Some sort of mapping function

[xmo-odoo]

One of the biggest issues of markupsafe is that most of the world is not markupsafe-aware, so when performing text transformations through markupsafe-unaware (but safe) APIs. In my experience, regex and markdown are common issues.

They generally require type-checking the input, applying the transformation, then re-wrapping if necessary, commonly transforming a simple expression into several statements (or a more complicated conditional expression).

Would be nice if markupsafe provided a markupsafe.map utility which did that internally, applying a user-supplied transformation function then rewrapping the input to match the output.

[davidism]

This sounds outside the scope of MarkupSafe. I'm not clear what this would be compared to Markup(markdown.render(value)).

[xmo-odoo]

This sounds outside the scope of MarkupSafe. I'm not clear what this would be compared to Markup(markdown.render(value)).

It would check that the original value is a Markup and only re-wrap in that case: if the input of markdown rendering is markup-safe, then the output is generally markup-safe (barring odd extensions).

An unconditional wrapping would launder markup from an unescaped markdown doc into the output. While there are use-cases for that, untrusted input is a pretty big use case for markdown.