IRC log of htmlt on 2011-05-17

Timestamps are in UTC.

15:02:11 [RRSAgent]
RRSAgent has joined #htmlt
15:02:11 [RRSAgent]
logging to http://www.w3.org/2011/05/17-htmlt-irc
15:02:23 [krisk]
RRSAgent, make logs public
15:02:50 [krisk]
zakim, this is htmlt
15:02:50 [Zakim]
sorry, krisk, I do not see a conference named 'htmlt' in progress or scheduled at this time
15:03:12 [krisk]
If someone else joins the IRC we can potentially setup the conf call
15:03:29 [krisk]
Though we can use IRC (like normal)
15:04:41 [krisk]
Agenda -> http://lists.w3.org/Archives/Public/public-html-testsuite/2011May/0002.html
15:06:35 [krisk]
We have no new bugs on approved tests
15:06:47 [krisk]
Bug Link http://www.w3.org/Bugs/Public/buglist.cgi?query_format=advanced&short_desc_type=allwordssubstr&short_desc=&product=HTML+WG&component=testsuite&long_desc_type=allwordssubstr&long_desc=&bug_file_loc_type=allwordssubstr&bug_file_loc=&status_whiteboard_type=allwordssubstr&status_whiteboard=&keywords_type=allwords&keywords=&bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&emailtype1=substring&email1=&emailtype2=substring&email2=&bugidtype=include&bug_i
15:07:13 [krisk]
I wanted to continue to discussion from our last meeting
15:08:33 [krisk]
I re-read the IRC from our last meeting and was not 100% sure why you didn't want to move in this direction
15:08:59 [jgraham]
Because I don't think it makes any sense
15:09:54 [jgraham]
A testsuite is continuously developing. It makes sense to have a process of continuous improvement ("continuous integration") rather than a process based around arbitarily set dates and stabilisation and so on
15:10:11 [jgraham]
Particularly because there isn't any stabilisation to do
15:10:27 [jgraham]
It's a *testsuite*. In itself it can't regress
15:10:36 [jgraham]
Unless the spec is wrong or something
15:10:41 [jgraham]
s/wrong/changed/
15:11:05 [jgraham]
We already know which tests are considered approved
15:11:13 [jgraham]
We have a whole folder for that
15:11:26 [jgraham]
(I would like to switch to using branches, but different discussion)
15:11:50 [jgraham]
It doesn't do anything to magically fix the submitted-not-approved problem
15:12:09 [krisk]
It should help get more people to particpate
15:12:10 [jgraham]
Because there is still no incentive for people to spend their time reviewing tests
15:12:35 [krisk]
..and would allow them a timeframe to potentially plan for reviewing tests
15:12:36 [jgraham]
Whereas writing tests will continue because it is typically a byproduct of other activity
15:12:58 [jgraham]
The coverage problem is a technical issue that we haven't solved
15:13:31 [jgraham]
Why would more people participate because every few months we said "OK stop writing tests and review tests instead!"
15:14:13 [krisk]
Because a schedule would exists and one could plan/allocate time if they were aware of a schedule
15:14:15 [jgraham]
Or even if we don't do that and just pick arbitary revisions to call "alpha", "beta", etc.
15:14:37 [jgraham]
I don't see how a schedule would help anyone, really
15:15:03 [jgraham]
Do we know of anyone who is not submitting tests due to the lack of a schedule?
15:15:31 [jgraham]
It's certainly not a problem for Opera; we write tests when we implement features
15:15:47 [jgraham]
That seems highly unlikely to change
15:17:12 [jgraham]
Possibky your claim is that there is no technical motivation but the psychological effect of being told "next Wednesday we will call what we have alpha 2" will be helpful to motivate people?
15:17:21 [jgraham]
*Possibly
15:17:32 [jgraham]
s/motivation/benefit/
15:17:41 [krisk]
Well we could try it and see what occurs...
15:18:29 [jgraham]
Well of course if you want to call some revisions "releases" or "milestones" or whatever that is fine
15:18:45 [jgraham]
I don't think it should have any impact on people submitting tests though
15:19:40 [krisk]
if doesn't actually change particpation then we can abandon the doing "releases" or "milestones"
15:20:43 [jgraham]
I mean I am happy as long as there is no change in the process of the group because we are x days from a milestone
15:21:19 [krisk]
people can always submit tests just like today..
15:22:15 [krisk]
the only change would be to create a branch at a point in time and then ask for feedback on all tests at that point in time
15:22:17 [jgraham]
Fine. Then I expect the change to mostly be a no-op, but if you think it is helpful then I can't really complain
15:22:57 [jgraham]
Since we are continuously asking for feedback on all submitted but unapporoved tests that seems like it isn't much of a chnge
15:24:26 [krisk]
how are the parser tests going?
15:24:51 [jgraham]
They need to be updated a bit because I made some stuff in the harness private
15:25:02 [jgraham]
and I need to update to latest html5lib
15:26:48 [jgraham]
But that is no problem
15:27:35 [krisk]
were you able to get a basic parser test to pass in all browsers?
15:29:20 [krisk]
The document.write seems to work in all browsers but not the data URI (fails in chrome)
15:30:01 [jgraham]
Yeah, I think that is a chrome bug due to over-enthusiastic blocking of reading the DOM in data URI child frames
15:30:23 [jgraham]
I'm not really sure of a good way to work around that
15:31:31 [krisk]
it may not be a bug - rather by design
15:31:48 [krisk]
...for security
15:32:04 [jgraham]
Well the spec will have to rule one way or the other
15:32:21 [jgraham]
It doesn't make much sense as a security restriction to me
15:34:48 [krisk]
I have not tried safari - let me check...
15:42:18 [krisk]
seems to work...
15:42:33 [krisk]
I might be good to make the first test just pass on all browsers
15:43:06 [krisk]
...then you could know that the harness stuff works
15:45:13 [krisk]
I need to head off - shall we adjourn?
15:45:20 [jgraham]
Sure
15:45:31 [krisk]
rrsagent, generate minutes
15:45:31 [RRSAgent]
I have made the request to generate http://www.w3.org/2011/05/17-htmlt-minutes.html krisk
16:17:07 [Ms2ger]
Ms2ger has joined #HTMLT
17:15:23 [Zakim]
Zakim has left #htmlt