Allow HTML tags in link text

[remibetin]

Resolves w3c/wai-website#481 (and also addresses w3c/wai-std-gl-overview#21)

Description of the problem

When link text is enclosed in double square brackets, [[link_text]](/url/), we want the plugin to replace it with the title of the target resource, in the same language as the current page. When the target page has not been translated yet, we want the plugin to keep the link text that is between brackets.

That is generally working, but sometimes, translators prefer to keep part of the link text in English. In this case, they add <span lang="en"></span> markup in the link text.

The plugin does not handle this case correctly.

Example raised in w3c/wai-website#481

[[Χρησιμοποιώντας Υλικά του <span lang='en'>WAI</span>: Άδεια Χρήσης με Απόδοση (<span lang='en'>Attribution</span>)]](/about/using-wai-material/)

is rendered as:

Μπορείτε να χρησιμοποιήσετε αυτό το βίντεο εάν συμπεριλάβετε έναν σύνδεσμο προς αυτήν τη σελίδα. Περισσότερες πληροφορίες διατίθενται στο «Χρησιμοποιώντας Υλικά του WAI: Άδεια Χρήσης με Απόδοση (Attribution) (στα Αγγλικά)»(/WAI/about/using-wai-material/){: hreflang=”en”}.

Instead of:

Χρησιμοποιώντας Υλικά του WAI: Άδεια Χρήσης με Απόδοση (Attribution) (στα Αγγλικά)

Technical side of the problem

Double brackets links are currently processed in 2 steps:

• Step 1: [[link_text]](/url/#fragment)
  is turned into
  <<link_text ("in English" in the target language)>>(/WAI/url/#fragment){: hreflang="en"}

  wai-website-plugin/lib/wai-website-plugin.rb

  Lines 31 to 39 in 4392f43

  	document.content.gsub!(/\[\[([^\]\]]+?)\]\]\((?!\/TR)(?!\/WAI)(\/.*?\/)(?:#(.*?))?\)/i) do |match|
  	translatedpage = getPage document.site, Regexp.last_match[2], document.data['lang']
  	if Regexp.last_match[3].nil?
  	fragment = ''
  	else
  	fragment = '#' + Regexp.last_match[3]
  	end
  	if translatedpage.nil?
  	'<<' + Regexp.last_match[1] + inenglishtext +'>>({{ "' + Regexp.last_match[2] +'" | relative_url }}'+ fragment+')' + hreflang

• Step 2: <<link_text ("in English" in the target language)>>(/WAI/url/#fragment){: hreflang="en"}
  is expected to be turned into
  [link_text ("in English" in the target language)](/WAI/url/#fragment){: hreflang="en"}

  wai-website-plugin/lib/wai-website-plugin.rb

  Lines 68 to 70 in 4392f43

  	document.content.gsub!(/<<([^>>]+?)>>/i) do |match|
  	'[' + Regexp.last_match[1] +']'
  	end

Problem: the regular expression /<<([^>>]+?)>>/i tries to extract the link text but stops at first > character (the second > is superfluous). Including at first HTML tag.

Solution

[In Progress]