URL_REGEX doesn't match URLs in markdown link syntax

Issue Description

Component: OCP\IURLGenerator::URL_REGEX / Reference extraction in ReferenceManager

Current Behavior:

The URL_REGEX pattern used by ReferenceManager::extractReferences() fails to match URLs when they're embedded in markdown link syntax:

// lib/public/IURLGenerator.php
public const URL_REGEX_NO_MODIFIERS = '(\s|\n|^)(https?:\/\/)([-A-Z0-9+_.]+(?::[0-9]+)?(?:\/[-A-Z0-9+&@#%?=~_|\!:,.;()]*)*)(\s|\n|$)';

The regex requires whitespace/newline before the URL ((\s|\n|^)), which doesn't match markdown syntax where URLs are preceded by ](.

Test Case:

Plain URL (works):
"Check https://github.com/nextcloud/server/issues/55845 for details"
✅ Extracted: https://github.com/nextcloud/server/issues/55845

Markdown link (fails):
"Check [GH #55845](https://github.com/nextcloud/server/issues/55845) for details"
❌ Extracted: (nothing)

Root Cause:

In lib/private/Collaboration/Reference/ReferenceManager.php line 53:

public function extractReferences(string $text): array {
    preg_match_all(IURLGenerator::URL_REGEX, $text, $matches);
    // ...
}

The regex pattern doesn't account for markdown link syntax [label](url) where the URL is preceded by ]( instead of whitespace.

Impact:

Reference providers (GitHub, GitLab, Zammad, custom integrations) don't generate rich previews when users paste markdown-formatted links, forcing users to paste plain URLs which reduces text readability.

Affected Use Cases:

• [GH #55845](https://github.com/nextcloud/server/issues/55845) - GitHub issues
• [Ticket #12345](https://support.example.com/ticket/12345) - Support tickets
• [PROJ-123](https://jira.example.com/browse/PROJ-123) - JIRA issues
• Any markdown link in Text app, Talk, or Comments

Proposed Solution:

Update URL_REGEX_NO_MODIFIERS to also match URLs preceded by markdown syntax:

public const URL_REGEX_NO_MODIFIERS = '(\s|\n|^|\]\()(https?:\/\/)([-A-Z0-9+_.]+(?::[0-9]+)?(?:\/[-A-Z0-9+&@#%?=~_|\!:,.;()]*)*)(\s|\n|$|\))';
//                      ^^^^ added markdown start      ^^^ added closing paren

Alternative Solution:

Strip markdown syntax before URL extraction:

public function extractReferences(string $text): array {
    // Strip markdown link syntax: [label](url) → url
    $text = preg_replace('/\[([^\]]+)\]\(([^)]+)\)/', ' $2 ', $text);
    preg_match_all(IURLGenerator::URL_REGEX, $text, $matches);
    // ...
}

Backward Compatibility:

• Existing plain URL extraction continues to work
• New: Markdown links also extracted
• No breaking changes for reference providers

Related Code:

• lib/public/IURLGenerator.php - URL_REGEX definition
• lib/private/Collaboration/Reference/ReferenceManager.php - extractReferences()
• Frontend: core/src/OCP/comments.js (mentioned in comments as needing sync)

Nextcloud Version:

• Affects: All versions with Reference Provider system (NC 25+)
• Tested on: Nextcloud 30

Workaround:

Users must paste plain URLs without markdown formatting to trigger rich previews.

Start with URL_REGEX_NO_MODIFIERS in lib/public/IURLGenerator.php and extractReferences() in lib/private/Collaboration/Reference/ReferenceManager.php. Check the plain URL and markdown-link examples from the issue, then verify that extraction still handles existing URLs and also returns URLs enclosed by markdown link syntax; review core/src/OCP/comments.js for the mentioned synchronization concern.