Bug 25833 – Page that performs syntax highlighting on <code> elements in JS looks wrong when loaded from web archive

• Status: NEW
• Component: Layout and Rendering

[Reiner Pittinger]

When saving a page with a <code>-element, the saved page is displayed incorrectly when viewing the saved archive.

Reproduce:

1. Go to http://guides.rubyonrails.org/i18n.html
2. Look at the correctly displayed code block (below "1.2 The Public I18n API").
3. Save the website as a webarchive.
4. Open the archive and look at the same document part again. Now you will see text like " class="comment">", which is actually a broken <span>-element.

[Mark Rowe (bdash)]

Looking in the web inspector, the markup for the code block on the original page looks like so:
<div class="code_container"><code class="ruby">
translate <span class="comment"># Lookup text translations</span>
localize  <span class="comment"># Localize Date and Time objects to local formats</span>
</code></div>

This matches what is in the document source.  When loaded from the web archive, the markup for the code block looks like so:
<div class="code_container"><code class="ruby">
translate <span><span class="keywords">class</span>=<span class="string">"comment"</span>&gt;<span class="comment"># Lookup text translations</span></span>
localize  <span><span class="keywords">class</span>=<span class="string">"comment"</span>&gt;<span class="comment"># Localize Date and Time objects to local formats</span></span>
</code></div>

Looking at the source of the main resource via View -> Source shows the original, correct markup.

[Mark Rowe (bdash)]

I was mistaken:  the markup in the original page looks like:
<div class="code_container"><code class="ruby">
translate # Lookup text translations
localize  # Localize Date and Time objects to local formats
</code></div>

Since the DOM being displayed contains extra elements related to syntax highlighting, I suspect there is syntax highlighting being performed via JavaScript.  When saving as a web archive we serialize from the DOM at that point in time.  When the document is then loaded once more the syntax highlighting is run on the already-highlighted element, which causes it to attempt to highlight the presentational markup that the earlier syntax highlighting introduced.

[Mark Rowe (bdash)]

I知 not sure whether this is really a bug.

[Mark Rowe (bdash)]

Brady, what do you think we should do with this?