securemark 0.290.2 → 0.291.0

package/src/parser/inline/media.test.ts

@@ -103,9 +103,9 @@ describe('Unit: parser/inline/media', () => {
     });
 
     it('attribute', () => {
-      assert.deepStrictEqual(inspect(parser('![]{/ __proto__}')), [['<a href="/" target="_blank"><img class="media invalid" data-src="/" alt="/"></a>'], '']);
-      assert.deepStrictEqual(inspect(parser('![]{/ constructor}')), [['<a href="/" target="_blank"><img class="media invalid" data-src="/" alt="/"></a>'], '']);
-      assert.deepStrictEqual(inspect(parser('![]{/ aspect-ratio}')), [['<a href="/" target="_blank"><img class="media invalid" data-src="/" alt="/"></a>'], '']);
+      assert.deepStrictEqual(inspect(parser('![]{/ __proto__}')), [['<a href="/" target="_blank"><img class="invalid" data-src="/" alt="/"></a>'], '']);
+      assert.deepStrictEqual(inspect(parser('![]{/ constructor}')), [['<a href="/" target="_blank"><img class="invalid" data-src="/" alt="/"></a>'], '']);
+      assert.deepStrictEqual(inspect(parser('![]{/ aspect-ratio}')), [['<a href="/" target="_blank"><img class="invalid" data-src="/" alt="/"></a>'], '']);
       assert.deepStrictEqual(inspect(parser('![]{/ nofollow}')), [['<a href="/" rel="nofollow" target="_blank"><img class="media" data-src="/" alt="/"></a>'], '']);
       assert.deepStrictEqual(inspect(parser('![]{/ width="4" height="3"}')), [['<a href="/" target="_blank"><img class="media" data-src="/" alt="/" width="4" height="3"></a>'], '']);
       assert.deepStrictEqual(inspect(parser('![]{/ 4x3}')), [['<a href="/" target="_blank"><img class="media" data-src="/" alt="/" width="4" height="3"></a>'], '']);

package/src/parser/inline/media.ts

@@ -7,7 +7,6 @@ import { unsafehtmlentity } from './htmlentity';
 import { txt, linebreak, str } from '../source';
 import { invalid } from '../util';
 import { ReadonlyURL } from 'spica/url';
-import { push } from 'spica/array';
 import { html, define } from 'typed-dom/dom';
 
 const optspec = {
@@ -18,7 +17,7 @@ const optspec = {
 } as const;
 Object.setPrototypeOf(optspec, null);
 
-export const media: MediaParser = lazy(() => constraint(State.media, false, validate(['![', '!{'], creation(10, open(
+export const media: MediaParser = lazy(() => constraint(State.media, validate(['![', '!{'], creation(10, open(
   '!',
   bind(verify(fmap(tails([
     dup(surround(
@@ -31,7 +30,7 @@ export const media: MediaParser = lazy(() => constraint(State.media, false, vali
       ']',
       true,
       ([, ns = []], rest, context) =>
-        context.linebreak === undefined
+        context.linebreak === 0
           ? [ns, rest]
           : undefined,
       undefined,
@@ -68,7 +67,7 @@ export const media: MediaParser = lazy(() => constraint(State.media, false, vali
     el.setAttribute('alt', text);
     if (!sanitize(el, uri)) return [[el], rest];
     assert(!el.matches('.invalid'));
-    define(el, attributes('media', push([], el.classList), optspec, params));
+    define(el, attributes('media', optspec, params));
     assert(el.matches('img') || !el.matches('.invalid'));
     // Awaiting the generic support for attr().
     if (el.hasAttribute('aspect-ratio')) {
@@ -99,8 +98,17 @@ const bracket: MediaParser.TextParser.BracketParser = lazy(() => recursion(Recur
 ])));
 
 const option: MediaParser.ParameterParser.OptionParser = lazy(() => union([
-  fmap(str(/^[^\S\n]+[1-9][0-9]*x[1-9][0-9]*(?=[^\S\n]|})/), ([opt]) => [` width="${opt.slice(1).split('x')[0]}"`, ` height="${opt.slice(1).split('x')[1]}"`]),
-  fmap(str(/^[^\S\n]+[1-9][0-9]*:[1-9][0-9]*(?=[^\S\n]|})/), ([opt]) => [` aspect-ratio="${opt.slice(1).split(':').join('/')}"`]),
+  surround(
+    open(/^[^\S\n]+/, str(/^[1-9][0-9]*/)),
+    str(/^[x:]/),
+    str(/^[1-9][0-9]*(?=[^\S\n]|})/),
+    false,
+    ([[a], [b], [c]], rest) => [
+      b === 'x'
+        ? [`width="${a}"`, `height="${c}"`]
+        : [`aspect-ratio="${a}/${c}"`],
+      rest
+    ]),
   linkoption,
 ]));
 

package/src/parser/inline/reference.ts

@@ -1,30 +1,87 @@
 import { ReferenceParser } from '../inline';
 import { State, Backtrack, Command } from '../context';
-import { union, subsequence, some, precedence, state, constraint, surround, setBacktrack, lazy } from '../../combinator';
+import { eval, exec } from '../../combinator/data/parser';
+import { union, subsequence, some, precedence, state, constraint, surround, isBacktrack, setBacktrack, lazy } from '../../combinator';
 import { inline } from '../inline';
+import { textlink } from './link';
 import { str } from '../source';
 import { blank, trimBlankStart, trimBlankNodeEnd } from '../visibility';
 import { html, defrag } from 'typed-dom/dom';
-import { unshift } from 'spica/array';
+import { unshift, push } from 'spica/array';
 import { invalid } from '../util';
 
-export const reference: ReferenceParser = lazy(() => constraint(State.reference, false, surround(
+export const reference: ReferenceParser = lazy(() => constraint(State.reference, surround(
   str('[['),
-  precedence(1, state(State.annotation | State.reference | State.media,
+  precedence(1, state(State.annotation | State.reference,
   subsequence([
     abbr,
     trimBlankStart(some(inline, ']', [[']', 1]])),
   ]))),
   ']]',
   false,
-  ([, ns], rest, context) =>
-    context.linebreak === undefined &&
-    trimBlankNodeEnd(ns).length > 0
-      ? [[html('sup', attributes(ns), [html('span', defrag(ns))])], rest]
-      : undefined,
+  ([, ns], rest, context) => {
+    if (context.linebreak === 0) {
+      return [[html('sup', attributes(ns), [html('span', defrag(trimBlankNodeEnd(ns)))])], rest];
+    }
+    else {
+      const head = context.recent!.reduce((a, b) => a + b.length, rest.length);
+      setBacktrack(context, [2 | Backtrack.link], head, 2);
+    }
+  },
   ([as, bs], rest, context) => {
+    const { recent } = context;
+    const head = recent!.reduce((a, b) => a + b.length, rest.length);
     if (rest[0] !== ']') {
-      setBacktrack(context, [2 | Backtrack.bracket], context.recent!.reduce((a, b) => a + b.length, 0), 2);
+      setBacktrack(context, [2 | Backtrack.bracket], head, 2);
+    }
+    else if (context.linebreak! > 0) {
+      setBacktrack(context, [2 | Backtrack.link], head, 2);
+    }
+    else {
+      assert(rest[0] === ']');
+      if (context.state! & State.annotation) {
+        push(bs, [rest[0]]);
+      }
+      const source = rest.slice(1);
+      let result: ReturnType<typeof textlink>;
+      if (source[0] !== '{') {
+        setBacktrack(context, [2 | Backtrack.link], head - 1);
+        result = [[], source];
+      }
+      else {
+        assert(source.length > 0);
+        result = !isBacktrack(context, [1 | Backtrack.link], source)
+          ? textlink({ source, context })
+          : undefined;
+        context.recent = recent;
+        if (!result) {
+          setBacktrack(context, [2 | Backtrack.link], head - 1);
+          result = [[], source];
+        }
+      }
+      assert(result);
+      if (exec(result) === '') {
+        setBacktrack(context, [2 | Backtrack.link], head);
+      }
+      else {
+        assert(context.state! ^ State.link);
+        const next = surround(
+          '',
+          some(inline, ']', [[']', 1]]),
+          str(']'),
+          true,
+          ([, cs = [], ds], rest) =>
+            [push(cs, ds), rest],
+          ([, cs = []], rest) => {
+            setBacktrack(context, [2 | Backtrack.link], head);
+            return [cs, rest];
+          })
+          ({ source: exec(result), context });
+        if (context.state! & State.annotation && next) {
+          push(push(bs, eval(result)), eval(next));
+          rest = exec(next);
+        }
+      }
     }
     return context.state! & State.annotation
       ? [unshift(as, bs), rest]

package/src/parser/inline/ruby.ts

@@ -18,15 +18,16 @@ export const ruby: RubyParser = lazy(() => bind(
         return isTightNodeStart(ns) ? [ns, rest] : undefined;
       },
       undefined,
-      [3 | Backtrack.ruby | 1, Backtrack.link, 1 | Backtrack.bracket])),
+      [3 | Backtrack.ruby, 1 | Backtrack.bracket])),
     dup(surround(
       '(', rtext, ')',
       false, undefined, undefined,
-      [3 | Backtrack.ruby | 1, Backtrack.link, 1 | Backtrack.bracket])),
+      [3 | Backtrack.ruby, 1 | Backtrack.bracket])),
   ]),
   ([texts, rubies], rest, context) => {
     if (rubies === undefined) {
-      return void setBacktrack(context, [2 | Backtrack.ruby], context.recent!.reduce((a, b) => a + b.length, 0));
+      const head = context.recent!.reduce((a, b) => a + b.length, rest.length);
+      return void setBacktrack(context, [2 | Backtrack.ruby], head);
     }
     switch (true) {
       case rubies.length <= texts.length:
@@ -65,7 +66,7 @@ const rtext: RubyParser.TextParser = ({ source, context }) => {
   const acc = [''];
   let state = false;
   while (source !== '') {
-    if (!/^(?:\\[^\n]|[^\\[\](){}<>"\n])/.test(source)) break;
+    if (!/^(?:\\[^\n]|[^\\[\](){}<>"$#\n])/.test(source)) break;
     assert(source[0] !== '\n');
     switch (source[0]) {
       // @ts-expect-error
@@ -73,7 +74,7 @@ const rtext: RubyParser.TextParser = ({ source, context }) => {
         const result = unsafehtmlentity({ source, context });
         if (result) {
           acc[acc.length - 1] += eval(result)[0];
-          source = exec(result, source.slice(1));
+          source = exec(result) ?? source.slice(1);
           state ||= acc.at(-1)!.trimStart() !== '';
           continue;
         }

package/src/parser/inline/shortmedia.ts

@@ -5,7 +5,7 @@ import { url } from './autolink/url';
 import { media } from './media';
 import { linebreak } from '../source';
 
-export const shortmedia: ShortMediaParser = constraint(State.media, false, rewrite(
+export const shortmedia: ShortMediaParser = constraint(State.media, rewrite(
   open('!', url),
   convert(
     source => `!{ ${source.slice(1)} }`,

package/src/parser/inline.ts

@@ -1,5 +1,5 @@
 import { MarkdownParser } from '../../markdown';
-import { union, verify, lazy } from '../combinator';
+import { union, lazy } from '../combinator';
 import { annotation } from './inline/annotation';
 import { reference } from './inline/reference';
 import { template } from './inline/template';
@@ -47,12 +47,10 @@ export import ShortMediaParser = InlineParser.ShortMediaParser;
 export import BracketParser = InlineParser.BracketParser;
 export import AutolinkParser = InlineParser.AutolinkParser;
 
-export const inline: InlineParser = lazy(() => verify(union([
+export const inline: InlineParser = lazy(() => union([
   input => {
-    const { source, context } = input;
+    const { source } = input;
     if (source === '') return;
-    context.depth ??= 0;
-    ++context.depth;
     switch (source.slice(0, 2)) {
       case '((':
         return annotation(input);
@@ -106,20 +104,7 @@ export const inline: InlineParser = lazy(() => verify(union([
   bracket,
   autolink,
   text
-]), (_, rest, context) => {
-  --context.depth!;
-  assert([rest]);
-  // ヒープを効率的に削除可能な場合は削除する。
-  // ヒープサイズは括弧類など特定の構文が完成しなかった場合にしか増加しないため
-  // ブロックごとに平均数ノード以下となることから削除せずとも平均的にはあまり影響はない。
-  //if (context.depth === 0) {
-  //  const { backtracks } = context;
-  //  while (backtracks.peek()?.key! > rest.length) {
-  //    backtracks.extract();
-  //  }
-  //}
-  return true;
-})) as any;
+])) as any;
 
 export { indexee } from './inline/extension/indexee';
 export { indexer } from './inline/extension/indexer';

package/src/parser/segment.test.ts

@@ -1,15 +1,16 @@
 import { segment } from './segment';
+import { Command } from './context';
 
 describe('Unit: parser/segment', () => {
   describe('segment', () => {
     it('huge input', () => {
       const result = segment(`${'\n'.repeat(10 * 1000 ** 2)}`).next().value?.split('\n', 1)[0];
-      assert(result?.startsWith('\x07Too large input'));
+      assert(result?.startsWith(`${Command.Error}Too large input`));
     });
 
     it('huge segment', () => {
       const result = segment(`${'\n'.repeat(1000 ** 2 - 1)}`).next().value?.split('\n', 1)[0];
-      assert(result?.startsWith('\x07Too large segment'));
+      assert(result?.startsWith(`${Command.Error}Too large segment`));
     });
 
     it('basic', () => {

package/src/parser/segment.ts

@@ -18,8 +18,8 @@ const parser: SegmentParser = union([
   codeblock,
   mathblock,
   extension,
-  some(contentline, MAX_SEGMENT_SIZE * 2),
-  some(emptyline, MAX_SEGMENT_SIZE * 2),
+  some(contentline, MAX_SEGMENT_SIZE + 1),
+  some(emptyline, MAX_SEGMENT_SIZE + 1),
 ]);
 
 export function* segment(source: string): Generator<string, undefined, undefined> {

package/src/parser/source/escapable.ts

@@ -34,7 +34,7 @@ export const escsource: EscapableSourceParser = ({ source, context }) => {
               return [[source.slice(0, 2)], source.slice(2)];
           }
         case '\n':
-          context.linebreak ??= source.length;
+          context.linebreak ||= source.length;
           return [[html('br')], source.slice(1)];
         default:
           assert(source[0] !== '\n');

package/src/parser/source/text.ts

@@ -4,7 +4,7 @@ import { union, consume, focus } from '../../combinator';
 import { str } from './str';
 import { html } from 'typed-dom/dom';
 
-export const delimiter = /[\s\x00-\x7F（）［］｛｝“”‘’「」『』]|\S#|[0-9A-Za-z]>/u;
+export const delimiter = /[\s\x00-\x7F（）［］｛｝]|\S#|[0-9A-Za-z]>/u;
 export const nonWhitespace = /[\S\n]|$/u;
 export const nonAlphanumeric = /[^0-9A-Za-z]|\S#|[0-9A-Za-z]>|$/u;
 const repeat = str(/^(.)\1*/);
@@ -36,7 +36,7 @@ export const text: TextParser = ({ source, context }) => {
               return [[source.slice(1, 2)], source.slice(2)];
           }
         case '\n':
-          context.linebreak ??= source.length;
+          context.linebreak ||= source.length;
           return [[html('br')], source.slice(1)];
         case '*':
         case '`':

package/src/parser/source/unescapable.ts

@@ -22,7 +22,7 @@ export const unescsource: UnescapableSourceParser = ({ source, context }) => {
           consume(1, context);
           return [[source.slice(1, 2)], source.slice(2)];
         case '\n':
-          context.linebreak ??= source.length;
+          context.linebreak ||= source.length;
           return [[html('br')], source.slice(1)];
         default:
           assert(source[0] !== '\n');

package/src/parser/util.ts

@@ -4,12 +4,22 @@ import { Parser, Result, Ctx, Node, Context, eval, exec } from '../combinator/da
 import { convert } from '../combinator';
 import { define } from 'typed-dom/dom';
 
-export function lineable<P extends Parser<HTMLElement | string>>(parser: P, fillTrailingLinebreak?: boolean): P;
-export function lineable<N extends HTMLElement | string>(parser: Parser<N>, fillTrailingLinebreak = false): Parser<N> {
+export function lineable<P extends Parser<HTMLElement | string>>(parser: P, trim?: 0 | 1 | -1): P;
+export function lineable<N extends HTMLElement | string>(parser: Parser<N>, trim = -1): Parser<N> {
   return convert(
-    source => `\r${source}${fillTrailingLinebreak && source.at(-1) !== '\n' ? '\n' : ''}`,
+    source => `\r${
+      trim === 0
+        ? source
+        : trim > 0
+          ? source.at(-1) === '\n'
+            ? source
+            : source + '\n'
+          : source.at(-1) === '\n'
+            ? source.slice(0, -1)
+            : source
+    }`,
     parser,
-    !fillTrailingLinebreak);
+    trim === 0);
 }
 
 export function repeat<P extends Parser<HTMLElement | string>>(symbol: string, parser: P, cons: (nodes: Node<P>[], context: Context<P>) => Node<P>[], termination?: (acc: Node<P>[][], rest: string, prefix: number, postfix: number, state: boolean) => Result<string | Node<P>>): P;

package/src/parser/visibility.ts

@@ -51,6 +51,7 @@ export function blankWith(starts: '' | '\n', delimiter?: string | RegExp): RegEx
   return new RegExp(String.raw
     `^(?:(?=${starts})(?:\\?\s|&(?:${invisibleHTMLEntityNames.join('|')});|<wbr[^\S\n]*>)${
       // 空行除去
+      // 完全な空行はエスケープ済みなので再帰的バックトラックにはならない。
       starts && '+'
     })?${
       typeof delimiter === 'string'