reftest framework and HTML5 parser testing

[Henri Sivonen]

Would it be feasible to modify the reftest framework to run the browser
over a long list of URIs from the Web and render each page twice for
comparing the two renderings: first with a pref set to false and then
with the pref set to true?

That is, would it be feasible to automate discovery of what kind of
visual changes the use of the HTML5 parsing algorithm causes compared to
using the old HTML parser?

--
Henri Sivonen
hsiv...@iki.fi
http://hsivonen.iki.fi/

[Robert O'Callahan]

On 21/4/09 12:09 AM, Henri Sivonen wrote:
> Would it be feasible to modify the reftest framework to run the browser
> over a long list of URIs from the Web and render each page twice for
> comparing the two renderings: first with a pref set to false and then
> with the pref set to true?
>
> That is, would it be feasible to automate discovery of what kind of
> visual changes the use of the HTML5 parsing algorithm causes compared to
> using the old HTML parser?

Yes, that would be pretty easy.

Rob

[L. David Baron]

On Monday 2009-04-20 14:09 +0200, Henri Sivonen wrote:
> Would it be feasible to modify the reftest framework to run the browser
> over a long list of URIs from the Web and render each page twice for
> comparing the two renderings: first with a pref set to false and then
> with the pref set to true?

It's technically feasible, although probably undesirable. What you
want is probably easier done using the mochitest framework, since we
already have a mechanism for taking reftest snapshots inside
mochitests. See:
http://mxr.mozilla.org/mozilla-central/search?string=WindowSnapshot.js

For examples of changing prefs in mochitests, see, e.g.,
http://mxr.mozilla.org/mozilla-central/source/layout/style/test/test_dont_use_document_colors.html?force=1

-David

--
L. David Baron http://dbaron.org/
Mozilla Corporation http://www.mozilla.com/