I called into the conf call incase someone else dials in

Though like most meeting I suspect this will be on IRC

Agenda Item #1 Bugs on approved tests

Looks like we have two bugs

First is http://www.w3.org/Bugs/Public/show_bug.cgi?id=13767

<Ms2ger> We should probably just remove these tests

I agree as well

jgraham/gsnedders do you object?

<jgraham> (looking)

<jgraham> Yep, delete them

So we have consensus that these tests should be removed

Next bug http://www.w3.org/Bugs/Public/show_bug.cgi?id=13875

This seem like more of an ask for the w3c systems team?

Can you explain more Ms2ger?

Php seems to be universal in terms of support and documentation

<jgraham> Well it is not universal in that Mozilla don't use it for their testing

<jgraham> OTOH, there are clearly cases that won't be covered by less than a full blown language

<Ms2ger> Right; we will need a full blown language for some tests

<Ms2ger> I think it's still useful to limit the use of such a language to the cases where it's actually necessary

I'm not opposed to have another way to create tests if we are limited by what we have in place today

This is why MS helped get the w3c a Hg server

<jgraham> I don't object to having a simple method for common cases if it is trivial to implement via apache

<jgraham> The W3C server is useless for our internal testing

Ms2ger can you add an example test that can't be done with php that is possible using the httpd.js ?

<jgraham> I don't know what it is like at Microsoft, but here using external servers for tests is forbidden for reliability reasons

<Ms2ger> This isn't about tests that can't be done using PHP

<Ms2ger> It's about tests that are easier to write without it

<Ms2ger> The text/plain test you commented on, for example, could be done with a one line ^headers^ files

<jgraham> Ms2ger: Can you write an implementation using only common apache modules?

<Ms2ger> I don't know, I hope so

<jgraham> I think we should revisit this once the implementability has been demonstrated

<Ms2ger> Agreed

Microsoft doesn't use apache or php internally for testing

<jgraham> Intresting. So you are screwed either way :)

Though translating from asp.net to php is trivial

<jgraham> But what Ms2ger suggests is likely to also be trivial in IIS (I guess)

<Ms2ger> Presumably as trivial as in Apache

<jgraham> Yeah, as trivial or not I suppose I mean

<jgraham> t

so what is httpd.js an appache module?

<jgraham> A webserver written entirely in javascript aiui

<jgraham> Nothing to do with apache

<Ms2ger> What jgraham said :)

One way to help is to submit this to Hg and send email to the list saying mozilla would like to give this to the w3c for possible use for testing

<jgraham> (the point of bringing up apache is that I assume it is relevant to the W3C systems team)

I'm not against this just need more information

<Ms2ger> My idea was to re-implement the file format for apache

<jgraham> Right, lets see if that is possible first them come back to this bug

<jgraham> +'

<Ms2ger> (Though I have to admit I don't have a lot of experience with server side stuff)

Ok ms2ger can you add more to the bug?

<Ms2ger> Will do

Agenda item #2 PHP

I just wanted to note that I pushed a php script and it worked just fine

<Ms2ger> Excellent

http://w3c-test.org/html/tests/submission/Microsoft/async/delay.php

It's used for some async script tests

async is pretty important for browser interop

since if all browsers run script differently a page is never going to work the same across browsers

<Ms2ger> I'll see if I can get someone to review them

I was going to send a private email to phl to let be aware of the php in case the systems folks wants to review

though it's a very trivial php file

<Ms2ger> I thought they got alerted automatically, but might be good to confirm that with plh

Agenda item #3 Test submissions

I submitted some tests from microsoft

some use async_test

for example http://w3c-test.org/html/tests/submission/Microsoft/async/async_003.htm

Beyond if the test is a normative requirement I would be good to have feedback on the use of async_test

I also didn't find a clean way to have a test timeout and fail

I though the setup({timeout:100}) would do this but it didn't

<jgraham> If a test times out it times out

<Ms2ger> You're always running the timeout function before that kicks in, right?

<jgraham> There isn't a way to say "a timeout is equivalent to a fail"

<Ms2ger> I'd say timeouts are always failures

<Ms2ger> But that's a question for the runner, really

<jgraham> Yes

<jgraham> That's my view

<jgraham> It is OK to write a test where timeout implies fail because any sane runner will realise that a timeout isn't a pass result

IMHO I added a timer to the async_tests that forces a fail before the setup({timeout}) occurs

<jgraham> That is OK too ofc

<jgraham> I don't have a strong opinion on the right style here

Thanks that is why I am asking

<Ms2ger> I think it would be better to get rid of the setTimeout, though... I'll have a look if I can come up with something later

<jgraham> The only problem with splicit in-test timeouts is that they can cause problems if you run the tests on unexpectedly slow platforms

<jgraham> e.g. mobile devices

<jgraham> *explicit

<Ms2ger> Or our test servers during California working hours

<jgraham> Really?

<jgraham> That's one problem I don't recall seeing :)

<Ms2ger> Maybe your test servers weren't written in the nineties :)

<Ms2ger> krisk, anything else about Microsoft's submissions?

I also submitted some sandbox tests, which have another problem...

If you are familiar with the sandbox attribute one feature is that the sandbox content can't run script

So for some of the sandbox tests I had to make the self describing tests

<jgraham> Sounds reasonable

Though for ones that do run script I used the test harness

The other part of sandbox attribute is pluggins

We used pdf as a pluggin since it seems to be widely available - but it's not universal (mobile)

<jgraham> Uh, yeah dunno what to do with plugins

<jgraham> Isn't there a NPAPI test plugin?

<Ms2ger> I believe Mozilla has one

<jgraham> Could we use something like that?

<Ms2ger> I don't know anything about it, though

<Ms2ger> I can ask around if it would be useful cross-browser

<jgraham> I think we might have/use something similar but I know nothing about it. Or even if it exists really. I don't like plugins :)

I'm open to using something else as long as it works on Windows

<Ms2ger> I'm open to anything as long as it works on Windows, Linux, and OSX

<Ms2ger> (Picky!)

I just wanted to raise awareness on why a test uses pdf

<jgraham> Sure. But I think depending on an external plugin that we can't put in hg is a bug

agree

<Ms2ger> I'll report to the list if I figure out something about Mozilla's plugin

thanks

shall we adjorn?

<Ms2ger> Infraware submitted testharness.js-using tests, btw

<Ms2ger> But the email didn't make it to the list

<Ms2ger> I emailed plh about that

good to hear

This is scribe.perl Revision: 1.136  of Date: 2011/05/12 12:01:43  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

No ScribeNick specified.  Guessing ScribeNick: krisk
Inferring Scribes: krisk

WARNING: No "Topic:" lines found.


WARNING: No "Present: ... " found!
Possibly Present: Microsoft Ms2ger jgraham krisk
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy


WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting

Agenda: http://lists.w3.org/Archives/Public/public-html-testsuite/2011Aug/0014.html

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Got date from IRC log name: 23 Aug 2011
Guessing minutes URL: http://www.w3.org/2011/08/23-htmlt-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.


WARNING: No "Topic: ..." lines found!  
Resulting HTML may have an empty (invalid) <ol>...</ol>.

Explanation: "Topic: ..." lines are used to indicate the start of 
new discussion topics or agenda items, such as:
<dbooth> Topic: Review of Amy's report