--- Comment #13 from Rob Lanphier <> 2010-10-05 16:38:12 UTC 
Hi Priyanka, I think checking for a cache hit seems like a fine idea, assuming
it doesn't make the logic too convoluted.

Hi Platonides, here's what's going on.  Thanks to the fix in bug 24124, there
will be times the parsed content is pulled from the parser cache rather than
rendered directly.  In those cases, nothing would change.  However, if we have
a cache miss, then the goal would be to render a placeholder (e.g. an empty div
or even <div id="rendered_version">Loading...</div>), and then use Javascript
to replace the div with rendered version.  The visual effect of this would be
that the revision would always be rendered, but the rendering of the revision
(which is often completely below the fold) would no longer block the much
faster rendering of the diff (always above the fold).

Based on the last conversation I had with Priyanka, I think we all agree that
removing the cut-n-pasted code from renderNewRevision would be a very good
thing, but outside the scope of addressing this issue.

Configure bugmail:
------- You are receiving this mail because: -------
You are on the CC list for the bug.

Wikibugs-l mailing list

Reply via email to