See also: IRC log
Kris: lets get going. thanks for showing up, to the first meeting
plh: CVS - status w3c system team working on something besides CVS (longterm). System folks are still debating over git vs Mercurial. participants here might be more familiar with the use mercurial.
Kris: Will this work for folks in the long term?
<gsnedders> And I know the general preference is within Opera in Mercurial
Anne: Mercurial will work for opera
ACTION: plh to come back with a status update for moving away from CVS
plh: We have a tracker but not the irc bot yet. http://www.w3.org/html/wg/test/track/
Kris: Can we get a wiki setup? I'll take an action to setup some basic content on the wiki
Mike: OK, I can get it set up this week at http://www.w3.org/html/wg/test/wiki
krisk:Other infra items? besides moving away from
CVS - we should be all set with infrastructure items
Kris: Call to have one or more folks create sample test cases for the group
<gsnedders> There's all the html5lib tests
krisk: Lets add list of links to the wiki?
Anne:http://esw.w3.org/topic/HtmlTestMaterial
<gsnedders> Some of that is outdated, though
Kris: anyone want to voluneer for updating wiki with this information?
<jgriffin> I'll update it
<gsnedders> http://wiki.whatwg.org/wiki/Test_cases
<gsnedders> http://wiki.whatwg.org/wiki/Testsuite
krisk: good to have one spot with all the various test case
<plh> http://omocha.w3.org/wiki/newformat
<jgriffin> yes, I agree
Kris: csswg seems to have agreement on this type of format
plh: not necessarily optimal for htmlwg since having the metadata in the test can temper it
Kris: as the spec changes this format will be tough to keep up
<jgraham> I think that format has too many requirements per file
plh: It depends on what we're trying to achieve with the metadata as wel. Going from the spec to the tests?
Kris: agree
<gsnedders> The other problem is a lot of tests like HTML parsing tests test a large part of the spec
Kris: having meta data in each test case won't
work [parser example]
... the 3 items we need to track are pointer back to spec, harness, test case
status (submitted, reviewed, invalid, approved)
plh: would be nice to have a link back to the spec
krisk: as the spec changes we'll never keep the test up to pointing back to the spec
<gsnedders> Relying upon section numbers seems terrible (As they are highly volitale)
<annevk> gsnedders, yeah, we should prolly get stable IDs for sections
<gsnedders> Or just rely upon section names, as they're far more stable
<MikeSmith> I wonder if we if should try to put together a canonical list of section names/feature names
<annevk> MikeSmith, would be nice for people giving presentations :)
<MikeSmith> using section number probably ain't good, as gsnedders points out
krisk: sounds like with a canonical list and
folders we can close this item?
... I'll take an action item to post this to the wiki
... format /submitted/featurename/submitter/harness/tests
/submitted/parser/microsoft/reftest/parser01.htm
other items?
next meeting technically is 12/22 - does this work for folks?
plh: lets only meet if we have items to talk about
krisk: agree