fix: preserve text nested inside <br> by html.parser (#244)

[gaoflow]

Summary

Fixes #244.

Python's html.parser has an edge case where text following <br /> (written with a space before the slash) is parsed as child content of the <br> element instead of a sibling text node, when the preceding sibling is a bare <br> tag.

from bs4 import BeautifulSoup
soup = BeautifulSoup('Hello<br>cruel<br />world', 'html.parser')
list(soup.children)
# ['Hello', <br/>, 'cruel', <br>world</br>]   # 'world' is nested inside <br>!

convert_br received the nested text via its text parameter but silently discarded it, causing characters to disappear from the output:

from markdownify import markdownify as md
md('Hello<br>cruel<br />world')
# Before: 'Hello  \ncruel  '   # 'world' is lost
# After:  'Hello  \ncruel  \nworld'

Fix

Append text to the returned line-break marker in all branches of convert_br. When <br> is used correctly as a void element, text is always an empty string, so the change is backward-compatible. The _inline branch is handled similarly for consistency.

Test

Added two assertions to test_br covering the mixed <br> / <br /> pattern that triggers the html.parser nesting behaviour.

This pull request was prepared with the assistance of AI, under my direction and review.