W3C home > Mailing lists > Public > public-html-testsuite@w3.org > September 2010

Reflection tests

From: Aryeh Gregor <Simetrical+w3c@gmail.com>
Date: Sun, 26 Sep 2010 19:54:20 -0400
Message-ID: <AANLkTinACAMhzYQ+eEJZ_3gAvPT1svL3r_P4XEM4Wt7U@mail.gmail.com>
To: public-html-testsuite@w3.org
I've been working on a test suite in my spare time for a while,
testing reflected attributes:

http://aryeh.name/tests/reflection.html

Since I'm an HTMLWG member, I'd check it into Mercurial, but there are
two things that make me hesitate:

1) It doesn't use the existing framework for pure JS tests, but
instead uses something I wrote myself.  I could try to port it to the
existing framework, but it would probably require some changes to the
framework to work smoothly.  There are over 15,000 tests, and I have
plenty more to add.  They're extremely repetitive, since they're
generated programmatically, and if I didn't write code in my framework
to filter out or consolidate the more repetitive failures, the results
would be a lot harder to make sense of.

2) I'm still actively developing the tests, and they're all contained
in one file, so it's not like we could easily chop them up into pieces
and call some stable and others not.

So I'm interested in hearing what people think.  Should I check them
into hg regardless and just make it clear they still need work done?
Should I wait until they're ported to the test framework that other
stuff is using?  Should I wait until they're complete and stable?
Also, how will review be conducted with this many tests?  A
source-code review would make much more sense than trying to review
each actual test.
Received on Monday, 27 September 2010 00:02:39 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 September 2010 00:02:40 GMT