Turning the tide for tidy

[forthrin]

Sadly, the "granddaddy of HTML tools" seems to have retired to the old folks' home for good, with no commits or responses from maintainers going on three years.

Would thus like to explore how Nokogiri can transform street HTML to valid HTML5.

1. ENSURE required tags (<!DOCTYPE> and <title>)
2. PRETTIFY (indent, mostly)
3. REMOVE disallowed tags and attributes

Possibly also:

4. PREVENT external auto-loads (<script>, <iframe>, <img src>, url())
5. COMPLETE relative URLs (and best-guess base[href], link[href][rel=canonical])

Already wrote much of this, but are some available OotB? (Can't get indentation to work for one.)

[flavorjones]

If you want to be constructive, please write some failing tests demonstrating specific behaviors that you want to see. What you've written here spans well-formedness, formatting, sanitization, and post-processing -- all of which could be a separate project.

I would also suggest if you haven't to look at rgrove/sanitize which does a good job of providing a flexible API for sanitization; and flavorjones/loofah which provides an API for generic tree transformations.

[forthrin]

Will check how the references fly. But does Nokogiri indent out of the box? indent: does nothing.