Eliminate HTML from translations

Summary

Per this review from @reprazent, and in the spirit of looking forward to https://gitlab.com/gitlab-org/gitlab/issues/19485, all translated text should be HTML-free.

Improvements

1. Obviously, https://gitlab.com/gitlab-org/gitlab/issues/19485 will be easier to implement.
2. Translators won't have to know HTML

Risks

The only risk is basically if - during the refactor - the current HTML + text is not removed from the translated string appropriately. This could fall into two categories:

1. Simply broken. A missing tag or incorrect text
2. Incorrect interpolation. Use locations could accidentally not pass the extracted parameter, or pass an incomplete value

Involved components

Primarily: gitlab.pot. However, that file is constructed from a multitude of places. A search for /<\w*?> in gitlab.pot returned 105 results.