See also: IRC log
<joanie> agenda: this
<joanie> agenda: be done
<richardschwerdtfeger> RRSAgen, make log public
<richardschwerdtfeger> chair: Joanie
<richardschwerdtfeger> not following
<richardschwerdtfeger> chair: Joan Marie Diggs
<joanie> chair: Joanmarie_Diggs
<scribe> scribe: MichaelC
<joanie> QUESTION 2: "SHOULD NOT expose" is not a "MUST NOT expose." Are only "MUST" and "MUST NOT" statements for user agents to be tested? Joanie added test placeholders in the meantime.
jd: questions need to be addressed before writing testable statements
first is asked and answered
<clown> ANSWER: Yes, only MUST and MUST NOT statements are tested. However, an implication of "SHOULD NOT expose unless the columnheader descends from a grid" is "MUST expose if the columnheader descends from a grid". At least I think that's a requirement, but a second opinion is needed.
jd: q ^; clown: a ^
clown: think must expose aria-readonly if columnheader descends from grid
rs: if it´s false?
clown: either way
if must is implied, then there are other tests we have to work out
rs: does readonly propagate?]
clown: yes
rs: agree with answer; just one test
clown: because of an unless elsewhere, this interpretation for columnheader and rowheader
jd: think same question and answer for aria-required, so update similiarly
ss: so need several test cases now?
30
clown: <updates tracking page to reflect these answers and their implications>
jd: gotta propagate stuff
clown: was waiting for second opinion, so now have it
rs: assign this work to Stefan now
ss: <sigh> yes
<munging about the proper edits, and wiki tips and tricks>
jd: other testable statement stuff?
rs: clown, heard from @@?
clown: yes, @@ under debate
stop editing core-aam and just work on tests
rs: great if possible
clown: can multitask now if you like
jd: Stefan and Rich are heroes
rs: helped Bryan on wiki editing
jd: anything else?
ss: how do we create eventing?
rs: at very bottom there´s an example
e.g., event type=selection change
jd: example at top to demonstrate
<joanie> https://www.w3.org/wiki/index.php?title=ARIA_1.1_Testable_Statements#Example_.28Dont_Generate.29
example text, and table drives the test runner
can do a series
ss: just for ATK?
jd: for those platforms that have corresponding features
in relevant sections
ss: ok will try, see what I break
for combobox and aria-invalid id
how?
rs: pick an ID that´s not in the DOM
then it should fail to map
ss: what put in table?
jd: propertyAccessible does not exist or something
or is false
rs: ¨is not¨ is not the way to say it?
<now into too much details to scribe>
assume you´re testing null reaction on change of state?
where is the state change indicated?
rs: oh, still have to put that in
otherwise it looks right?
jd: yes
ss: so I can use this as example of property change?
jd: yes
ss: so in combobox and invalid id, put in ¨not generated¨ something
there should be a clearly described test case
for these sort of situations
jd: will work up more examples
<jongund> yes
jg: so JD building the framework to run tests
rather than try to keep up with that, focus on test adapter
and let you plug into framework
jd: ok
to explain
I´m working in a branch
that´s viewable but pretty shaky right now
trying to come up with common architecture
keep finding things need to change
within that branch, a given ATTA should work
but the code underneath it isn´t the official architecture or stable
it won´t become official until Shane accepts a pull request
so makes sense to leave me loose on this for now
at the moment I´m trying to figure out IA2 architecture
if JG et al could help me with that stuff, it´ll help
jg: we have stuff, now need to get IA2 states and relations
but have 3/4 of what is needed
can work on this scope
but I don´t wanna get involved in refactoring
I could get going on UIA
jd: that would also be awsome
doesn´t look like Microsoft will help with that, in spite of statement at TPAC
jg: ok, think we know what we need to focus on now
jd: can you work in python 3?
jg: using python 2 because WPT uses that
jd: think they´re converting
test adapters are independent of WPT
they can be run completely separately though results are boring
jg: ok, could try python 3
though there are variations of it
jd: >=3.4 <=3.6 is the range I´m using now
I´m mainly using 3.4 but testing in later
jg: ok, can use 3.4
jd: don´t feel bound to it, there are hiccups
would be useful for me to work with you on features of the code
jg: using pyatspi interfaces
jd: <gory details>
leading towards more ATK-like stuff
<joanie> https://github.com/Spec-Ops/web-platform-tests/blob/atk-atspi-atta/wai-aria/tools/atta_base.py
<joanie> https://github.com/Spec-Ops/web-platform-tests/blob/atk-atspi-atta/wai-aria/tools/atta_base.py#L128
scribe: so in sum, that´s what I need
jg: so need is ready?
jd: yes, says we found doc and has specified URI
which is done via a11y tree
so one thing we need is how to get the URI from the ATK interface
the current version of the code is effectively a todo list of things we still need to know how to do
jg: great, very helpful
jd: really glad you´re working on this
jg: hope the work delta between IA2 and UIA won´t be too bad
<jongund> I need to leave for another meeting
jd: because of JG work in IA2, the need is now mainly for UIA
<jongund> Talk to you next week
MS did say they´d provide manual testers
RS, wanna follow up with them on that?
though we have 200+ things to test
either an ATTA or manual testers from them really needed
RS: main goal is CR exit
though still want to do comprehensive testing
for 1.1
1.2 also starting up, browsers want it but aren´t contributing much atm
rs: next week expect to have a draft run by the WG
for review and approval as a First Public Working Draft
Personalization Semantics editors´ draft
js: also COGA issue papers and gap analysis will come out soon
mc: from APA and AG
This is scribe.perl Revision: 1.152 of Date: 2017/02/06 11:04:15 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: Irssi_ISO8601_Log_Text_Format (score 1.00) Succeeded: s/pyscatk/pyatspi/ Found embedded ScribeOptions: -final *** RESTARTING DUE TO EMBEDDED OPTIONS *** Present: Rich_Scherdtfeger Joanmarie_Diggs jongund janina Joseph_Scheuhammer MichaelC Regrets: Michiel_Bijl Bryan_Garaventa Léonie_Watson Matt_King Found Scribe: MichaelC Inferring ScribeNick: MichaelC Found Date: 16 Mar 2017 Guessing minutes URL: http://www.w3.org/2017/03/16-aria-minutes.html People with action items: WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]