Incorrect parsing of Setext-style headings

[saf-dmitry]

Python-Markdown wrongly recognizes following as H1 and H2 headings, respectively:

Some text
=-=-=-=-=

Some text
-=-=-=-=-

This is because the regex for detection of Setext-style headings handles runs of the = and - characters incorrectly. Below the improved version, which uses (?:[=]+|[-]+) group instead of [=-]+:

RE = re.compile(r'^.*?\n(?:[=]+|[-]+)[ ]*(\n|$)', re.MULTILINE)

[Jah-yee]

Fixed in PR #1605 — regex changed from [=-]+ to (?:[=]+|[-]+) to match only homogeneous runs. Thanks for the clear issue!