Fix HTML escaping problems

[daniel-rikowski]

This fixes #50 resp. awesome-print/awesome_print#199

Also this prevents the ActiveRecord formatters from emitting unescaped #<ClassName#object_id> or Class < Superclass strings, which couldn't be catched by the caller even with manual escaping.

Existing code using workarounds like <%= raw ap ... %> or <%= (ap ...).html_safe %> will continue to work.

[paddor]

Thanks for the PR. I haven't touched Rails in forever. Can this be verified in a test case?

[HarlemSquirrel]

@daniel-rikowski Do you have some time to resolve the conflicts? Thank you!

[daniel-rikowski]

I rebased my pull request, but now tests are failing. It looks like 25352d2 added an explicit test to enforce that the returned HTML is not marked as safe. It looks like it declares #50 to be the intended behaviour.

I'm not sure what to do here...

[HarlemSquirrel]

I wrote a test to verify how it works today. Could you update the test to verify this change?
Thank you!

[daniel-rikowski]

Oh, sorry, I didn't notice #65 🙈

I updated the test and added a new one for the ActiveRecord subclass bug, i.e. the < in ChildModel < ParentModel wasn't escaped as < when using html: true.

[HarlemSquirrel]

Thank you!