			start_offs = 0;
		}

		/* Walk backwards to find the last suitable breaking point. */
		while (--num_chars > this->cur_range_offset && !log_attribs[num_chars].fSoftBreak && !log_attribs[num_chars].fWhiteSpace) {}

		if (num_chars == this->cur_range_offset) {
			/* Didn't find any suitable word break point, just break on the last cluster boundary. */
			num_chars = last_cluster;
		}

		/* Eat any whitespace characters before the breaking point. */
		while (num_chars - 1 > this->cur_range_offset && log_attribs[num_chars - 1].fWhiteSpace) num_chars--;
		/* Count whitespace after the breaking point. */
		while (num_chars + whitespace_count < (int)log_attribs.size() && log_attribs[num_chars + whitespace_count].fWhiteSpace) whitespace_count++;

		/* Get last run that corresponds to the number of characters to show. */
		for (std::vector<UniscribeRun>::iterator run = start_run; run != last_run; run++) {
			num_chars -= run->len;

			if (num_chars <= 0) {
				remaining_offset = num_chars + run->len + 1;
				last_run = run + 1;
				assert(remaining_offset - 1 > 0);
				break;
			}
		}
	}

	/* Build display order from the runs. */
	std::vector<BYTE> bidi_level;
	for (std::vector<UniscribeRun>::iterator r = start_run; r != last_run; r++) {
		bidi_level.push_back(r->sa.s.uBidiLevel);
	}
	std::vector<INT> vis_to_log(bidi_level.size());
	if (FAILED(ScriptLayout((int)bidi_level.size(), &bidi_level[0], &vis_to_log[0], nullptr))) return nullptr;

	/* Create line. */
	std::unique_ptr<UniscribeLine> line(new UniscribeLine());

	int cur_pos = 0;
	for (std::vector<INT>::iterator l = vis_to_log.begin(); l != vis_to_log.end(); l++) {
		std::vector<UniscribeRun>::iterator i_run = start_run + *l;
		UniscribeRun run = *i_run;

		/* Partial run after line break (either start or end)? Reshape run to get the first/last glyphs right. */
		if (i_run == last_run - 1 && remaining_offset <= (last_run - 1)->len) {
			run.len = remaining_offset - 1;

			if (!UniscribeShapeRun(this->text_buffer, run)) return nullptr;
		}
		if (i_run == start_run && this->cur_range_offset > 0) {
			assert(run.len - this->cur_range_offset > 0);
			run.pos += this->cur_range_offset;
			run.len -= this->cur_range_offset;

			if (!UniscribeShapeRun(this->text_buffer, run)) return nullptr;
		}

		line->emplace_back(run, cur_pos);
		cur_pos += run.total_advance;
	}

	if (remaining_offset + whitespace_count - 1 < (last_run - 1)->len) {
		/* We didn't use up all of the last run, store remainder for the next line. */
		this->cur_range_offset = remaining_offset + whitespace_count - 1;
		this->cur_range = last_run - 1;
		assert(this->cur_range->len > this->cur_range_offset);
	} else {
		this->cur_range_offset = 0;
		this->cur_range = last_run;
	}

	return line;
}

/**
 * Get the height of the line.
 * @return The maximum height of the line.
 */
int UniscribeParagraphLayout::UniscribeLine::GetLeading() const
{
	int leading = 0;
	for (const auto &run : *this) {
		leading = std::max(leading, run.GetLeading());
	}

	return leading;
}

/**
 * Get the width of this line.
 * @return The width of the line.
 */
int UniscribeParagraphLayout::UniscribeLine::GetWidth() const
{
	int length = 0;
	for (const auto &run : *this) {
		length += run.GetAdvance();
	}

	return length;
}

UniscribeParagraphLayout::UniscribeVisualRun::UniscribeVisualRun(const UniscribeRun &range, int x) : glyphs(range.ft_glyphs), char_to_glyph(range.char_to_glyph), start_pos(range.pos), total_advance(range.total_advance), font(range.font)
{
	this->num_glyphs = (int)glyphs.size();
	this->positions.reserve(this->num_glyphs + 1);

	int advance = x;
	for (int i = 0; i < this->num_glyphs; i++) {
		this->positions.emplace_back(range.offsets[i].du + advance, range.offsets[i].dv);

		advance += range.advances[i];
	}
	/* End-of-run position. */
	this->positions.emplace_back(advance, 0);
}

UniscribeParagraphLayout::UniscribeVisualRun::UniscribeVisualRun(UniscribeVisualRun&& other) noexcept
								: glyphs(std::move(other.glyphs)), positions(std::move(other.positions)), char_to_glyph(std::move(other.char_to_glyph)),
								  start_pos(other.start_pos), total_advance(other.total_advance), num_glyphs(other.num_glyphs), font(other.font),
								  glyph_to_char(std::move(other.glyph_to_char))
{
}

const std::vector<int> &UniscribeParagraphLayout::UniscribeVisualRun::GetGlyphToCharMap() const
{
	if (this->glyph_to_char.empty()) {
		this->glyph_to_char.resize(this->GetGlyphCount());

		/* The char to glyph array contains the first glyph index of the cluster that is associated
		 * with each character. It is possible for a cluster to be formed of several chars. */
		for (int c = 0; c < (int)this->char_to_glyph.size(); c++) {
			/* If multiple chars map to one glyph, only refer back to the first character. */
			if (this->glyph_to_char[this->char_to_glyph[c]] == 0) this->glyph_to_char[this->char_to_glyph[c]] = c + this->start_pos;
		}

		/* We only marked the first glyph of each cluster in the loop above. Fill the gaps. */
		int last_char = this->glyph_to_char[0];
		for (int g = 0; g < this->GetGlyphCount(); g++) {
			if (this->glyph_to_char[g] != 0) last_char = this->glyph_to_char[g];
			this->glyph_to_char[g] = last_char;
		}
	}

	return this->glyph_to_char;
}


/* virtual */ void UniscribeStringIterator::SetString(const char *s)
{
	const char *string_base = s;

	this->utf16_to_utf8.clear();
	this->str_info.clear();
	this->cur_pos = 0;

	/* Uniscribe operates on UTF-16, thus we have to convert the input string.
	 * To be able to return proper offsets, we have to create a mapping at the same time. */
	std::vector<wchar_t> utf16_str;     ///< UTF-16 copy of the string.
	while (*s != '\0') {
		size_t idx = s - string_base;

		char32_t c = Utf8Consume(&s);
		if (c < 0x10000) {
			utf16_str.push_back((wchar_t)c);
		} else {
			/* Make a surrogate pair. */
			utf16_str.push_back((wchar_t)(0xD800 + ((c - 0x10000) >> 10)));
			utf16_str.push_back((wchar_t)(0xDC00 + ((c - 0x10000) & 0x3FF)));
			this->utf16_to_utf8.push_back(idx);
		}
		this->utf16_to_utf8.push_back(idx);
	}
	this->utf16_to_utf8.push_back(s - string_base);

	/* Query Uniscribe for word and cluster break information. */
	this->str_info.resize(utf16_to_utf8.size());

	if (!utf16_str.empty()) {
		/* Itemize string into language runs. */
		std::vector<SCRIPT_ITEM> runs = UniscribeItemizeString(&utf16_str[0], (int32_t)utf16_str.size());

		for (std::vector<SCRIPT_ITEM>::const_iterator run = runs.begin(); !runs.empty() && run != runs.end() - 1; run++) {
			/* Get information on valid word and character break.s */
			int len = (run + 1)->iCharPos - run->iCharPos;
			std::vector<SCRIPT_LOGATTR> attr(len);
			ScriptBreak(&utf16_str[run->iCharPos], len, &run->a, &attr[0]);

			/* Extract the information we're interested in. */
			for (size_t c = 0; c < attr.size(); c++) {
				/* First character of a run is always a valid word break. */
				this->str_info[c + run->iCharPos].word_stop = attr[c].fWordStop || c == 0;
				this->str_info[c + run->iCharPos].char_stop = attr[c].fCharStop;
			}
		}
	}

	/* End-of-string is always a valid stopping point. */
	this->str_info.back().char_stop = true;
	this->str_info.back().word_stop = true;
}

/* virtual */ size_t UniscribeStringIterator::SetCurPosition(size_t pos)
{
	/* Convert incoming position to an UTF-16 string index. */
	size_t utf16_pos = 0;
	for (size_t i = 0; i < this->utf16_to_utf8.size(); i++) {
		if (this->utf16_to_utf8[i] == pos) {
			utf16_pos = i;
			break;
		}
	}

	/* Sanitize in case we get a position inside a grapheme cluster. */
	while (utf16_pos > 0 && !this->str_info[utf16_pos].char_stop) utf16_pos--;
	this->cur_pos = utf16_pos;

	return this->utf16_to_utf8[this->cur_pos];
}

/* virtual */ size_t UniscribeStringIterator::Next(IterType what)
{
	assert(this->cur_pos <= this->utf16_to_utf8.size());
	assert(what == StringIterator::ITER_CHARACTER || what == StringIterator::ITER_WORD);

	if (this->cur_pos == this->utf16_to_utf8.size()) return END;

	do {
		this->cur_pos++;
	} while (this->cur_pos < this->utf16_to_utf8.size() && (what  == ITER_WORD ? !this->str_info[this->cur_pos].word_stop : !this->str_info[this->cur_pos].char_stop));

	return this->cur_pos == this->utf16_to_utf8.size() ? END : this->utf16_to_utf8[this->cur_pos];
}

/*virtual */ size_t UniscribeStringIterator::Prev(IterType what)
{
	assert(this->cur_pos <= this->utf16_to_utf8.size());
	assert(what == StringIterator::ITER_CHARACTER || what == StringIterator::ITER_WORD);

	if (this->cur_pos == 0) return END;

	do {
		this->cur_pos--;
	} while (this->cur_pos > 0 && (what == ITER_WORD ? !this->str_info[this->cur_pos].word_stop : !this->str_info[this->cur_pos].char_stop));

	return this->utf16_to_utf8[this->cur_pos];
}