SVG Working Group Teleconference -- 22 Jul 2008

<trackbot> Date: 22 July 2008

<scribe> scribe: erik

<scribe> scribeNick: ed_

1.2T testsuite

<aemmons> http://www.w3.org/Graphics/SVG/Group/wiki/TestCoverage

<aemmons> http://lists.w3.org/Archives/Public/public-svg-wg/2008JulSep/0064.html

AE: this is the status update doug sent out

DS: just a short update

<ChrisL> I just checked in struct-cond-08-t.svg which tests requiredFonts

AE: so you changed the table a bit, and took out some of the items to a separate table
... we have a list of tests to be done
... I've done timer
... any other updates?

CL: added a test for requiredFonts
... opera passes, bitflash fails

CM: mine can be moved to the bottom as well

CL: when I was doing the test, we have an untested condition if you have a list of requiredFonts

AE: add that to the list?

CL: yes
... requiredFonts with listvalues

AE: updating the table now
... any other updates?
... guessing AG hasn't had time to do his share

<ChrisL> Are the instructions at http://www.w3.org/Graphics/SVG/Group/repository/docs/SVGTestSuite-howto.html still the correct ones?

AE: are the assignments ok with you CL?

DS: the ones I mentioned in my email, i'll just do those

AE: requiredFeatures and requiredExtensions are covered, yes?

DS: someone had written that there was only struct-condition-1-t
... but since then there has been tests added
... each of those attributes have been tested except requiredFonts
... there are two tests basically

<ChrisL> $ grep requiredExtensions *.svg

<ChrisL> struct-cond-01-t.svg: <rect fill="red" x="0" y="0" width="220" height="150"

<ChrisL> requiredExtensions="http://example.org/bogus" />

AE: there is some work to be done then?

DS: yes, as I mentioned in the mail

CL: requiredFormats is covered by struct-cond-204

AE: CM do you have time to do some in the next few days?
... the connection ones maybe?

CM: those will need some serverside stuff, but sure I can give it a shot

AE: just let AG know that you're working on them
... until thursday, do people intend have all tests done?

CM: not all

DS: plan to

CL: took me a while to get started with the new cvs base
... we need to update the testsuite howto page
... does anyone have a pointer to the new one?

DS: AG was going to update it

CL: right, but this hasn't been updated in a while

DS: i'll move this to the wiki and update it

CL: yes, we need to have the shortnames correct since people will use different names otherwise

DS: doing the update right now

AE: great progress everyone
... we still have outstanding actions on the old tests, and we need to review and add the new tests
... we should review the tests before the f2f
... good that we have this table with tests to be reviewed

CL: i thought we would list the tests we've made, to show that we've made progress

AE: we still have the list of recently reviewed tests
... looking a bit at the open actions against the testsuite
... most are in our member-only tracker

<aemmons> http://www.w3.org/Graphics/SVG/Group/track/products/13

AE: you can go on forever with some of these comments
... we need to be prudent and pragmatic and I don't think we have to address all of them to get the testsuite out

CL: some of them are cosmetic yes, like change this color to something else

AE: it would be nice to clear out these open issues
... we do have alot of open actions
... so we need to triage them
... which is again something that has to be done between now and the f2f
... tedious work, but it is something we need to focus on
... should we do a marathon telcon, with testsuite as the focus?

CL: sure

AE: what's a good date for that?
... I'm away next week

ED: i'm away the first couple of weeks in august

AE: other people, beginning of august?

CL: fine with me

AE: starting at the normal telcon time?

CL: maybe an hour earlier?
... and going for four hours

AE: typically an hour of discussion and an hour of actions

<scribe> ACTION: AE to mail the svgwg-list about the marathon telcon [recorded in http://www.w3.org/2008/07/22-svg-minutes.html#action01]

<trackbot> Created ACTION-2106 - Mail the svgwg-list about the marathon telcon [on Andrew Emmons - due 2008-07-29].

AE: do we think we will be ready with the testsuite after the f2f?

CM: what's the remaining work?

AE: review, approve and generate references
... and we have actions against old tests too
... we haven't published the one we made in nice

CL: yes, it is in the public cvs now though
... but we should make a release with a zip file
... there is the problem with moving files to a new repository
... and revision numbers, but you can do a script to replace the dollarsigns so that cvs don't replace those

CM: wasn't this already done for the testsuite?

CL: it's a non-changing snapshot

AE: so we can use the tricks that mikesmith suggested

CL: I still think that just making a script to replace the dollarsigns is easier

AE: does someone have the time to publish the testsuite? and is that more important than doing new tests?

DS: and also editing D3E

AE: I think it's more important to work on the new tests
... but it would be nice if we could push it out before the f2f

CL: let's see where we are in a few weeks
... and maybe push it out then

<script> and externalResourcesRequired="" (ACTION-2063)

<heycam> http://lists.w3.org/Archives/Public/public-svg-wg/2008AprJun/0149.html

<shepazu> http://www.w3.org/2007/03/mth/harness

DS: the css wg has been working with the mobile web people, to make a testsuite harness
... just take a look and we can see if we can move in a similar direction
... they have their own testresults
... obviously this is really small, but they have their own implementation report
... using this would allow us to step through the testsuite more quickly
... though it would need an svg harness, this one uses a html harness
... so this probably wouldn't work in batik or bitflash, but we could make a similar one in svg
... if we could do this quickly then it might be worth the effort

CL: we could generate a new harness similar to this

DS: maybe after we've done the testsuite

AE: any improvement we can do to our testsuite will be good

DS: the w3c has been working on a new testsuite licensing policy
... it's being review
... allows making changes to tests, and republishing

AE: so, back to the script and eRR issue again...

CM: there are three questions
... does eRR apply to script elements
... spec is a bit unclear

CL: the intention was to require the resources to be available
... so the paragraph that mentions rendering may be an oversight
... usually you want to have all resources just like if you had it locally

CM: second question, should eRR=true require some resource to be there?
... so, that if you had eRR there on anything that causes a 404 require other script elements not to run?

CL: eRR can be on any subtree, so you only stop the subtree

CM: so if you had script elements before and after but eRR on only one subtree
... you'd get the other one
... executed first, and the other one when it becomes available
... in HTML it blocks if you have a defer attribute on it

CL: scripts in html are in the head section?

DS: they can be anywhere

<ChrisL> yeah, of course they can (d'oh!)

CM: opera implements eRR?

ED: yes, 9.5 does, the interaction with scripts may need some testing though

CM: the wording in the spec about how the tree gets constructed in progressive rendering
... if scripts are executed when inserted we may need to tweak the wording

DS: yes, we don't want to differ from html

CM: right, so batik parses the whole document at one time, but opera and firefox execute script as they are inserted

RESOLUTION: when start-tags get parsed they get appended to the most recently opened element that is still actually open (the top of the stack)

<scribe> ACTION: Cameron to update the 1.2T spec with the reworded section on progressive rendering and tree construction [recorded in http://www.w3.org/2008/07/22-svg-minutes.html#action02]

<trackbot> Created ACTION-2107 - Update the 1.2T spec with the reworded section on progressive rendering and tree construction [on Cameron McCormack - due 2008-07-29].

CM: the last part is about firing load events on script elements
... the scripts can refer to some external resource
... the spec doesn't take into account the case where eRR=true but there is no reference

AE: even if the reference doesn't exist the event is fired

CM: what does it mean that it's not visible in the dom?

AE: accessible via the dom

CL: maybe it has to do with shadowtrees?

CM: if you have some subtree outside the document which you do modifications to do they still generate events?

<ChrisL> you can traverse from a shadow tree but not to it, so it would make sense in that context.

DS: i don't think those gets fired

CM: maybe that's what the sentence is trying to say

DS: maybe that's something for D3E
... it's not a mutation event since the document tree is not actually changed

CM: so you don't get any events if you create and append elements outside the tree (in that tree)?

<scribe> ACTION: DS to make sure DOM 3 Events talks about mutation events outside of the document [recorded in http://www.w3.org/2008/07/22-svg-minutes.html#action03]

<trackbot> Created ACTION-2108 - Make sure DOM 3 Events talks about mutation events outside of the document [on Doug Schepers - due 2008-07-29].

<aemmons> "The mutation and mutation name event modules are designed to allow notification of any changes to the structure of a document"

AE: I think it is pretty clear that you don't get events fired outside the tree
... are we comfortable with the wording as is?

CL: maybe talking about elements that have been creaed but not yet appended
... you wouldn't expect events to be sent to it right?

AE: so maybe it should say if it is in the document tree

CL: yes that's better

CM: ok, so the first issue about script elements without a reference but with eRR=false, do you get a load event?

AE: it would still fire
... I remember we discussed this about a year ago

RESOLUTION: change wording from "accesible in the DOM" to be "in the document tree"

<scribe> ACTION: cameron to make changes for script elements and eRR as discussed in these minutes [recorded in http://www.w3.org/2008/07/22-svg-minutes.html#action04]

<trackbot> Created ACTION-2109 - Make changes for script elements and eRR as discussed in these minutes [on Cameron McCormack - due 2008-07-29].

<aemmons> http://lists.w3.org/Archives/Public/public-svg-wg/2008JulSep/0008.html

http://lists.w3.org/Archives/Public/public-svg-wg/2008JulSep/0008.html

Use of org.w3c.svg package in Java bindings

CM: in the java appendix the packagename has changed from svg 1.1
... was that deliberate?
... [reads the mail]

AE: so JSR-287 should be backwards compatible with 226

CM: any objections to changing it?

<scribe> ACTION: Cameron to change the packagename from org.w3c.svg to org.w3c.dom.svg in 1.2T [recorded in http://www.w3.org/2008/07/22-svg-minutes.html#action05]

<trackbot> Created ACTION-2110 - Change the packagename from org.w3c.svg to org.w3c.dom.svg in 1.2T [on Cameron McCormack - due 2008-07-29].

<aemmons> http://www.w3.org/mid/20080704024005.GB6027@arc.mcc.id.au

exceptions for non-finite floats

CM: I had an action to make various methods throw NOT_SUPPORTED_ERR when NaN or INF are passed to methods that take a float argument

<heycam> ACTION-2082?

<trackbot> ACTION-2082 -- Cameron McCormack to fix exceptions in the specification for setFloat values -- due 2008-07-10 -- CLOSED

<trackbot> http://www.w3.org/Graphics/SVG/WG/track/actions/2082

AE: so you're asking if get(Presentation)Trait(NS) throws a different exception
... I'd be hesitant to change anything there if it didn't already have an exception

CM: these ones had exceptions already

AE: setFloatTrait has type-mismatch-err

CM: ok, so if JSR says that maybe just leave those as is

<heycam> http://dev.w3.org/SVG/profiles/1.2T/publish/svgudom.html#FloatValues

CM: I've added some wording to cover the other cases

AE: ok, good, so I'm fine with those changes

<aemmons> http://lists.w3.org/Archives/Public/public-svg-wg/2008JulSep/0045.html

Unwanted restriction for JPEG images that have colour profiles

CL: so only one comment saying this was good, and it doesn't affect conformance
... any objections to adding this wording?

DS,AE,CM: looks ok

<scribe> ACTION: CL to remove the unwanted restriction for JPEG images that have colour profiles [recorded in http://www.w3.org/2008/07/22-svg-minutes.html#action06]

<trackbot> Created ACTION-2111 - Remove the unwanted restriction for JPEG images that have colour profiles [on Chris Lilley - due 2008-07-29].

CL: so the css wg has just released a css3 draft that removes some parts that svg UA:s implemented but which weren't implemented in html/css UA:s

- DRAFT -

SVG Working Group Teleconference

22 Jul 2008

Attendees

Contents