Fix more edge-case embed formatting

m2d/converters/emoji.js

@@ -40,6 +40,7 @@ function encodeEmoji(input, shortcode) {
 			"%E2%AD%90", // ⭐
 			"%F0%9F%90%88", // 🐈
 			"%E2%9D%93", // ❓
+			"%F0%9F%8F%86", // 🏆️
 		]
 
 		discordPreferredEncoding =

m2d/converters/event-to-message.js

@@ -110,6 +110,24 @@ turndownService.addRule("inlineLink", {
 	}
 })
 
+turndownService.addRule("listItem", {
+	filter: "li",
+	replacement: function (content, node, options) {
+		content = content
+		.replace(/^\n+/, "") // remove leading newlines
+		.replace(/\n+$/, "\n") // replace trailing newlines with just a single one
+		.replace(/\n/gm, "\n  ") // indent
+		var prefix = options.bulletListMarker + " "
+		var parent = node.parentNode
+		if (parent.nodeName === "OL") {
+			var start = parent.getAttribute("start")
+			var index = Array.prototype.indexOf.call(parent.children, node)
+			prefix = (start ? Number(start) + index : index + 1) + ".  "
+		}
+		return prefix + content + (node.nextSibling && !/\n$/.test(content) ? "\n" : "")
+	}
+})
+
 /** @type {string[]} SPRITE SHEET EMOJIS FEATURE: mxc urls for the currently processing message */
 let endOfMessageEmojis = []
 turndownService.addRule("emoji", {

m2d/converters/event-to-message.test.js

@@ -368,7 +368,7 @@ test("event2message: lists are bridged correctly", async t => {
 			messagesToEdit: [],
 			messagesToSend: [{
 				username: "cadence [they]",
-				content: "*   line one\n*   line two\n*   line three\n    *   nested one\n    *   nested two\n*   line four",
+				content: "* line one\n* line two\n* line three\n  * nested one\n  * nested two\n* line four",
 				avatar_url: undefined
 			}]
 		}
@@ -498,6 +498,47 @@ test("event2message: quotes have an appropriate amount of whitespace", async t =
 	)
 })
 
+test("event2message: lists have appropriate line breaks", async t => {
+	t.deepEqual(
+		await eventToMessage({
+			content: {
+				body: 'i am not certain what you mean by "already exists with as discord". my goals are\n' +
+					'* bridgeing specific channels with existing matrix rooms\n' +
+					'  * optionally maybe entire "servers"\n' +
+					'* offering the bridge as a public service ',
+				format: 'org.matrix.custom.html',
+				formatted_body: '<p>i am not certain what you mean by "already exists with as discord". my goals are</p>\n' +
+					'<ul>\n' +
+					'<li>bridgeing specific channels with existing matrix rooms\n' +
+					'<ul>\n' +
+					'<li>optionally maybe entire "servers"</li>\n' +
+					'</ul>\n' +
+					'</li>\n' +
+					'<li>offering the bridge as a public service</li>\n' +
+					'</ul>\n',
+				'm.mentions': {},
+				msgtype: 'm.text'
+			},
+			room_id: '!cBxtVRxDlZvSVhJXVK:cadence.moe',
+			sender: '@Milan:tchncs.de',
+			type: 'm.room.message',
+		}, {}, {
+			api: {
+				getStateEvent: async () => ({displayname: "Milan"})
+			}
+		}),
+		{
+			messagesToDelete: [],
+			messagesToEdit: [],
+			messagesToSend: [{
+				username: "Milan",
+				content: `i am not certain what you mean by "already exists with as discord". my goals are\n\n* bridgeing specific channels with existing matrix rooms\n  * optionally maybe entire "servers"\n* offering the bridge as a public service`,
+				avatar_url: undefined
+			}]
+		}
+	)
+})
+
 test("event2message: m.emote plaintext works", async t => {
 	t.deepEqual(
 		await eventToMessage({

m2d/converters/utils.js

@@ -16,6 +16,7 @@ const BLOCK_ELEMENTS = [
 	"NOSCRIPT", "OL", "OUTPUT", "P", "PRE", "SECTION", "SUMMARY", "TABLE", "TBODY", "TD",
 	"TFOOT", "TH", "THEAD", "TR", "UL"
 ]
+const NEWLINE_ELEMENTS = BLOCK_ELEMENTS.concat(["BR"])
 
 /**
  * Determine whether an event is the bridged representation of a discord message.
@@ -63,7 +64,71 @@ function getEventIDHash(eventID) {
 	return signedHash
 }
 
+class MatrixStringBuilder {
+	constructor() {
+		this.body = ""
+		this.formattedBody = ""
+	}
+
+	/**
+	 * @param {string} body
+	 * @param {string} formattedBody
+	 * @param {any} [condition]
+	 */
+	add(body, formattedBody, condition = true) {
+		if (condition) {
+			if (formattedBody == undefined) formattedBody = body
+			this.body += body
+			this.formattedBody += formattedBody
+		}
+		return this
+	}
+
+	/**
+	 * @param {string} body
+	 * @param {string} [formattedBody]
+	 * @param {any} [condition]
+	 */
+	addLine(body, formattedBody, condition = true) {
+		if (condition) {
+			if (formattedBody == undefined) formattedBody = body
+			if (this.body.length && this.body.slice(-1) !== "\n") this.body += "\n"
+			this.body += body
+			const match = this.formattedBody.match(/<\/?([a-zA-Z]+[a-zA-Z0-9]*)[^>]*>\s*$/)
+			if (this.formattedBody.length && (!match || !NEWLINE_ELEMENTS.includes(match[1].toUpperCase()))) this.formattedBody += "<br>"
+			this.formattedBody += formattedBody
+		}
+		return this
+	}
+
+	/**
+	 * @param {string} body
+	 * @param {string} [formattedBody]
+	 * @param {any} [condition]
+	 */
+	addParagraph(body, formattedBody, condition = true) {
+		if (condition) {
+			if (formattedBody == undefined) formattedBody = body
+			if (this.body.length && this.body.slice(-1) !== "\n") this.body += "\n\n"
+			this.body += body
+			formattedBody = `<p>${formattedBody}</p>`
+			this.formattedBody += formattedBody
+		}
+		return this
+	}
+
+	get() {
+		return {
+			msgtype: "m.text",
+			body: this.body,
+			format: "org.matrix.custom.html",
+			formatted_body: this.formattedBody
+		}
+	}
+}
+
 module.exports.BLOCK_ELEMENTS = BLOCK_ELEMENTS
 module.exports.eventSenderIsFromDiscord = eventSenderIsFromDiscord
 module.exports.getPublicUrlForMxc = getPublicUrlForMxc
 module.exports.getEventIDHash = getEventIDHash
+module.exports.MatrixStringBuilder = MatrixStringBuilder