html-to-markdown-rs 3.11.0

High-performance HTML to Markdown converter using the astral-tl parser. Part of the Xberg ecosystem.
Documentation
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
//! Handler for link elements (a, anchor).
//!
//! Converts HTML anchor tags to Markdown links with support for:
//! - Standard Markdown link syntax `[label](href "title")`
//! - Autolinks for simple URLs like `<https://example.com>`
//! - Link label escaping for special Markdown characters
//! - Heading-in-link special handling (wraps link around heading)
//! - Visitor callbacks for custom link processing
//! - Metadata collection for links (links, URLs, titles, rel attributes)
//! - Block-level content within links (via inline context)

use crate::converter::utility::content::{collect_link_label_text, escape_link_label, normalize_link_label};
use crate::converter::utility::preprocessing::sanitize_markdown_url;
use crate::options::ConversionOptions;
use std::borrow::Cow;
#[cfg(feature = "metadata")]
use std::collections::BTreeMap;
use tl::{NodeHandle, Parser};

type Context = crate::converter::Context;
type DomContext = crate::converter::DomContext;

/// Handler for anchor/link elements: `<a>`.
///
/// Processes anchor tags to generate Markdown links:
/// - Detects autolinks (link text matches href)
/// - Extracts and normalizes link labels
/// - Handles nested headings within links
/// - Escapes special characters in labels
/// - Collects metadata when feature is enabled
/// - Supports visitor callbacks for custom processing
///
/// # Link Label Extraction
/// For links with block-level content, extracts text separately.
/// Collapses newlines and normalizes whitespace per Markdown spec.
///
/// # Autolinks
/// When `autolinks` option is enabled, detects links where the text equals
/// the href (e.g., `<a href="https://example.com">https://example.com</a>`)
/// and outputs as `<https://example.com>` instead.
///
/// # Note
/// This function references helper functions from converter.rs
/// which must be accessible (pub(crate)) for this module to work correctly.
pub fn handle(
    node_handle: &NodeHandle,
    parser: &Parser,
    output: &mut String,
    options: &ConversionOptions,
    ctx: &Context,
    depth: usize,
    dom_ctx: &DomContext,
) {
    use crate::converter::block::heading::{heading_allows_inline_images, push_heading};
    use crate::converter::utility::content::normalized_tag_name;
    // ~keep reason: serialize_node is only used when the visitor feature is active.
    #[allow(unused_imports)]
    use crate::converter::utility::serialization::serialize_node;
    use crate::converter::{find_single_heading_child, get_text_content, walk_node};

    let Some(node) = node_handle.get(parser) else {
        return;
    };

    let tl::Node::Tag(tag) = node else {
        return;
    };

    let href_attr = tag.attributes().get("href").flatten().map(|v| {
        let decoded = crate::text::decode_html_entities(&v.as_utf8_str());
        sanitize_markdown_url(&decoded).into_owned()
    });
    // ~keep Hold title as a Cow borrowed from the tag's attribute bytes (Tier-2
    // ~keep hot-spot pass III): avoids per-link String allocation when no entity
    // ~keep decoding is needed.  Consumers below use `.as_deref()` (`Option<&str>`),
    // ~keep which works the same for either Cow variant.
    let title = tag.attributes().get("title").flatten().map(|v| v.as_utf8_str());

    if let Some(href) = href_attr {
        if ctx.in_link {
            let children = tag.children();
            for child_handle in children.top().iter() {
                walk_node(child_handle, parser, output, options, ctx, depth + 1, dom_ctx);
            }
            return;
        }

        let owned_children: Vec<tl::NodeHandle>;
        let children: &[tl::NodeHandle] = if let Some(c) = dom_ctx.children_of(node_handle.get_inner()) {
            c.as_slice()
        } else {
            owned_children = tag.children().top().iter().copied().collect();
            owned_children.as_slice()
        };
        let (inline_label, _block_nodes, saw_block) = collect_link_label_text(children, parser, dom_ctx);

        // ~keep Without block descendants the sweep above already visited exactly the nodes
        // ~keep `get_text_content` would and decoded them the same way, so its text is reused
        // ~keep rather than walking the `<a>` subtree a second time.
        let text_source: Cow<'_, str> = if saw_block {
            Cow::Owned(get_text_content(node_handle, parser, dom_ctx))
        } else {
            Cow::Borrowed(inline_label.as_str())
        };
        let normalized_text = crate::text::normalize_whitespace_cow(text_source.as_ref());
        let raw_text = normalized_text.trim();

        // ~keep Check if this should be rendered as an autolink.
        // ~keep GFM requires an absolute URI with a scheme (e.g. `https://…`, `mailto:…`);
        // ~keep bare paths or filenames must use the full `[text](href)` form.
        let is_autolink = options.autolinks
            && !options.default_title
            && !href.is_empty()
            && has_uri_scheme(href.as_str())
            && (raw_text == href || (href.starts_with("mailto:") && raw_text == &href[7..]));

        if is_autolink {
            output.push('<');
            if href.starts_with("mailto:") && raw_text == &href[7..] {
                output.push_str(raw_text);
            } else {
                output.push_str(&href);
            }
            output.push('>');
            return;
        }

        if let Some((heading_level, heading_handle)) = find_single_heading_child(*node_handle, parser) {
            if let Some(heading_node) = heading_handle.get(parser) {
                if let tl::Node::Tag(heading_tag) = heading_node {
                    let heading_name = normalized_tag_name(heading_tag.name().as_utf8_str()).into_owned();
                    let mut heading_text = String::new();
                    let heading_ctx = Context {
                        in_heading: true,
                        convert_as_inline: true,
                        heading_allow_inline_images: heading_allows_inline_images(
                            &heading_name,
                            &ctx.keep_inline_images_in,
                        ),
                        ..ctx.clone()
                    };
                    walk_node(
                        &heading_handle,
                        parser,
                        &mut heading_text,
                        options,
                        &heading_ctx,
                        depth + 1,
                        dom_ctx,
                    );
                    let trimmed_heading = heading_text.trim();
                    if !trimmed_heading.is_empty() {
                        let escaped_label = escape_link_label(trimmed_heading);
                        let mut link_buffer = String::new();
                        append_markdown_link(
                            &mut link_buffer,
                            &escaped_label,
                            href.as_str(),
                            title.as_deref(),
                            raw_text,
                            options,
                            ctx.reference_collector.as_ref(),
                        );
                        push_heading(output, ctx, options, heading_level, link_buffer.as_str());
                        return;
                    }
                }
            }
        }

        let mut label = if saw_block {
            let mut content = String::new();
            let link_ctx = Context {
                inline_depth: ctx.inline_depth + 1,
                convert_as_inline: true,
                in_link: true,
                ..ctx.clone()
            };
            for child_handle in children {
                let mut child_buf = String::new();
                walk_node(
                    child_handle,
                    parser,
                    &mut child_buf,
                    options,
                    &link_ctx,
                    depth + 1,
                    dom_ctx,
                );
                if !child_buf.trim().is_empty()
                    && !content.is_empty()
                    && !content.chars().last().is_none_or(char::is_whitespace)
                    && !child_buf.chars().next().is_none_or(char::is_whitespace)
                {
                    content.push(' ');
                }
                content.push_str(&child_buf);
            }
            if content.trim().is_empty() {
                normalize_link_label(&inline_label)
            } else {
                normalize_link_label(&content)
            }
        } else {
            let mut content = String::new();
            let link_ctx = Context {
                inline_depth: ctx.inline_depth + 1,
                in_link: true,
                ..ctx.clone()
            };
            for child_handle in children {
                walk_node(
                    child_handle,
                    parser,
                    &mut content,
                    options,
                    &link_ctx,
                    depth + 1,
                    dom_ctx,
                );
            }
            normalize_link_label(&content)
        };

        // ~keep `raw_text` is already the whole-subtree text when `saw_block`, so this single
        // ~keep fallback covers both the block and inline cases.
        if label.is_empty() && !raw_text.is_empty() {
            label = normalize_link_label(raw_text);
        }

        if label.is_empty() && !href.is_empty() && !children.is_empty() {
            label.clone_from(&href);
        }

        let escaped_label = escape_link_label(&label);

        #[cfg(feature = "visitor")]
        if let Some(ref visitor_handle) = ctx.visitor {
            use crate::visitor::{NodeContext, NodeType, VisitResult};

            let node_id = node_handle.get_inner();
            let parent_tag = dom_ctx.parent_tag_name(node_id, parser);
            let index_in_parent = dom_ctx.get_sibling_index(node_id).unwrap_or(0);

            let node_ctx = NodeContext::with_lazy_attributes(
                NodeType::Link,
                Cow::Borrowed("a"),
                tag,
                depth,
                index_in_parent,
                parent_tag.map(Cow::Borrowed),
                true,
            );

            let visit_result = {
                let mut visitor = visitor_handle.lock().expect("visitor mutex poisoned");
                visitor.visit_link(&node_ctx, &href, &label, title.as_deref())
            };
            match visit_result {
                VisitResult::Continue => append_markdown_link(
                    output,
                    &escaped_label,
                    href.as_str(),
                    title.as_deref(),
                    label.as_str(),
                    options,
                    ctx.reference_collector.as_ref(),
                ),
                VisitResult::Custom(custom) => output.push_str(&custom),
                VisitResult::Skip => {}
                VisitResult::Error(err) => {
                    if ctx.visitor_error.borrow().is_none() {
                        *ctx.visitor_error.borrow_mut() = Some(err);
                    }
                }
                VisitResult::PreserveHtml => output.push_str(&serialize_node(node_handle, parser)),
            }
        } else {
            append_markdown_link(
                output,
                &escaped_label,
                href.as_str(),
                title.as_deref(),
                label.as_str(),
                options,
                ctx.reference_collector.as_ref(),
            );
        }

        #[cfg(not(feature = "visitor"))]
        append_markdown_link(
            output,
            &escaped_label,
            href.as_str(),
            title.as_deref(),
            label.as_str(),
            options,
            ctx.reference_collector.as_ref(),
        );

        #[cfg(feature = "metadata")]
        if ctx.metadata_wants_links {
            if let Some(ref collector) = ctx.metadata_collector {
                let rel_attr = tag
                    .attributes()
                    .get("rel")
                    .flatten()
                    .map(|v| v.as_utf8_str().to_string());
                let mut attributes_map = BTreeMap::new();
                for (key, value_opt) in tag.attributes().iter() {
                    let key_str = key.to_string();
                    if key_str == "href" {
                        continue;
                    }

                    let value = value_opt.map(|v| v.to_string()).unwrap_or_default();
                    attributes_map.insert(key_str, value);
                }
                collector.borrow_mut().add_link(
                    href.clone(),
                    label,
                    title.as_deref().map(str::to_string),
                    rel_attr,
                    attributes_map,
                );
            }
        }
    } else {
        let children = tag.children();
        for child_handle in children.top().iter() {
            walk_node(child_handle, parser, output, options, ctx, depth + 1, dom_ctx);
        }
    }
}

/// Check whether `href` begins with a syntactically valid RFC 3986 URI scheme.
///
/// A scheme matches `ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )` followed by `:`.
/// Bare paths and filenames (e.g. `foobar.png`) fail this check and must be rendered
/// as `[text](href)` rather than as autolinks per GFM §6.5.
#[must_use]
pub fn has_uri_scheme(href: &str) -> bool {
    let mut bytes = href.bytes();
    match bytes.next() {
        Some(b) if b.is_ascii_alphabetic() => {}
        _ => return false,
    }
    for b in bytes {
        match b {
            b':' => return true,
            b'A'..=b'Z' | b'a'..=b'z' | b'0'..=b'9' | b'+' | b'-' | b'.' => {}
            _ => return false,
        }
    }
    false
}

/// Percent-encode a URL destination.
///
/// Encodes every character that is not an RFC 3986 unreserved character (`A-Z`, `a-z`, `0-9`,
/// `-`, `_`, `.`, `~`) or a forward slash (`/`). This produces a destination that all
/// Markdown parsers handle correctly even when the original URL contains `<`, `>`, spaces,
/// or parentheses.
#[must_use]
pub fn percent_encode_url(url: &str) -> String {
    let mut encoded = String::with_capacity(url.len() * 2);
    for byte in url.bytes() {
        match byte {
            b'A'..=b'Z' | b'a'..=b'z' | b'0'..=b'9' | b'-' | b'_' | b'.' | b'~' | b'/' => {
                encoded.push(byte as char);
            }
            other => {
                encoded.push('%');
                let hi = char::from_digit(u32::from(other >> 4), 16)
                    .unwrap_or('0')
                    .to_ascii_uppercase();
                let lo = char::from_digit(u32::from(other & 0x0f), 16)
                    .unwrap_or('0')
                    .to_ascii_uppercase();
                encoded.push(hi);
                encoded.push(lo);
            }
        }
    }
    encoded
}

/// Check whether every `)` in `href` is matched by a preceding `(`, and every `(` is closed.
///
/// A raw (non-bracketed) Markdown link destination may contain parentheses only if they form a
/// properly nested, balanced pair — a plain count of `(` versus `)` is not sufficient, since e.g.
/// `")("` has equal counts but is not balanced (CommonMark 6.3). Unbalanced parentheses must be
/// backslash-escaped or the destination must be wrapped in angle brackets.
#[must_use]
fn parens_are_balanced(href: &str) -> bool {
    let mut depth: i32 = 0;
    for c in href.chars() {
        match c {
            '(' => depth += 1,
            ')' => {
                depth -= 1;
                if depth < 0 {
                    return false;
                }
            }
            _ => {}
        }
    }
    depth == 0
}

/// Escape a Markdown title's backslashes and double quotes for interpolation into a
/// double-quoted title `"..."`.
///
/// Backslashes are escaped *before* quotes: a title ending in a literal `\` would otherwise
/// make the following delimiter's `\"` read as an escaped quote instead of the closing
/// delimiter, letting the title (and the destination that follows) run into whatever content
/// comes next in the document.
#[must_use]
pub fn escape_markdown_title(text: &str) -> std::borrow::Cow<'_, str> {
    if !text.contains('\\') && !text.contains('"') {
        return std::borrow::Cow::Borrowed(text);
    }
    std::borrow::Cow::Owned(text.replace('\\', "\\\\").replace('"', "\\\""))
}

/// Append a Markdown link destination (the `(...)` portion, without the enclosing parens) to
/// `output`, honoring `url_escape_style`.
///
/// Shared by [`append_markdown_link`] (for `<a href>`) and the image/graphic handlers, so a
/// destination gets the same treatment — empty-destination handling, percent-encoding,
/// space-triggered angle-bracket wrapping with backslash-safe `<`/`>` escaping inside it, and
/// paren-balance escaping — no matter which element produced it.
pub fn append_url_destination(
    output: &mut String,
    dest: &str,
    url_escape_style: crate::options::validation::UrlEscapeStyle,
) {
    if dest.is_empty() {
        output.push_str("<>");
    } else if url_escape_style == crate::options::validation::UrlEscapeStyle::Percent {
        let encoded = percent_encode_url(dest);
        output.push_str(&encoded);
    } else if dest.contains(' ') || dest.contains('\n') {
        // ~keep angle-bracket destinations may contain raw parentheses, but a raw `<`, `>`, or
        // ~keep an unescaped `\` (which would otherwise merge with the next escaped char and
        // ~keep un-escape it) terminates the wrap early, so all three must be escaped inside it
        // ~keep (CommonMark 6.3).
        output.push('<');
        for c in dest.chars() {
            match c {
                '\\' => output.push_str("\\\\"),
                '<' => output.push_str("\\<"),
                '>' => output.push_str("\\>"),
                other => output.push(other),
            }
        }
        output.push('>');
    } else if parens_are_balanced(dest) {
        output.push_str(dest);
    } else {
        let escaped_dest = dest.replace('(', "\\(").replace(')', "\\)");
        output.push_str(&escaped_dest);
    }
}

/// Format and append a Markdown link to the output string.
///
/// Generates the link syntax: `[label](href "title")`
/// Handles special cases:
/// - Empty href renders as `[label]()`
/// - With `UrlEscapeStyle::Angle` (default): hrefs with spaces/newlines get wrapped in angle
///   brackets: `[label](<URL with spaces>)`
/// - With `UrlEscapeStyle::Percent`: every non-unreserved character is percent-encoded
/// - Unbalanced parentheses in href get escaped when using `Angle` style
/// - Titles are wrapped in quotes and quotes inside are escaped
/// - When `default_title` option is true and `raw_text` equals href, adds href as title
///
/// # Arguments
/// * `output` - Output buffer to append the link to
/// * `label` - The link text (already escaped)
/// * `href` - The URL/destination
/// * `title` - Optional link title attribute
/// * `raw_text` - Original unprocessed text (for `default_title` option)
/// * `options` - Conversion options
pub fn append_markdown_link(
    output: &mut String,
    label: &str,
    href: &str,
    title: Option<&str>,
    raw_text: &str,
    options: &ConversionOptions,
    reference_collector: Option<&crate::converter::reference_collector::ReferenceCollectorHandle>,
) {
    if options.link_style == crate::options::validation::LinkStyle::Reference && !href.is_empty() {
        if let Some(collector) = reference_collector {
            let ref_num = collector.borrow_mut().get_or_insert(href, title);
            output.push('[');
            output.push_str(label);
            output.push_str("][");
            output.push_str(&ref_num.to_string());
            output.push(']');
            return;
        }
    }

    output.push('[');
    output.push_str(label);
    output.push_str("](");

    append_url_destination(output, href, options.url_escape_style);

    if let Some(title_text) = title {
        output.push_str(" \"");
        output.push_str(&escape_markdown_title(title_text));
        output.push('"');
    } else if options.default_title && raw_text == href {
        output.push_str(" \"");
        output.push_str(&escape_markdown_title(href));
        output.push('"');
    }

    output.push(')');
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::options::validation::UrlEscapeStyle;

    fn opts_with_style(style: UrlEscapeStyle) -> ConversionOptions {
        ConversionOptions::builder().url_escape_style(style).build()
    }

    #[test]
    fn has_uri_scheme_accepts_http() {
        assert!(has_uri_scheme("http://example.com"));
        assert!(has_uri_scheme("https://example.com/path"));
    }

    #[test]
    fn has_uri_scheme_accepts_mailto() {
        assert!(has_uri_scheme("mailto:a@b.com"));
    }

    #[test]
    fn has_uri_scheme_accepts_uncommon_schemes() {
        assert!(has_uri_scheme("ftp://host"));
        assert!(has_uri_scheme("ssh://host"));
        assert!(has_uri_scheme("data:text/plain,foo"));
        assert!(has_uri_scheme("file:///etc/hosts"));
    }

    #[test]
    fn has_uri_scheme_rejects_bare_paths() {
        assert!(!has_uri_scheme("foobar.png"));
        assert!(!has_uri_scheme("/relative/path"));
        assert!(!has_uri_scheme("../up.html"));
        assert!(!has_uri_scheme("#fragment"));
    }

    #[test]
    fn has_uri_scheme_rejects_leading_digit_or_punct() {
        assert!(!has_uri_scheme("9scheme:foo"));
        assert!(!has_uri_scheme(":no-scheme"));
        assert!(!has_uri_scheme(""));
    }

    #[test]
    fn issue_397_filename_with_extension_is_not_autolinked() {
        assert!(!has_uri_scheme("foobar.png"));
    }

    #[test]
    fn percent_encode_url_leaves_unreserved_chars_unchanged() {
        let result = percent_encode_url("/path-to_file.html~");
        assert_eq!(result, "/path-to_file.html~");
    }

    #[test]
    fn percent_encode_url_encodes_spaces() {
        assert_eq!(percent_encode_url("/file (1).pdf"), "/file%20%281%29.pdf");
    }

    #[test]
    fn percent_encode_url_encodes_angle_brackets() {
        assert_eq!(percent_encode_url("/file <draft>.pdf"), "/file%20%3Cdraft%3E.pdf");
    }

    #[test]
    fn percent_encode_url_full_issue_example() {
        assert_eq!(
            percent_encode_url("/file (1) <draft>.pdf"),
            "/file%20%281%29%20%3Cdraft%3E.pdf"
        );
    }

    #[test]
    fn append_markdown_link_angle_plain_url_unchanged() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(&mut out, "text", "/file.pdf", None, "text", &options, None);
        assert_eq!(out, "[text](/file.pdf)");
    }

    #[test]
    fn append_markdown_link_angle_wraps_space_in_angle_brackets() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(&mut out, "file", "/file (1).pdf", None, "file", &options, None);
        assert_eq!(out, "[file](</file (1).pdf>)");
    }

    #[test]
    fn append_markdown_link_percent_encodes_spaces() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Percent);
        append_markdown_link(&mut out, "file", "/file (1).pdf", None, "file", &options, None);
        assert_eq!(out, "[file](/file%20%281%29.pdf)");
    }

    #[test]
    fn append_markdown_link_percent_encodes_angle_brackets() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Percent);
        append_markdown_link(&mut out, "file", "/file <draft>.pdf", None, "file", &options, None);
        assert_eq!(out, "[file](/file%20%3Cdraft%3E.pdf)");
    }

    #[test]
    fn append_markdown_link_percent_full_issue_example() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Percent);
        append_markdown_link(&mut out, "file", "/file (1) <draft>.pdf", None, "file", &options, None);
        assert_eq!(out, "[file](/file%20%281%29%20%3Cdraft%3E.pdf)");
    }

    #[test]
    fn parens_are_balanced_accepts_nested_parens() {
        assert!(parens_are_balanced("wiki/Rust_(programming_language)"));
        assert!(parens_are_balanced("no/parens/here"));
    }

    #[test]
    fn parens_are_balanced_rejects_equal_counts_out_of_order() {
        // ~keep equal open/close counts are not sufficient for balance: a `)` before its `(`
        // ~keep is the exact naive-count bug this check replaces.
        assert!(!parens_are_balanced("a)b(c"));
    }

    #[test]
    fn parens_are_balanced_rejects_unmatched_open_or_close() {
        assert!(!parens_are_balanced("a(b"));
        assert!(!parens_are_balanced("a)b"));
    }

    #[test]
    fn append_markdown_link_angle_leaves_balanced_parens_unescaped_when_href_has_parens() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(
            &mut out,
            "Rust",
            "https://en.wikipedia.org/wiki/Rust_(programming_language)",
            None,
            "Rust",
            &options,
            None,
        );
        assert_eq!(out, "[Rust](https://en.wikipedia.org/wiki/Rust_(programming_language))");
    }

    #[test]
    fn append_markdown_link_angle_escapes_out_of_order_parens_when_href_has_parens() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(
            &mut out,
            "link",
            "http://example.com/a)(b",
            None,
            "link",
            &options,
            None,
        );
        assert_eq!(out, "[link](http://example.com/a\\)\\(b)");
    }

    #[test]
    fn append_markdown_link_angle_escapes_gt_inside_wrap_when_href_has_space_and_gt() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(&mut out, "text", "/my file >.pdf", None, "text", &options, None);
        assert_eq!(out, "[text](</my file \\>.pdf>)");
    }

    #[test]
    fn append_markdown_link_angle_produces_empty_angle_brackets_when_href_is_empty() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(&mut out, "text", "", None, "text", &options, None);
        assert_eq!(out, "[text](<>)");
    }

    #[test]
    fn append_markdown_link_percent_preserves_title() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Percent);
        append_markdown_link(
            &mut out,
            "link",
            "/path with spaces",
            Some("My Title"),
            "link",
            &options,
            None,
        );
        assert_eq!(out, "[link](/path%20with%20spaces \"My Title\")");
    }

    #[test]
    fn escape_markdown_title_escapes_backslash_before_quote_so_the_closing_quote_is_not_swallowed() {
        // ~keep audit #24 finding 8: a title ending in a literal `\` must not let a following `\"`
        // (backslash escaping the delimiter's quote) read as an escaped quote instead of the
        // closing delimiter.
        assert_eq!(escape_markdown_title("foo\\"), "foo\\\\");
        assert_eq!(escape_markdown_title("say \"hi\"\\"), "say \\\"hi\\\"\\\\");
    }

    #[test]
    fn append_markdown_link_escapes_a_trailing_backslash_in_title_so_the_closing_quote_is_not_swallowed() {
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(&mut out, "text", "/url", Some("foo\\"), "text", &options, None);
        assert_eq!(out, "[text](/url \"foo\\\\\")");
    }

    #[test]
    fn append_markdown_link_escapes_a_trailing_backslash_in_default_title_so_the_closing_quote_is_not_swallowed() {
        let mut out = String::new();
        let mut options = opts_with_style(UrlEscapeStyle::Angle);
        options.default_title = true;
        append_markdown_link(&mut out, "text", "http://a\\", None, "http://a\\", &options, None);
        assert_eq!(out, "[text](http://a\\ \"http://a\\\\\")");
    }

    #[test]
    fn append_markdown_link_escapes_a_backslash_inside_the_angle_bracket_wrap_so_it_cannot_unescape_a_delimiter() {
        // ~keep audit #24 finding 8: inside an angle-bracket-wrapped destination, an unescaped `\`
        // immediately before an escaped `<`/`>` merges with it into a single `\\` escape pair,
        // un-escaping the delimiter and terminating the destination early.
        let mut out = String::new();
        let options = opts_with_style(UrlEscapeStyle::Angle);
        append_markdown_link(&mut out, "text", "/my file\\>.pdf", None, "text", &options, None);
        assert_eq!(out, "[text](</my file\\\\\\>.pdf>)");
    }
}