new description list rendering

src/description-list.ts

@@ -3,14 +3,19 @@ import { marked } from 'marked';
 import { MarkdownOptions } from './render';
 
 export interface DescriptionListToken extends marked.Tokens.Generic {
+	type: 'description_list';
 	items: (DescriptionTermToken | DescriptionDetailToken)[];
 }
 
+export type DescriptionElemToken = DescriptionTermToken | DescriptionDetailToken;
+
 export interface DescriptionTermToken extends marked.Tokens.Generic {
+	type: 'description_term';
 	text: string;
 }
 
 export interface DescriptionDetailToken extends marked.Tokens.Generic {
+	type: 'description_detail';
 	text: string;
 }
 
@@ -18,72 +23,107 @@ export function description_list_ext(renderer: marked.Renderer, opts: MarkdownOp
 	return {
 		name: 'description_list',
 		level: 'block',
-		start: (src) => src.match(/^:[:#-]/)?.index,
+		start: (src) => src.match(/^: /)?.index,
 		tokenizer(src, tokens) {
-			const rule = /^(?::[:#-](?:\s[^\n]*)?(?:\n|$))+/;
-			const match = rule.exec(src);
+			const start = src.match(/^: /)?.index;
+			const lines = src.slice(start).split(/\n/g);
+			const token: DescriptionListToken = {
+				type: 'description_list',
+				raw: '',
+				items: [ ]
+			};
 
-			if (match) {
-				const token: DescriptionListToken = {
-					type: 'description_list',
-					raw: match[0],
-					items: [ ]
-				};
+			let current: DescriptionElemToken;
+			const render_current = () => {
+				if (current) {
+					this.lexer.blockTokens(current.text, current.tokens);
+					current = null;
+				}
+			};
 
-				const items = token.raw.trim().split('\n');
-				const raw_buffer: string[] = [ ];
-				const text_buffer: string[] = [ ];
+			for (const line of lines) {
+				// Skip empty lines
+				if (! line.trim()) {
+					token.raw += line + '\n';
 
-				const flush_buffer = () => {
-					if (! raw_buffer.length) {
-						return;
+					if (current) {
+						current.raw += line + '\n';
+						current.text += '\n';
 					}
 
-					// Grab the second character from the first line to determine the
-					// token type (should be "#" or "-")
-					const type = raw_buffer[0][1] === '#' ? 'description_term' : 'description_detail';
-					
-					const sub_token: (DescriptionTermToken | DescriptionDetailToken) = {
-						type,
-						raw: raw_buffer.join('\n'),
-						text: text_buffer.join('\n'),
-						tokens: [ ],
-					};
-
-					raw_buffer.length = 0;
-					text_buffer.length = 0;
-
-					this.lexer.blockTokens(sub_token.text, sub_token.tokens);
-					token.items.push(sub_token);
-				};
-
-				for (const line of items) {
-					const rule = /^:([:#-])(?:\s([^\n]*))?(?:\n|$)/;
-					const match = rule.exec(line);
-		
-					if (match) {
-						if (match[1] !== ':') {
-							flush_buffer();
-						}
-
-						raw_buffer.push(match[0]);
-						text_buffer.push(match[2]);
-					}
+					continue;
 				}
 
-				flush_buffer();
-				
+				// If the line starts immediately with a colon, it is a <dt>
+				if (line.startsWith(': ')) {
+					render_current();
+					token.raw += line + '\n';
+					token.items.push(
+						current = {
+							type: 'description_term',
+							raw: line,
+							text: line.slice(2),
+							tokens: [ ],
+						}
+					);
+					continue;
+				}
+
+				// If the line starts with a colon after an indent, it is a <dd>
+				if (line.startsWith('  : ')) {
+					render_current();
+					token.raw += line + '\n';
+					token.items.push(
+						current = {
+							type: 'description_detail',
+							raw: line,
+							text: line.slice(4),
+							tokens: [ ],
+						}
+					);
+					continue;
+				}
+
+				// If the line starts with (at least) two indents, it is a child
+				// of the current element
+				if (line.startsWith('    ')) {
+					token.raw += line + '\n';
+					current.raw += '\n' + line;
+					current.text += '\n' + line.slice(current.type === 'description_term' ? 2 : 4);
+					continue;
+				}
+
+				// If the line starts with one indent, it is a child of the current
+				// <dt> (but is not allowed after a <dd>)
+				if (line.startsWith('  ')) {
+					if (current.type !== 'description_term') {
+						render_current();
+						break;
+					}
+
+					token.raw += line + '\n';
+					current.raw += '\n' + line;
+					current.text += '\n' + line.slice(2);
+					continue;
+				}
+
+				// If the line starts any other way, it is the start of new content
+				// and we are done parsing
+				render_current();
+				break;
+			}
+			
+			render_current();
+
+			if (token.items.length) {
+				console.log(token);
 				return token;
 			}
 		},
 		renderer(token: DescriptionListToken) {
 			const items = token.items.map((item) => {
 				const tag = item.type === 'description_term' ? 'dt' : 'dd';
-				return `
-					<${tag}>
-						${this.parser.parse(item.tokens)}
-					</${tag}>
-				`;
+				return `<${tag}>${this.parser.parse(item.tokens)}</${tag}>`;
 			});
 
 			return `<dl>${items.join('')}</dl>`;