`header_id_from_text` used by extras "header-ids" doesn't parse markdown

I wanted to make a heading a link:

```
# [link](https://github.com)
```
# [link](https://github.com)

but the id returned includes the URL text. In this case it should be just "link". Would it be expensive to parse `text` in the function before the call to `_slugify`?
