Fix web page loader to handle nested iframes correctly

[dmitrivMS]

Fixes #283248

[Copilot]

Pull request overview

This PR fixes the web page loader to correctly extract content from nested iframes, addressing issue #283248 where JavaDoc content embedded in iframes was not being retrieved. The fix enables the Page domain, retrieves the complete frame tree, and extracts accessibility content from all frames (main frame and all nested iframes) rather than just the main frame.

Key Changes:

• Enhanced accessibility tree extraction to iterate through all frames using Chrome DevTools Protocol's Page.getFrameTree command
• Refactored the accessibility domain to handle multiple root nodes (one per frame) instead of a single tree
• Added comprehensive test coverage for nested iframe scenarios

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
src/vs/platform/webContentExtractor/electron-main/webPageLoader.ts	Modified extractAccessibilityTreeContent to enable Page domain, fetch frame tree, flatten it using BFS, and extract accessibility nodes from all frames with proper cancellation token handling
src/vs/platform/webContentExtractor/electron-main/cdpAccessibilityDomain.ts	Refactored createNodeTree to createNodeTrees to return multiple root nodes for multi-frame scenarios, updated convertAXTreeToMarkdown to process and combine content from all trees
src/vs/platform/webContentExtractor/test/electron-main/webPageLoader.test.ts	Extracted common mock setup into setupDebuggerMock helper function, added comprehensive test for nested iframe content extraction, refactored all existing tests to use the new helper