Escape link/image destinations and image alt text (#261)

[assinscreedFC]

Summary

Fixes #261.

convert_a, convert_img and convert_video dropped raw attribute values and generated labels into Markdown link/image syntax without escaping them for that context. As reported, this lets a destination or alt value truncate the syntax, so downstream Markdown parsers read different HTML than the original — including substituting an attacker-controlled URL:

md('<img src="/safe" alt="](http://attacker)">')
# before: '![](http://attacker)](/safe)'  -> re-parses to <img src="http://attacker">
# after:  '![\](http://attacker)](/safe)' -> alt is literal text, src preserved

Changes

Two helpers in markdownify/__init__.py:

• escape_link_destination(url) — wraps a destination containing whitespace or parentheses in angle brackets (CommonMark allows spaces and () inside <...>), escaping any </>. Applied to href (convert_a), src (convert_img, convert_video) and poster (convert_video).
• escape_link_text(text) — escapes [ and ] (and backslashes) in alt text taken verbatim from HTML, so it can't close the [...] label early. Applied to alt in convert_img.

Verified against all reporter cases:

input	output
<img src="/a" alt="]">	![\]](/a)
<img src="/a b" alt="x">	
<img src="/safe" alt="](http://attacker)">	](/safe)
<a href="/a)b">click</a>	[click](</a)b>)

Unaffected inputs (no spaces/parens/brackets) are emitted unchanged, so existing output is preserved.

Scope note

I left link text for convert_a/convert_video alone, since that is already-rendered Markdown from child nodes (escaped via self.escape()), not a verbatim attribute. Happy to extend if you'd prefer.

Test plan

• Added test_a_escapes_destination, test_img_escapes_link_syntax, test_video_escapes_destination in tests/test_conversions.py.
• pytest — 86 passed.
• flake8 --ignore=E501,W503 markdownify tests — clean.