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
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
/* 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 https://mozilla.org/MPL/2.0/. */

//! Text layout.

use std::borrow::ToOwned;
use std::collections::LinkedList;
use std::sync::Arc;

use app_units::Au;
use base::text::is_bidi_control;
use fonts::{
    self, ByteIndex, FontIdentifier, FontMetrics, FontRef, RunMetrics, ShapingFlags,
    ShapingOptions, LAST_RESORT_GLYPH_ADVANCE,
};
use log::{debug, warn};
use range::Range;
use style::computed_values::text_rendering::T as TextRendering;
use style::computed_values::white_space_collapse::T as WhiteSpaceCollapse;
use style::computed_values::word_break::T as WordBreak;
use style::logical_geometry::{LogicalSize, WritingMode};
use style::properties::style_structs::Font as FontStyleStruct;
use style::properties::ComputedValues;
use style::values::generics::font::LineHeight;
use style::values::specified::text::{TextTransform, TextTransformCase};
use unicode_bidi as bidi;
use unicode_script::Script;
use xi_unicode::LineBreakLeafIter;

use crate::context::LayoutFontContext;
use crate::fragment::{
    Fragment, ScannedTextFlags, ScannedTextFragmentInfo, SpecificFragmentInfo,
    UnscannedTextFragmentInfo,
};
use crate::inline::{InlineFragmentNodeFlags, InlineFragments};
use crate::linked_list::split_off_head;
use crate::text_run::TextRun;

/// Returns the concatenated text of a list of unscanned text fragments.
fn text(fragments: &LinkedList<Fragment>) -> String {
    // FIXME: Some of this work is later duplicated in split_first_fragment_at_newline_if_necessary
    // and transform_text.  This code should be refactored so that the all the scanning for
    // newlines is done in a single pass.

    let mut text = String::new();

    for fragment in fragments {
        if let SpecificFragmentInfo::UnscannedText(ref info) = fragment.specific {
            if fragment.white_space_collapse() != WhiteSpaceCollapse::Collapse {
                text.push_str(&info.text);
            } else {
                text.push_str(&info.text.replace('\n', " "));
            }
        }
    }
    text
}

/// A stack-allocated object for scanning an inline flow into `TextRun`-containing `TextFragment`s.
pub struct TextRunScanner {
    pub clump: LinkedList<Fragment>,
}

impl TextRunScanner {
    pub fn new() -> TextRunScanner {
        TextRunScanner {
            clump: LinkedList::new(),
        }
    }

    pub fn scan_for_runs(
        &mut self,
        font_context: &LayoutFontContext,
        mut fragments: LinkedList<Fragment>,
    ) -> InlineFragments {
        debug!(
            "TextRunScanner: scanning {} fragments for text runs...",
            fragments.len()
        );
        debug_assert!(!fragments.is_empty());

        // Calculate bidi embedding levels, so we can split bidirectional fragments for reordering.
        let text = text(&fragments);
        let para_level = fragments
            .front()
            .unwrap()
            .style
            .writing_mode
            .to_bidi_level();
        let bidi_info = bidi::BidiInfo::new(&text, Some(para_level));

        // Optimization: If all the text is LTR, don't bother splitting on bidi levels.
        let bidi_levels = if bidi_info.has_rtl() {
            Some(&bidi_info.levels[..])
        } else {
            None
        };

        // FIXME(pcwalton): We want to be sure not to allocate multiple times, since this is a
        // performance-critical spot, but this may overestimate and allocate too much memory.
        let mut new_fragments = Vec::with_capacity(fragments.len());
        let mut last_whitespace = false;
        let mut paragraph_bytes_processed = 0;

        // The first time we process a text run we will set this
        // linebreaker. There is no way for the linebreaker to start
        // with an empty state; you must give it its first input immediately.
        //
        // This linebreaker is shared across text runs, so we can know if
        // there is a break at the beginning of a text run or clump, e.g.
        // in the case of FooBar<span>Baz</span>
        let mut linebreaker = None;

        while !fragments.is_empty() {
            // Create a clump.
            split_first_fragment_at_newline_if_necessary(&mut fragments);
            self.clump.append(&mut split_off_head(&mut fragments));
            while !fragments.is_empty() &&
                self.clump
                    .back()
                    .unwrap()
                    .can_merge_with_fragment(fragments.front().unwrap())
            {
                split_first_fragment_at_newline_if_necessary(&mut fragments);
                self.clump.append(&mut split_off_head(&mut fragments));
            }

            // Flush that clump to the list of fragments we're building up.
            last_whitespace = self.flush_clump_to_list(
                font_context,
                &mut new_fragments,
                &mut paragraph_bytes_processed,
                bidi_levels,
                last_whitespace,
                &mut linebreaker,
            );
        }

        debug!("TextRunScanner: complete.");
        InlineFragments {
            fragments: new_fragments,
        }
    }

    /// A "clump" is a range of inline flow leaves that can be merged together into a single
    /// fragment. Adjacent text with the same style can be merged, and nothing else can.
    ///
    /// The flow keeps track of the fragments contained by all non-leaf DOM nodes. This is necessary
    /// for correct painting order. Since we compress several leaf fragments here, the mapping must
    /// be adjusted.
    fn flush_clump_to_list(
        &mut self,
        font_context: &LayoutFontContext,
        out_fragments: &mut Vec<Fragment>,
        paragraph_bytes_processed: &mut usize,
        bidi_levels: Option<&[bidi::Level]>,
        mut last_whitespace: bool,
        linebreaker: &mut Option<LineBreakLeafIter>,
    ) -> bool {
        debug!(
            "TextRunScanner: flushing {} fragments in range",
            self.clump.len()
        );

        debug_assert!(!self.clump.is_empty());
        match self.clump.front().unwrap().specific {
            SpecificFragmentInfo::UnscannedText(_) => {},
            _ => {
                debug_assert!(
                    self.clump.len() == 1,
                    "WAT: can't coalesce non-text nodes in flush_clump_to_list()!"
                );
                out_fragments.push(self.clump.pop_front().unwrap());
                return false;
            },
        }

        // Concatenate all of the transformed strings together, saving the new character indices.
        let mut mappings: Vec<RunMapping> = Vec::new();
        let runs = {
            let font_group;
            let compression;
            let text_transform;
            let letter_spacing;
            let word_spacing;
            let text_rendering;
            let word_break;
            {
                let in_fragment = self.clump.front().unwrap();
                let font_style = in_fragment.style().clone_font();
                let inherited_text_style = in_fragment.style().get_inherited_text();
                font_group = font_context.font_group(font_style);
                compression = match in_fragment.white_space_collapse() {
                    WhiteSpaceCollapse::Collapse => CompressionMode::CompressWhitespaceNewline,
                    WhiteSpaceCollapse::Preserve | WhiteSpaceCollapse::BreakSpaces => {
                        CompressionMode::CompressNone
                    },
                    WhiteSpaceCollapse::PreserveBreaks => CompressionMode::CompressWhitespace,
                };
                text_transform = inherited_text_style.text_transform;
                letter_spacing = inherited_text_style.letter_spacing;
                word_spacing = inherited_text_style
                    .word_spacing
                    .to_length()
                    .map(|l| l.into())
                    .unwrap_or_else(|| {
                        let space_width = font_group
                            .write()
                            .find_by_codepoint(font_context, ' ', None)
                            .and_then(|font| {
                                font.glyph_index(' ')
                                    .map(|glyph_id| font.glyph_h_advance(glyph_id))
                            })
                            .unwrap_or(LAST_RESORT_GLYPH_ADVANCE);
                        inherited_text_style
                            .word_spacing
                            .to_used_value(Au::from_f64_px(space_width))
                    });
                text_rendering = inherited_text_style.text_rendering;
                word_break = inherited_text_style.word_break;
            }

            // First, transform/compress text of all the nodes.
            let (mut run_info_list, mut run_info) = (Vec::new(), RunInfo::new());
            let mut insertion_point = None;

            for (fragment_index, in_fragment) in self.clump.iter().enumerate() {
                debug!("  flushing {:?}", in_fragment);
                let mut mapping = RunMapping::new(&run_info_list[..], fragment_index);
                let text;
                let selection;
                match in_fragment.specific {
                    SpecificFragmentInfo::UnscannedText(ref text_fragment_info) => {
                        text = &text_fragment_info.text;
                        selection = text_fragment_info.selection;
                    },
                    _ => panic!("Expected an unscanned text fragment!"),
                };
                insertion_point = match selection {
                    Some(range) if range.is_empty() => {
                        // `range` is the range within the current fragment. To get the range
                        // within the text run, offset it by the length of the preceding fragments.
                        Some(range.begin() + ByteIndex(run_info.text.len() as isize))
                    },
                    _ => None,
                };

                let (mut start_position, mut end_position) = (0, 0);
                for (byte_index, character) in text.char_indices() {
                    if !character.is_control() {
                        let font =
                            font_group
                                .write()
                                .find_by_codepoint(font_context, character, None);

                        let bidi_level = match bidi_levels {
                            Some(levels) => levels[*paragraph_bytes_processed],
                            None => bidi::Level::ltr(),
                        };

                        // Break the run if the new character has a different explicit script than the
                        // previous characters.
                        //
                        // TODO: Special handling of paired punctuation characters.
                        // http://www.unicode.org/reports/tr24/#Common
                        let script = Script::from(character);
                        let compatible_script = is_compatible(script, run_info.script);
                        if compatible_script && !is_specific(run_info.script) && is_specific(script)
                        {
                            run_info.script = script;
                        }

                        let selected = match selection {
                            Some(range) => range.contains(ByteIndex(byte_index as isize)),
                            None => false,
                        };

                        // Now, if necessary, flush the mapping we were building up.
                        let flush_run = !run_info.has_font(&font) ||
                            run_info.bidi_level != bidi_level ||
                            !compatible_script;
                        let new_mapping_needed = flush_run || mapping.selected != selected;

                        if new_mapping_needed {
                            // We ignore empty mappings at the very start of a fragment.
                            // The run info values are uninitialized at this point so
                            // flushing an empty mapping is pointless.
                            if end_position > 0 {
                                mapping.flush(
                                    &mut mappings,
                                    &mut run_info,
                                    text,
                                    compression,
                                    text_transform,
                                    &mut last_whitespace,
                                    &mut start_position,
                                    end_position,
                                );
                            }
                            if !run_info.text.is_empty() {
                                if flush_run {
                                    run_info.flush(&mut run_info_list, &mut insertion_point);
                                    run_info = RunInfo::new();
                                }
                                mapping = RunMapping::new(&run_info_list[..], fragment_index);
                            }
                            run_info.font = font;
                            run_info.bidi_level = bidi_level;
                            run_info.script = script;
                            mapping.selected = selected;
                        }
                    }

                    // Consume this character.
                    end_position += character.len_utf8();
                    *paragraph_bytes_processed += character.len_utf8();
                }

                // Flush the last mapping we created for this fragment to the list.
                mapping.flush(
                    &mut mappings,
                    &mut run_info,
                    text,
                    compression,
                    text_transform,
                    &mut last_whitespace,
                    &mut start_position,
                    end_position,
                );
            }

            // Push the final run info.
            run_info.flush(&mut run_info_list, &mut insertion_point);

            // Per CSS 2.1 § 16.4, "when the resultant space between two characters is not the same
            // as the default space, user agents should not use ligatures." This ensures that, for
            // example, `finally` with a wide `letter-spacing` renders as `f i n a l l y` and not
            // `fi n a l l y`.
            let mut flags = ShapingFlags::empty();
            if letter_spacing.0.px() != 0. {
                flags.insert(ShapingFlags::IGNORE_LIGATURES_SHAPING_FLAG);
            }
            if text_rendering == TextRendering::Optimizespeed {
                flags.insert(ShapingFlags::IGNORE_LIGATURES_SHAPING_FLAG);
                flags.insert(ShapingFlags::DISABLE_KERNING_SHAPING_FLAG)
            }
            if word_break == WordBreak::KeepAll {
                flags.insert(ShapingFlags::KEEP_ALL_FLAG);
            }
            let options = ShapingOptions {
                letter_spacing: if letter_spacing.0.px() == 0. {
                    None
                } else {
                    Some(Au::from(letter_spacing.0))
                },
                word_spacing,
                script: Script::Common,
                flags,
            };

            let mut result = Vec::with_capacity(run_info_list.len());
            for run_info in run_info_list {
                let mut options = options;
                options.script = run_info.script;
                if run_info.bidi_level.is_rtl() {
                    options.flags.insert(ShapingFlags::RTL_FLAG);
                }

                // If no font is found (including fallbacks), there's no way we can render.
                let font = match run_info
                    .font
                    .or_else(|| font_group.write().first(font_context))
                {
                    Some(font) => font,
                    None => {
                        result.push(None);
                        continue;
                    },
                };

                let (run, break_at_zero) = TextRun::new(
                    font,
                    run_info.text,
                    &options,
                    run_info.bidi_level,
                    linebreaker,
                );
                result.push(Some((
                    ScannedTextRun {
                        run: Arc::new(run),
                        insertion_point: run_info.insertion_point,
                    },
                    break_at_zero,
                )))
            }
            result
        };

        // Make new fragments with the runs and adjusted text indices.
        debug!("TextRunScanner: pushing {} fragment(s)", self.clump.len());
        let mut mappings = mappings.into_iter().peekable();
        let mut prev_fragments_to_meld = Vec::new();

        for (logical_offset, old_fragment) in
            std::mem::take(&mut self.clump).into_iter().enumerate()
        {
            let mut is_first_mapping_of_this_old_fragment = true;
            loop {
                match mappings.peek() {
                    Some(mapping) if mapping.old_fragment_index == logical_offset => {},
                    Some(_) | None => {
                        if is_first_mapping_of_this_old_fragment {
                            // There were no mappings for this unscanned fragment. Transfer its
                            // flags to the previous/next sibling elements instead.
                            if let Some(ref mut last_fragment) = out_fragments.last_mut() {
                                last_fragment.meld_with_next_inline_fragment(&old_fragment);
                            }
                            prev_fragments_to_meld.push(old_fragment);
                        }
                        break;
                    },
                };
                let mapping = mappings.next().unwrap();
                let run = runs[mapping.text_run_index].clone();
                let (scanned_run, break_at_zero) = match run {
                    Some(run) => run,
                    None => {
                        warn!("Could not find found for TextRun!");
                        continue;
                    },
                };

                let mut byte_range = Range::new(
                    ByteIndex(mapping.byte_range.begin() as isize),
                    ByteIndex(mapping.byte_range.length() as isize),
                );

                let mut flags = ScannedTextFlags::empty();
                if !break_at_zero && mapping.byte_range.begin() == 0 {
                    // If this is the first segment of the text run,
                    // and the text run doesn't break at zero, suppress line breaks
                    flags.insert(ScannedTextFlags::SUPPRESS_LINE_BREAK_BEFORE)
                }
                let text_size = old_fragment.border_box.size;

                let requires_line_break_afterward_if_wrapping_on_newlines = scanned_run.run.text
                    [mapping.byte_range.begin()..mapping.byte_range.end()]
                    .ends_with('\n');

                if requires_line_break_afterward_if_wrapping_on_newlines {
                    byte_range.extend_by(ByteIndex(-1)); // Trim the '\n'
                    flags.insert(
                        ScannedTextFlags::REQUIRES_LINE_BREAK_AFTERWARD_IF_WRAPPING_ON_NEWLINES,
                    );
                }

                if mapping.selected {
                    flags.insert(ScannedTextFlags::SELECTED);
                }

                let insertion_point =
                    if mapping.contains_insertion_point(scanned_run.insertion_point) {
                        scanned_run.insertion_point
                    } else {
                        None
                    };

                let mut new_text_fragment_info = Box::new(ScannedTextFragmentInfo::new(
                    scanned_run.run,
                    byte_range,
                    text_size,
                    insertion_point,
                    flags,
                ));

                let new_metrics = new_text_fragment_info.run.metrics_for_range(&byte_range);
                let writing_mode = old_fragment.style.writing_mode;
                let bounding_box_size = bounding_box_for_run_metrics(&new_metrics, writing_mode);
                new_text_fragment_info.content_size = bounding_box_size;

                let mut new_fragment = old_fragment.transform(
                    bounding_box_size,
                    SpecificFragmentInfo::ScannedText(new_text_fragment_info),
                );

                let is_last_mapping_of_this_old_fragment = !matches!(
                    mappings.peek(),
                    Some(mapping) if mapping.old_fragment_index == logical_offset
                );

                if let Some(ref mut context) = new_fragment.inline_context {
                    for node in &mut context.nodes {
                        if !is_last_mapping_of_this_old_fragment {
                            node.flags
                                .remove(InlineFragmentNodeFlags::LAST_FRAGMENT_OF_ELEMENT);
                        }
                        if !is_first_mapping_of_this_old_fragment {
                            node.flags
                                .remove(InlineFragmentNodeFlags::FIRST_FRAGMENT_OF_ELEMENT);
                        }
                    }
                }

                for prev_fragment in prev_fragments_to_meld.drain(..) {
                    new_fragment.meld_with_prev_inline_fragment(&prev_fragment);
                }

                is_first_mapping_of_this_old_fragment = false;
                out_fragments.push(new_fragment)
            }
        }

        last_whitespace
    }
}

#[inline]
fn bounding_box_for_run_metrics(
    metrics: &RunMetrics,
    writing_mode: WritingMode,
) -> LogicalSize<Au> {
    // TODO: When the text-orientation property is supported, the block and inline directions may
    // be swapped for horizontal glyphs in vertical lines.
    LogicalSize::new(
        writing_mode,
        metrics.bounding_box.size.width,
        metrics.bounding_box.size.height,
    )
}

/// Returns the metrics of the font represented by the given `FontStyleStruct`.
///
/// `#[inline]` because often the caller only needs a few fields from the font metrics.
///
/// # Panics
///
/// Panics if no font can be found for the given font style.
#[inline]
pub fn font_metrics_for_style(
    font_context: &LayoutFontContext,
    style: crate::ServoArc<FontStyleStruct>,
) -> FontMetrics {
    let font_group = font_context.font_group(style);
    let font = font_group.write().first(font_context);
    font.as_ref().unwrap().metrics.clone()
}

/// Returns the line block-size needed by the given computed style and font size.
pub fn line_height_from_style(style: &ComputedValues, metrics: &FontMetrics) -> Au {
    let font = style.get_font();
    let font_size = font.font_size.computed_size();
    match font.line_height {
        LineHeight::Normal => metrics.line_gap,
        LineHeight::Number(l) => Au::from(font_size * l.0),
        LineHeight::Length(l) => Au::from(l),
    }
}

fn split_first_fragment_at_newline_if_necessary(fragments: &mut LinkedList<Fragment>) {
    if fragments.is_empty() {
        return;
    }

    let new_fragment = {
        let first_fragment = fragments.front_mut().unwrap();
        let string_before;
        let selection_before;
        {
            if first_fragment.white_space_collapse() == WhiteSpaceCollapse::Collapse {
                return;
            }

            let unscanned_text_fragment_info = match first_fragment.specific {
                SpecificFragmentInfo::UnscannedText(ref mut unscanned_text_fragment_info) => {
                    unscanned_text_fragment_info
                },
                _ => return,
            };

            let position = match unscanned_text_fragment_info.text.find('\n') {
                Some(position) if position < unscanned_text_fragment_info.text.len() - 1 => {
                    position
                },
                Some(_) | None => return,
            };

            string_before = unscanned_text_fragment_info.text[..(position + 1)].to_owned();
            unscanned_text_fragment_info.text = unscanned_text_fragment_info.text[(position + 1)..]
                .to_owned()
                .into_boxed_str();
            let offset = ByteIndex(string_before.len() as isize);
            match unscanned_text_fragment_info.selection {
                Some(ref mut selection) if selection.begin() >= offset => {
                    // Selection is entirely in the second fragment.
                    selection_before = None;
                    selection.shift_by(-offset);
                },
                Some(ref mut selection) if selection.end() > offset => {
                    // Selection is split across two fragments.
                    selection_before = Some(Range::new(selection.begin(), offset));
                    *selection = Range::new(ByteIndex(0), selection.end() - offset);
                },
                _ => {
                    // Selection is entirely in the first fragment.
                    selection_before = unscanned_text_fragment_info.selection;
                    unscanned_text_fragment_info.selection = None;
                },
            };
        }
        first_fragment.transform(
            first_fragment.border_box.size,
            SpecificFragmentInfo::UnscannedText(Box::new(UnscannedTextFragmentInfo::new(
                string_before.into_boxed_str(),
                selection_before,
            ))),
        )
    };

    fragments.push_front(new_fragment);
}

/// Information about a text run that we're about to create. This is used in `scan_for_runs`.
struct RunInfo {
    /// The text that will go in this text run.
    text: String,
    /// The insertion point in this text run, if applicable.
    insertion_point: Option<ByteIndex>,
    /// The font that the text should be rendered with.
    font: Option<FontRef>,
    /// The bidirection embedding level of this text run.
    bidi_level: bidi::Level,
    /// The Unicode script property of this text run.
    script: Script,
}

impl RunInfo {
    fn new() -> RunInfo {
        RunInfo {
            text: String::new(),
            insertion_point: None,
            font: None,
            bidi_level: bidi::Level::ltr(),
            script: Script::Common,
        }
    }

    /// Finish processing this RunInfo and add it to the "done" list.
    ///
    /// * `insertion_point`: The position of the insertion point, in characters relative to the start
    ///   of this text run.
    fn flush(mut self, list: &mut Vec<RunInfo>, insertion_point: &mut Option<ByteIndex>) {
        if let Some(idx) = *insertion_point {
            let char_len = ByteIndex(self.text.len() as isize);
            if idx <= char_len {
                // The insertion point is in this text run.
                self.insertion_point = insertion_point.take()
            } else {
                // Continue looking for the insertion point in the next text run.
                *insertion_point = Some(idx - char_len)
            }
        }
        list.push(self);
    }

    fn has_font(&self, font: &Option<FontRef>) -> bool {
        fn identifier_and_pt_size(font: &Option<FontRef>) -> Option<(FontIdentifier, Au)> {
            font.as_ref()
                .map(|font| (font.identifier().clone(), font.descriptor.pt_size))
        }

        identifier_and_pt_size(&self.font) == identifier_and_pt_size(font)
    }
}

/// A mapping from a portion of an unscanned text fragment to the text run we're going to create
/// for it.
#[derive(Clone, Copy, Debug)]
struct RunMapping {
    /// The range of byte indices within the text fragment.
    byte_range: Range<usize>,
    /// The index of the unscanned text fragment that this mapping corresponds to.
    old_fragment_index: usize,
    /// The index of the text run we're going to create.
    text_run_index: usize,
    /// Is the text in this fragment selected?
    selected: bool,
}

impl RunMapping {
    /// Given the current set of text runs, creates a run mapping for the next fragment.
    /// `run_info_list` describes the set of runs we've seen already.
    fn new(run_info_list: &[RunInfo], fragment_index: usize) -> RunMapping {
        RunMapping {
            byte_range: Range::new(0, 0),
            old_fragment_index: fragment_index,
            text_run_index: run_info_list.len(),
            selected: false,
        }
    }

    /// Flushes this run mapping to the list. `run_info` describes the text run that we're
    /// currently working on. `text` refers to the text of this fragment.
    #[allow(clippy::too_many_arguments)]
    fn flush(
        mut self,
        mappings: &mut Vec<RunMapping>,
        run_info: &mut RunInfo,
        text: &str,
        compression: CompressionMode,
        text_transform: TextTransform,
        last_whitespace: &mut bool,
        start_position: &mut usize,
        end_position: usize,
    ) {
        let was_empty = *start_position == end_position;
        let old_byte_length = run_info.text.len();
        *last_whitespace = transform_text(
            &text[(*start_position)..end_position],
            compression,
            *last_whitespace,
            &mut run_info.text,
        );

        // Account for `text-transform`. (Confusingly, this is not handled in "text
        // transformation" above, but we follow Gecko in the naming.)
        let is_first_run = *start_position == 0;
        apply_style_transform_if_necessary(
            &mut run_info.text,
            old_byte_length,
            text_transform,
            *last_whitespace,
            is_first_run,
        );
        *start_position = end_position;

        let new_byte_length = run_info.text.len();
        let is_empty = new_byte_length == old_byte_length;

        // Don't save mappings that contain only discarded characters.
        // (But keep ones that contained no characters to begin with, since they might have been
        // generated by an empty flow to draw its borders/padding/insertion point.)
        if is_empty && !was_empty {
            return;
        }

        self.byte_range = Range::new(old_byte_length, new_byte_length - old_byte_length);
        mappings.push(self)
    }

    /// Is the insertion point for this text run within this mapping?
    ///
    /// NOTE: We treat the range as inclusive at both ends, since the insertion point can lie
    /// before the first character *or* after the last character, and should be drawn even if the
    /// text is empty.
    fn contains_insertion_point(&self, insertion_point: Option<ByteIndex>) -> bool {
        match insertion_point.map(ByteIndex::to_usize) {
            None => false,
            Some(idx) => self.byte_range.begin() <= idx && idx <= self.byte_range.end(),
        }
    }
}

/// Accounts for `text-transform`.
///
/// FIXME(#4311, pcwalton): Title-case mapping can change length of the string;
/// case mapping should be language-specific; `full-width`;
/// use graphemes instead of characters.
fn apply_style_transform_if_necessary(
    string: &mut String,
    first_character_position: usize,
    text_transform: TextTransform,
    last_whitespace: bool,
    is_first_run: bool,
) {
    match text_transform.case() {
        TextTransformCase::None => {},
        TextTransformCase::Uppercase => {
            let original = string[first_character_position..].to_owned();
            string.truncate(first_character_position);
            for ch in original.chars().flat_map(|ch| ch.to_uppercase()) {
                string.push(ch);
            }
        },
        TextTransformCase::Lowercase => {
            let original = string[first_character_position..].to_owned();
            string.truncate(first_character_position);
            for ch in original.chars().flat_map(|ch| ch.to_lowercase()) {
                string.push(ch);
            }
        },
        TextTransformCase::Capitalize => {
            let original = string[first_character_position..].to_owned();
            string.truncate(first_character_position);

            let mut capitalize_next_letter = is_first_run || last_whitespace;
            for character in original.chars() {
                // FIXME(#4311, pcwalton): Should be the CSS/Unicode notion of a *typographic
                // letter unit*, not an *alphabetic* character:
                //
                //    http://dev.w3.org/csswg/css-text/#typographic-letter-unit
                if capitalize_next_letter && character.is_alphabetic() {
                    string.push(character.to_uppercase().next().unwrap());
                    capitalize_next_letter = false;
                    continue;
                }

                string.push(character);

                // FIXME(#4311, pcwalton): Try UAX29 instead of just whitespace.
                if character.is_whitespace() {
                    capitalize_next_letter = true
                }
            }
        },
    }
}

#[derive(Clone)]
struct ScannedTextRun {
    run: Arc<TextRun>,
    insertion_point: Option<ByteIndex>,
}

/// Can a character with script `b` continue a text run with script `a`?
fn is_compatible(a: Script, b: Script) -> bool {
    a == b || !is_specific(a) || !is_specific(b)
}

/// Returns true if the script is not invalid or inherited.
fn is_specific(script: Script) -> bool {
    script != Script::Common && script != Script::Inherited
}

#[derive(Clone, Copy, Debug, Eq, PartialEq)]
#[allow(clippy::enum_variant_names)]
pub enum CompressionMode {
    CompressNone,
    CompressWhitespace,
    CompressWhitespaceNewline,
}

// ported from Gecko's nsTextFrameUtils::TransformText.
//
// High level TODOs:
//
// * Issue #113: consider incoming text state (arabic, etc)
//               and propagate outgoing text state (dual of above)
//
// * Issue #114: record skipped and kept chars for mapping original to new text
//
// * Untracked: various edge cases for bidi, CJK, etc.
pub fn transform_text(
    text: &str,
    mode: CompressionMode,
    incoming_whitespace: bool,
    output_text: &mut String,
) -> bool {
    let out_whitespace = match mode {
        CompressionMode::CompressNone => {
            for ch in text.chars() {
                if is_discardable_char(ch, mode) {
                    // TODO: record skipped char
                } else {
                    // TODO: record kept char
                    if ch == '\t' {
                        // TODO: set "has tab" flag
                    }
                    output_text.push(ch);
                }
            }
            false
        },

        CompressionMode::CompressWhitespace | CompressionMode::CompressWhitespaceNewline => {
            let mut in_whitespace: bool = incoming_whitespace;
            for ch in text.chars() {
                // TODO: discard newlines between CJK chars
                let mut next_in_whitespace: bool = is_in_whitespace(ch, mode);

                if !next_in_whitespace {
                    if is_always_discardable_char(ch) {
                        // revert whitespace setting, since this char was discarded
                        next_in_whitespace = in_whitespace;
                    // TODO: record skipped char
                    } else {
                        // TODO: record kept char
                        output_text.push(ch);
                    }
                } else {
                    /* next_in_whitespace; possibly add a space char */
                    if in_whitespace {
                        // TODO: record skipped char
                    } else {
                        // TODO: record kept char
                        output_text.push(' ');
                    }
                }
                // save whitespace context for next char
                in_whitespace = next_in_whitespace;
            } /* /for str::each_char */
            in_whitespace
        },
    };

    return out_whitespace;

    fn is_in_whitespace(ch: char, mode: CompressionMode) -> bool {
        match (ch, mode) {
            (' ', _) => true,
            ('\t', _) => true,
            ('\n', CompressionMode::CompressWhitespaceNewline) => true,
            (_, _) => false,
        }
    }

    fn is_discardable_char(ch: char, mode: CompressionMode) -> bool {
        if is_always_discardable_char(ch) {
            return true;
        }
        match mode {
            CompressionMode::CompressWhitespaceNewline => ch == '\n',
            _ => false,
        }
    }

    fn is_always_discardable_char(ch: char) -> bool {
        // TODO: check for soft hyphens.
        is_bidi_control(ch)
    }
}

#[test]
fn test_transform_compress_none() {
    let test_strs = [
        "  foo bar",
        "foo bar  ",
        "foo\n bar",
        "foo \nbar",
        "  foo  bar  \nbaz",
        "foo bar baz",
        "foobarbaz\n\n",
    ];

    let mode = CompressionMode::CompressNone;
    for &test in test_strs.iter() {
        let mut trimmed_str = String::new();
        transform_text(test, mode, true, &mut trimmed_str);
        assert_eq!(trimmed_str, test)
    }
}

#[test]
fn test_transform_compress_whitespace() {
    let test_strs = [
        ("  foo bar", "foo bar"),
        ("foo bar  ", "foo bar "),
        ("foo\n bar", "foo\n bar"),
        ("foo \nbar", "foo \nbar"),
        ("  foo  bar  \nbaz", "foo bar \nbaz"),
        ("foo bar baz", "foo bar baz"),
        ("foobarbaz\n\n", "foobarbaz\n\n"),
    ];

    let mode = CompressionMode::CompressWhitespace;
    for &(test, oracle) in test_strs.iter() {
        let mut trimmed_str = String::new();
        transform_text(test, mode, true, &mut trimmed_str);
        assert_eq!(&*trimmed_str, oracle)
    }
}

#[test]
fn test_transform_compress_whitespace_newline() {
    let test_strs = vec![
        ("  foo bar", "foo bar"),
        ("foo bar  ", "foo bar "),
        ("foo\n bar", "foo bar"),
        ("foo \nbar", "foo bar"),
        ("  foo  bar  \nbaz", "foo bar baz"),
        ("foo bar baz", "foo bar baz"),
        ("foobarbaz\n\n", "foobarbaz "),
    ];

    let mode = CompressionMode::CompressWhitespaceNewline;
    for &(test, oracle) in test_strs.iter() {
        let mut trimmed_str = String::new();
        transform_text(test, mode, true, &mut trimmed_str);
        assert_eq!(&*trimmed_str, oracle)
    }
}

#[test]
fn test_transform_compress_whitespace_newline_no_incoming() {
    let test_strs = [
        ("  foo bar", " foo bar"),
        ("\nfoo bar", " foo bar"),
        ("foo bar  ", "foo bar "),
        ("foo\n bar", "foo bar"),
        ("foo \nbar", "foo bar"),
        ("  foo  bar  \nbaz", " foo bar baz"),
        ("foo bar baz", "foo bar baz"),
        ("foobarbaz\n\n", "foobarbaz "),
    ];

    let mode = CompressionMode::CompressWhitespaceNewline;
    for &(test, oracle) in test_strs.iter() {
        let mut trimmed_str = String::new();
        transform_text(test, mode, false, &mut trimmed_str);
        assert_eq!(trimmed_str, oracle)
    }
}