fix(weixin): stop line wrapping from breaking markdown links

[KANIKIG]

What does this PR do?

Fixes a Weixin delivery-formatting regression where the copy-friendly long-line wrapper can split valid Markdown links in the middle of the [text](url) syntax. When that happens, WeChat receives broken Markdown and shows what should have been a clickable link as plain text.

This is most visible in cron/report/article-summary messages: the source Markdown is valid, but the Weixin adapter rewrites a long single-line link before delivery.

Root cause

Follow-up to #21258 / #19161. _wrap_copy_friendly_lines_for_weixin() intentionally wraps long non-code, non-table lines at WEIXIN_COPY_LINE_WIDTH so WeChat users can copy long prose more reliably.

That is still useful for plain text, but textwrap.wrap() does not understand Markdown inline syntax. A valid link such as:

[Designcenter NX中的分析载荷](https://...)

can be delivered as:

[Designcenter
NX中的分析载荷](https://...)

At that point WeChat no longer sees a valid Markdown link. The content is correct before formatting; it is corrupted during Weixin delivery formatting.

Related Issue

Follow-up to #21258 and #19161. I did not find an existing duplicate PR/issue for the interaction between Weixin copy-friendly wrapping and long Markdown links.

Related historical context, but not duplicates:

• fix(weixin): streaming cursor, media uploads, markdown links, blank messages #8665 / fix(weixin): preserve native markdown rendering #10338 preserved native Weixin Markdown behavior.
• fix(weixin): wrap long copy-unfriendly lines (salvage #19227) #21258 added the copy-friendly long-line wrapper for Weixin text.

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✅ Tests (adding or improving test coverage)

Changes Made

• gateway/platforms/weixin.py: skip copy-friendly wrapping for lines that contain Markdown links, reusing the existing _MARKDOWN_LINK_RE.
• tests/gateway/test_weixin.py: add a regression test for long Markdown links that exceed WEIXIN_COPY_LINE_WIDTH.

This keeps #21258's wrapping behavior for long plain text, issue-template/log-style prose, and other copy-unfriendly lines, while treating Markdown links as atomic inline content.

How to Test

1. Format a Weixin message containing a long Markdown link that exceeds WEIXIN_COPY_LINE_WIDTH.
2. Before this change, the wrapper can insert a newline inside [text](url) and break the link.
3. After this change, the Markdown link remains intact, while long plain text still wraps.

Tests run on Windows:

.\.venv\Scripts\python.exe -m pytest -p no:timeout -o "addopts=" tests/gateway/test_weixin.py::TestWeixinFormatting::test_format_message_wraps_long_plain_lines_for_copying tests/gateway/test_weixin.py::TestWeixinFormatting::test_format_message_does_not_wrap_long_markdown_links tests/gateway/test_weixin.py::TestWeixinFormatting::test_format_message_does_not_wrap_long_code_block_lines -q
# 3 passed

.\.venv\Scripts\python.exe scripts/run_tests_parallel.py tests/gateway/test_weixin.py -- -q -p no:timeout -o "addopts="
# 70 passed

.\.venv\Scripts\python.exe scripts/check-windows-footguns.py gateway/platforms/weixin.py tests/gateway/test_weixin.py
# No Windows footguns found

uv tool run --offline ruff@0.15.10 check gateway/platforms/weixin.py tests/gateway/test_weixin.py
# All checks passed

git diff --check
# passed

Checklist

Code

• I've read the Contributing Guide
• My commit messages follow Conventional Commits
• I searched for existing PRs to make sure this isn't a duplicate
• My PR contains only changes related to this fix
• I've run the relevant pytest suite and all targeted tests pass
• I've added tests for my changes
• I've tested on Windows

Documentation & Housekeeping

• Documentation update N/A — this preserves existing Weixin formatting semantics
• cli-config.yaml.example N/A — no config changes
• CONTRIBUTING.md / AGENTS.md N/A — no workflow or architecture changes
• Cross-platform impact considered — change is pure string formatting logic
• Tool descriptions/schemas N/A — no tool schema changes

[Copilot]

Pull request overview

Fixes a Weixin delivery formatting regression where the “copy-friendly” long-line wrapper could insert newlines inside valid Markdown link syntax ([text](url)), causing WeChat to render links as plain text.

Changes:

• Update Weixin’s copy-friendly wrapper to skip wrapping any line that contains an inline Markdown link (detected via existing _MARKDOWN_LINK_RE).
• Add a regression test ensuring long Markdown links are not modified/wrapped even when exceeding WEIXIN_COPY_LINE_WIDTH.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
gateway/platforms/weixin.py	Prevents the copy-friendly line wrapper from splitting lines containing Markdown links.
tests/gateway/test_weixin.py	Adds regression coverage for long inline Markdown links so they remain intact after formatting.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[teknium1]

Thanks for tracing this to the copy-friendly wrapper; the current-main path does pass long lines through textwrap.wrap() at gateway/platforms/weixin.py:726, so protecting inline link syntax is needed.

Problems

• gateway/platforms/weixin.py:722 skips wrapping the entire line when it contains any Markdown link. That removes the intended copy-friendly behavior for a long prose/report line containing even a short link; the wrapper was added for long copy-unfriendly lines in 7244a1f0d3c17631661fbf103440a3790ab0bab9.
• The new test only covers a nearly all-link line. It does not preserve the wrapper's behavior for prose surrounding an inline link.

Suggested changes

• Protect matched [label](url) spans as atomic tokens, but continue wrapping surrounding prose.
• Add a test with a space-containing link label embedded in long prose, asserting both that the link stays contiguous and that non-link prose is still wrapped.

Automated hermes-sweeper review.

[teknium1]

This bypasses the copy-friendly wrapper for the entire line, including arbitrary surrounding prose. Please protect only the matched Markdown-link span(s) and keep wrapping text before/after them; add coverage for a long prose line containing a short inline link.