fix(weixin): preserve native markdown rendering

[bernylinville]

Summary

• preserve raw Markdown in the Weixin adapter instead of rewriting tables, headings, and links
• keep Markdown tables and headings together during Weixin message chunking
• update Weixin tests and docs to reflect native Markdown rendering

Root cause

The Weixin adapter assumed WeChat could not render Markdown well, so format_message() rewrote outgoing content before delivery:

• tables were converted into pseudo key/value lists
• headings were rewritten into custom plain-text variants
• markdown links were rewritten/stripped

That behavior now degrades real Weixin output because the current client can render Markdown tables natively.

Testing

• source venv/bin/activate && python -m pytest tests/gateway/test_weixin.py -q

Closes #10308

[teknium1]

Merged as part of Batch-3 salvage: #11571

Your commit cherry-picked onto main (rewrote the author line to your GitHub identity bernylinville@devopsthink.org since the original commit had placeholder Your Name <your-email@example.com>). The change has community backing across #7468, #7617, and #9612 — all reporting that the legacy rewriting degrades modern WeChat output. Low-risk fallback if any client variant doesn't render markdown natively. Thanks for the systematic test + docs update.

Commit SHAs on main: 6ee65b4