Skip to content

AI-native knowledge base & Evernote/Obsidian alternative with instant sync & native MCP. Free on Cloudflare or self-hosted via Docker.

fix(extension): save the full text of a folded X post

tianma-if59b6ee2c7197
diff --git a/apps/extension/README.md b/apps/extension/README.md
index 778b3c27..99613233 100644
--- a/apps/extension/README.md
+++ b/apps/extension/README.md
@@ -9,7 +9,7 @@ Chrome, Edge, and Firefox Manifest V3 extension for saving the current webpage,
- Click the extension action to capture the selected content, or extract the article body with Mozilla Readability when there is no selection.
- Convert the extracted article HTML into Markdown with Turndown before uploading it.
- Right-click an image and choose **Save image to EdgeEver**. The image file is uploaded into a new note in the default notebook, with the source page linked underneath.
-- On X, right-click a post and choose **Save this post to EdgeEver**. The visible text, author, time, and photos already shown in that post are saved into one note. A single-post page can be saved immediately. On a timeline, the extension asks once for access to X so it can remember the post under the pointer.
+- On X, right-click a post and choose **Save this post to EdgeEver**. The post text, author, time, and photos already shown in that post are saved into one note. Text folded behind Show more is expanded first, so a long post is saved in full. A single-post page can be saved immediately. On a timeline, the extension asks once for access to X so it can remember the post under the pointer.
- When the page cannot hand over the image bytes, the extension asks once for access to that image's site. After that, later images from the same site save directly.
- Create a searchable EdgeEver memo with the source URL and a `web-clip` tag.
diff --git a/apps/extension/public/_locales/en/messages.json b/apps/extension/public/_locales/en/messages.json
index 84beeabc..22f690ed 100644
--- a/apps/extension/public/_locales/en/messages.json
+++ b/apps/extension/public/_locales/en/messages.json
@@ -259,7 +259,7 @@
"message": "Save this post to EdgeEver"
},
"tweetSaveHint": {
- "message": "On X, right-click a post and choose “Save this post to EdgeEver”. The visible text and photos go into one new note in the same default notebook."
+ "message": "On X, right-click a post and choose “Save this post to EdgeEver”. The post text, including text behind Show more, and the photos go into one new note in the same default notebook."
},
"savingTweet": {
"message": "Saving this post…"
diff --git a/apps/extension/public/_locales/ja/messages.json b/apps/extension/public/_locales/ja/messages.json
index 2e26801a..f19aa595 100644
--- a/apps/extension/public/_locales/ja/messages.json
+++ b/apps/extension/public/_locales/ja/messages.json
@@ -259,7 +259,7 @@
"message": "この投稿を EdgeEver に保存"
},
"tweetSaveHint": {
- "message": "X で投稿を右クリックし、「この投稿を EdgeEver に保存」を選びます。表示中の本文と画像は、上で選んだ既定のノートブックに 1 件のノートとして保存されます。"
+ "message": "X で投稿を右クリックし、「この投稿を EdgeEver に保存」を選びます。本文は「さらに表示」で折りたたまれた部分も含めて保存され、表示中の画像とともに、上で選んだ既定のノートブックへ 1 件のノートになります。"
},
"savingTweet": {
"message": "この投稿を保存しています…"
diff --git a/apps/extension/public/_locales/zh_CN/messages.json b/apps/extension/public/_locales/zh_CN/messages.json
index c7b183ef..1ebef711 100644
--- a/apps/extension/public/_locales/zh_CN/messages.json
+++ b/apps/extension/public/_locales/zh_CN/messages.json
@@ -259,7 +259,7 @@
"message": "保存这条推文到 EdgeEver"
},
"tweetSaveHint": {
- "message": "在 X 上右键一条推文,选择“保存这条推文到 EdgeEver”。已经显示的文字和图片会放进同一条新笔记,笔记本与上面的默认笔记本相同。"
+ "message": "在 X 上右键一条推文,选择“保存这条推文到 EdgeEver”。正文会完整保存,包括被「显示更多」折叠的部分;已经显示的图片也会放进同一条新笔记,笔记本与上面的默认笔记本相同。"
},
"savingTweet": {
"message": "正在保存这条推文……"
diff --git a/apps/extension/src/capture-tweet.ts b/apps/extension/src/capture-tweet.ts
index 4311b62c..e3732feb 100644
--- a/apps/extension/src/capture-tweet.ts
+++ b/apps/extension/src/capture-tweet.ts
@@ -17,6 +17,15 @@
const clean = (value: string) => value.replace(/\s+/g, " ").trim();
+ // Keep identical to visibleTweetAuthor in tweet-clip.ts.
+ const visibleTweetAuthor = (value: string) => {
+ const parts = value.replace(/\r/g, "").split("\n").map((part) => part.trim()).filter(Boolean);
+ const handlePart = parts.find((part) => /^@[A-Za-z0-9_]{1,15}$/.test(part));
+ const handle = handlePart ? handlePart.slice(1) : "";
+ const displayName = parts.find((part) => part !== handlePart && part !== "·" && part !== "•") ?? "";
+ return { displayName, handle };
+ };
+
const parseName = (rootNode: Element | null) => {
if (!rootNode) return { displayName: "", handle: "" };
const links = [...rootNode.querySelectorAll("a[href]")];
@@ -28,20 +37,100 @@
const text = clean(link.textContent || "");
return link !== handleLink && text.length > 0 && !text.startsWith("@");
});
- return { displayName: clean(nameLink?.textContent || ""), handle };
+ const displayName = clean(nameLink?.textContent || "");
+ if (displayName || handle) return { displayName, handle };
+ return visibleTweetAuthor((rootNode as HTMLElement).innerText || rootNode.textContent || "");
};
const hrefMatchesStatus = (href: string, id: string) => new RegExp(`/status/${id}(?:/|\\?|$)`).test(href);
- const ownStatusHref = (article: Element) => {
- const name = [...article.querySelectorAll('[data-testid="User-Name"]')].find((node) => !node.closest('[role="link"]'))
- ?? article.querySelector('[data-testid="User-Name"]');
- return name?.querySelector("time")?.closest("a")?.getAttribute("href") ?? "";
+ const quotedTime = (time: Element) => {
+ const name = time.closest('[data-testid="User-Name"]');
+ return Boolean(name?.closest('[role="link"]'));
+ };
+
+ const ownTime = (article: Element) => [...article.querySelectorAll("time")].find((time) => !quotedTime(time)) ?? null;
+
+ const ownStatusHref = (article: Element) => ownTime(article)?.closest("a")?.getAttribute("href") ?? "";
+
+ const findArticle = (preferredId: string) => {
+ const marked = document.querySelector('article[data-testid="tweet"][data-edgeever-tweet-target="1"]');
+ if (marked) return marked;
+ const id = preferredId || statusId;
+ if (!id) return null;
+ return [...document.querySelectorAll('article[data-testid="tweet"]')].find((article) => hrefMatchesStatus(ownStatusHref(article), id)) ?? null;
+ };
+
+ const isMainShowMore = (node: Element) => {
+ const card = node.closest('[role="link"]');
+ return !card || card === node;
+ };
+
+ const mainShowMore = (article: Element) => [...article.querySelectorAll('[data-testid="tweet-text-show-more-link"]')].find((node) => isMainShowMore(node)) ?? null;
+
+ const mainTextNode = (article: Element) => {
+ const texts = [...article.querySelectorAll('[data-testid="tweetText"]')];
+ return texts.find((node) => !node.closest('[role="link"]')) ?? texts[0] ?? null;
};
- const articleFromStatus = () => {
- if (!statusId) return null;
- return [...document.querySelectorAll('article[data-testid="tweet"]')].find((article) => hrefMatchesStatus(ownStatusHref(article), statusId)) ?? null;
+ const sleep = (ms: number) => new Promise((resolve) => setTimeout(resolve, ms));
+
+ const expandTruncatedText = async (article: Element) => {
+ const button = mainShowMore(article);
+ if (!button) return article;
+ const before = (mainTextNode(article)?.textContent || "").length;
+ const statusKey = ownStatusHref(article).match(/\/status\/(\d+)/)?.[1] ?? statusId;
+ (button as HTMLElement).click();
+ const started = Date.now();
+ const deadline = started + 4000;
+ let current = article;
+ while (Date.now() < deadline) {
+ await sleep(80);
+ const next = findArticle(statusKey) ?? (article.isConnected ? article : null);
+ if (!next) continue;
+ current = next;
+ const length = (mainTextNode(current)?.textContent || "").length;
+ const stillFolded = Boolean(mainShowMore(current));
+ if (!stillFolded && (length > before || Date.now() - started > 500)) break;
+ }
+ const photoDeadline = Date.now() + 800;
+ while (Date.now() < photoDeadline) {
+ const next = findArticle(statusKey) ?? (current.isConnected ? current : null);
+ if (next) current = next;
+ if (current.querySelector('[data-testid="tweetPhoto"] img, [data-testid="card.layoutLarge.media"] img')) break;
+ await sleep(80);
+ }
+ return current;
+ };
+
+ const visibleText = (node: Element | null) => {
+ if (!node) return "";
+ const parts: string[] = [];
+ const walk = (current: Node) => {
+ if (current.nodeType === Node.TEXT_NODE) {
+ parts.push(current.nodeValue || "");
+ return;
+ }
+ if (current.nodeType !== Node.ELEMENT_NODE) return;
+ const element = current as HTMLElement;
+ if (element.getAttribute("aria-hidden") === "true") return;
+ if (element.getAttribute("data-testid") === "tweet-text-show-more-link") return;
+ if (element.tagName === "BR") {
+ parts.push("\n");
+ return;
+ }
+ if (element.tagName === "IMG") {
+ const alt = element.getAttribute("alt");
+ if (alt) parts.push(alt);
+ return;
+ }
+ const block = element !== node && /^(block|flex|grid|list-item)$/.test(getComputedStyle(element).display);
+ if (block && parts.length > 0 && !parts[parts.length - 1]?.endsWith("\n")) parts.push("\n");
+ for (const child of element.childNodes) walk(child);
+ if (block && parts.length > 0 && !parts[parts.length - 1]?.endsWith("\n")) parts.push("\n");
+ };
+ walk(node);
+ return parts.join("").replace(/\u00a0/g, " ").replace(/\n{3,}/g, "\n\n").trim();
};
const isPhoto = (src: string) => {
@@ -71,14 +160,7 @@
}
};
- try {
- const marked = document.querySelector('article[data-testid="tweet"][data-edgeever-tweet-target="1"]');
- const article = marked ?? articleFromStatus();
- if (!article) {
- finish({ ok: false, reason: root.__edgeeverTweetTarget ? "not-found" : "needs-listener" });
- return;
- }
-
+ const readArticle = (article: Element) => {
const names = [...article.querySelectorAll('[data-testid="User-Name"]')];
const mainName = names.find((node) => !node.closest('[role="link"]')) ?? names[0] ?? null;
const quoteName = names.find((node) => node !== mainName && Boolean(node.closest('[role="link"]'))) ?? null;
@@ -98,32 +180,44 @@
images.push({ url: src, alt: clean(img.alt) });
if (images.length >= 6) break;
}
-
- const visibleText = (node: Element | null) => {
- if (!node) return "";
- return ((node as HTMLElement).innerText || node.textContent || "").trim();
+ return {
+ author,
+ quoted,
+ text: visibleText(mainText),
+ quotedText: visibleText(quoteText),
+ datetime: ownTime(article)?.getAttribute("datetime") ?? "",
+ statusUrl: absoluteStatus(ownStatusHref(article)),
+ images,
};
- const text = visibleText(mainText);
- const quotedText = visibleText(quoteText);
- const statusUrl = absoluteStatus(ownStatusHref(article));
- if (!author.displayName && !author.handle && !text && !quotedText && images.length === 0) {
+ };
+
+ void (async () => {
+ try {
+ const initial = findArticle(statusId);
+ if (!initial) {
+ finish({ ok: false, reason: root.__edgeeverTweetTarget ? "not-found" : "needs-listener" });
+ return;
+ }
+ const article = await expandTruncatedText(initial);
+ const read = readArticle(article);
+ if (!read.author.displayName && !read.author.handle && !read.text && !read.quotedText && read.images.length === 0) {
+ finish({ ok: false, reason: "not-found" });
+ return;
+ }
+ finish({
+ ok: true,
+ displayName: read.author.displayName,
+ handle: read.author.handle,
+ text: read.text,
+ quotedDisplayName: read.quoted.displayName,
+ quotedHandle: read.quoted.handle,
+ quotedText: read.quotedText,
+ datetime: read.datetime,
+ statusUrl: read.statusUrl,
+ images: read.images,
+ });
+ } catch {
finish({ ok: false, reason: "not-found" });
- return;
}
-
- finish({
- ok: true,
- displayName: author.displayName,
- handle: author.handle,
- text,
- quotedDisplayName: quoted.displayName,
- quotedHandle: quoted.handle,
- quotedText,
- datetime: mainName?.querySelector("time")?.getAttribute("datetime") ?? "",
- statusUrl,
- images,
- });
- } catch {
- finish({ ok: false, reason: "not-found" });
- }
+ })();
})();
diff --git a/apps/extension/src/tweet-clip.ts b/apps/extension/src/tweet-clip.ts
index 9641e020..8be093fb 100644
--- a/apps/extension/src/tweet-clip.ts
+++ b/apps/extension/src/tweet-clip.ts
@@ -84,6 +84,16 @@ export const cleanTweetText = (value: string) => {
return lines.join("\n").trim();
};
+// Quote cards on X put the name and @handle in plain text, not in links.
+// Keep the body of this function identical to visibleTweetAuthor in capture-tweet.ts.
+export const visibleTweetAuthor = (value: string) => {
+ const parts = value.replace(/\r/g, "").split("\n").map((part) => part.trim()).filter(Boolean);
+ const handlePart = parts.find((part) => /^@[A-Za-z0-9_]{1,15}$/.test(part));
+ const handle = handlePart ? handlePart.slice(1) : "";
+ const displayName = parts.find((part) => part !== handlePart && part !== "·" && part !== "•") ?? "";
+ return { displayName, handle };
+};
+
export const tweetAuthorLine = (displayName: string, handle: string) => {
const name = displayName.replace(/\s+/g, " ").trim();
const user = handle.replace(/^@/, "").replace(/\s+/g, "").trim();
diff --git a/apps/extension/tweet-clip.test.mjs b/apps/extension/tweet-clip.test.mjs
index 3a22b184..ccb84033 100644
--- a/apps/extension/tweet-clip.test.mjs
+++ b/apps/extension/tweet-clip.test.mjs
@@ -1,3 +1,4 @@
+import { readFileSync } from "node:fs";
import { describe, expect, test } from "bun:test";
import {
canonicalStatusUrl,
@@ -8,6 +9,7 @@ import {
tweetAuthorLine,
tweetNoteMarkdown,
tweetNoteTitle,
+ visibleTweetAuthor,
} from "./src/tweet-clip.ts";
const labels = {
@@ -58,6 +60,18 @@ describe("tweet note content", () => {
expect(markdown).not.toContain("pbs.twimg.com");
});
+ test("reads a quote author from the visible name block when it has no links", () => {
+ expect(visibleTweetAuthor("yetone\n@yetone\n·\n9月25日")).toEqual({ displayName: "yetone", handle: "yetone" });
+ expect(visibleTweetAuthor("Geek\n@geekbb\n·\n2小时")).toEqual({ displayName: "Geek", handle: "geekbb" });
+ expect(visibleTweetAuthor("")).toEqual({ displayName: "", handle: "" });
+ const page = readFileSync(new URL("./src/capture-tweet.ts", import.meta.url), "utf8");
+ const library = readFileSync(new URL("./src/tweet-clip.ts", import.meta.url), "utf8");
+ const parser = library.match(/const visibleTweetAuthor = \(value: string\) => \{[\s\S]*?\n\};/);
+ const compact = (source) => source.replace(/^[ \t]+/gm, "");
+ expect(parser?.[0]).toBeTruthy();
+ expect(compact(page)).toContain(compact(parser[0]));
+ });
+
test("accepts post photos and skips avatars, emoji, and video posters", () => {
expect(isTweetPhotoUrl("https://pbs.twimg.com/media/Abc?format=jpg&name=small")).toBe(true);
expect(isTweetPhotoUrl("https://pbs.twimg.com/card_img/Abc/img")).toBe(true);