Add more robust support for HTML5 anchor tags. by xStrom · Pull Request #16 · k3a/html2text

xStrom · 2023-05-04T21:42:41Z

This PR makes the anchor tag regex more robust so it can handle more of the HTML5 spec.

Old   a.*href=('([^']*?)'|"([^"]*?)")
New   ^(?i:a)(?:$|\s).*(?i:href)\s*=\s*('([^']*?)'|"([^"]*?)"|([^\s"'`=<>]+))

To break down the changes, in order:

Regex	Reasoning
`^`	Check that the tag actually starts with the letter `a`, as opposed to say in the case of `head`.
`(?i:a)`	Tags and attribute names are case insensitive, so we need a case-insensetive check.
`(?:$\|\s)`	We want only the letter `a` and not say `article`, so we check for the end the same way as `badTagnamesRE`.
`(?i:href)`	Case-insensetive check for attribute names as well.
`\s=\s`	The equals sign can be surrounded by zero or more spaces.
\|([^\s"'`=<>]+)	Attribute values don't have to be enclosed in quotes if they follow certain rules.

k3a · 2023-05-10T19:48:08Z

Thanks for the nice PR! I will make other tag regexpes case-insensitive as well in the followup commits.

Add more robust support for HTML5 anchor tags.

66c9ccd

k3a merged commit a58537e into k3a:master May 10, 2023

xStrom deleted the anchor branch May 10, 2023 21:23

Provide feedback