fix(editor): database behavier (#14394)

fix #13459
fix #13707
fix #13924

#### PR Dependency Tree


* **PR #14394** 👈

This tree was auto-generated by
[Charcoal](https://github.com/danerwilliams/charcoal)

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Improved URL paste: text is split into segments, inserted correctly,
and single-URL pastes create linked-page references.

* **UI Improvements**
  * Redesigned layout selector with compact dynamic options.
* Number-format options are always available in table headers and mobile
menus.

* **Bug Fixes**
  * More consistent paste behavior for mixed text+URL content.
  * Prevented recursive selection updates when exiting edit mode.

* **Tests**
* Added tests for URL splitting, paste insertion, number formatting, and
selection behavior.

* **Chores**
* Removed number-formatting feature flag; formatting now applied by
default.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->

blocksuite/affine/shared/src/__tests__/utils/url.unit.spec.ts

@@ -1,6 +1,6 @@
 import { describe, expect, test } from 'vitest';
 
-import { isValidUrl } from '../../utils/url.js';
+import { isValidUrl, splitTextByUrl } from '../../utils/url.js';
 
 describe('isValidUrl: determining whether a URL is valid is very complicated', () => {
   test('basic case', () => {
@@ -85,3 +85,55 @@ describe('isValidUrl: determining whether a URL is valid is very complicated', (
     expect(isValidUrl('http://127.0.0.1', 'http://127.0.0.1')).toEqual(true);
   });
 });
+
+describe('splitTextByUrl', () => {
+  test('should split text and keep url part as link segment', () => {
+    expect(splitTextByUrl('hi - https://google.com')).toEqual([
+      { text: 'hi - ' },
+      {
+        text: 'https://google.com',
+        link: 'https://google.com',
+      },
+    ]);
+  });
+
+  test('should support prefixed url token without swallowing prefix text', () => {
+    expect(splitTextByUrl('-https://google.com')).toEqual([
+      { text: '-' },
+      {
+        text: 'https://google.com',
+        link: 'https://google.com',
+      },
+    ]);
+  });
+
+  test('should trim tail punctuations from url token', () => {
+    expect(splitTextByUrl('visit https://google.com, now')).toEqual([
+      { text: 'visit ' },
+      {
+        text: 'https://google.com',
+        link: 'https://google.com',
+      },
+      { text: ', now' },
+    ]);
+  });
+
+  test('should convert domain token in plain text', () => {
+    expect(splitTextByUrl('google.com and text')).toEqual([
+      {
+        text: 'google.com',
+        link: 'https://google.com',
+      },
+      { text: ' and text' },
+    ]);
+  });
+
+  test('should normalize www domain token link while preserving display text', () => {
+    expect(splitTextByUrl('www.google.com')).toEqual([
+      {
+        text: 'www.google.com',
+        link: 'https://www.google.com',
+      },
+    ]);
+  });
+});

blocksuite/affine/shared/src/services/feature-flag-service.ts

@@ -2,7 +2,6 @@ import { type Store, StoreExtension } from '@blocksuite/store';
 import { type Signal, signal } from '@preact/signals-core';
 
 export interface BlockSuiteFlags {
-  enable_database_number_formatting: boolean;
   enable_database_attachment_note: boolean;
   enable_database_full_width: boolean;
   enable_block_query: boolean;
@@ -28,7 +27,6 @@ export class FeatureFlagService extends StoreExtension {
   static override key = 'feature-flag-server';
 
   private readonly _flags: Signal<BlockSuiteFlags> = signal({
-    enable_database_number_formatting: false,
     enable_database_attachment_note: false,
     enable_database_full_width: false,
     enable_block_query: false,

blocksuite/affine/shared/src/utils/url.ts

@@ -95,28 +95,107 @@ export function isValidUrl(str: string, baseUrl = location.origin) {
   return result?.allowed ?? false;
 }
 
+const URL_SCHEME_IN_TOKEN_REGEXP =
+  /(?:https?:\/\/|ftp:\/\/|sftp:\/\/|mailto:|tel:|www\.)/i;
+
+const URL_LEADING_DELIMITER_REGEXP = /^[-([{<'"~]+/;
+
+const URL_TRAILING_DELIMITER_REGEXP = /[)\]}>.,;:!?'"]+$/;
+
+export type UrlTextSegment = {
+  text: string;
+  link?: string;
+};
+
+function appendUrlTextSegment(
+  segments: UrlTextSegment[],
+  segment: UrlTextSegment
+) {
+  if (!segment.text) return;
+  const last = segments[segments.length - 1];
+  if (last && !last.link && !segment.link) {
+    last.text += segment.text;
+    return;
+  }
+  segments.push(segment);
+}
+
+function splitTokenByUrl(token: string, baseUrl: string): UrlTextSegment[] {
+  const schemeMatch = token.match(URL_SCHEME_IN_TOKEN_REGEXP);
+  const schemeIndex = schemeMatch?.index;
+  if (typeof schemeIndex === 'number' && schemeIndex > 0) {
+    return [
+      { text: token.slice(0, schemeIndex) },
+      ...splitTokenByUrl(token.slice(schemeIndex), baseUrl),
+    ];
+  }
+
+  const leading = token.match(URL_LEADING_DELIMITER_REGEXP)?.[0] ?? '';
+  const withoutLeading = token.slice(leading.length);
+  const trailing =
+    withoutLeading.match(URL_TRAILING_DELIMITER_REGEXP)?.[0] ?? '';
+  const core = trailing
+    ? withoutLeading.slice(0, withoutLeading.length - trailing.length)
+    : withoutLeading;
+
+  if (core && isValidUrl(core, baseUrl)) {
+    const segments: UrlTextSegment[] = [];
+    appendUrlTextSegment(segments, { text: leading });
+    appendUrlTextSegment(segments, { text: core, link: normalizeUrl(core) });
+    appendUrlTextSegment(segments, { text: trailing });
+    return segments;
+  }
+
+  return [{ text: token }];
+}
+
+/**
+ * Split plain text into mixed segments, where only URL segments carry link metadata.
+ * This is used by paste handlers so text like `example:https://google.com` keeps
+ * normal text while only URL parts are linkified.
+ */
+export function splitTextByUrl(text: string, baseUrl = location.origin) {
+  const chunks = text.match(/\s+|\S+/g);
+  if (!chunks) {
+    return [];
+  }
+
+  const segments: UrlTextSegment[] = [];
+  chunks.forEach(chunk => {
+    if (/^\s+$/.test(chunk)) {
+      appendUrlTextSegment(segments, { text: chunk });
+      return;
+    }
+    splitTokenByUrl(chunk, baseUrl).forEach(segment => {
+      appendUrlTextSegment(segments, segment);
+    });
+  });
+  return segments;
+}
+
 // https://en.wikipedia.org/wiki/Top-level_domain
 const COMMON_TLDS = new Set([
-  'com',
-  'org',
-  'net',
-  'edu',
-  'gov',
-  'co',
-  'io',
-  'me',
-  'moe',
-  'mil',
-  'top',
-  'dev',
-  'xyz',
-  'info',
   'cat',
-  'ru',
+  'co',
+  'com',
   'de',
+  'dev',
+  'edu',
+  'eu',
+  'gov',
+  'info',
+  'io',
   'jp',
-  'uk',
+  'me',
+  'mil',
+  'moe',
+  'net',
+  'org',
   'pro',
+  'ru',
+  'top',
+  'uk',
+  'xyz',
 ]);
 
 function isCommonTLD(url: URL) {