SVG Working Group Teleconference -- 31 Jul 2014

<trackbot> Date: 31 July 2014

<Smailus> Not me this time, I'm on mute.

<scribe> scribe: Nikos

<scribe> scribenick: nikos_

Should css pseudo-elements (first-line, first-letter) be required in SVG2?

heycam: are you asking specifically for these two?

ed: those are the two that occur in the css 3 spec
... not sure if more have been added in css 4
... FF has support for first-line and first-letter
... we have a bug on blink to support it
... do we want to have this in SVG or not?

heycam: I think we should have it
... I added support because it was easy to do

ed: I suspect Blink implementation will be simple

<TabAtkins> Have to define nesting compatibly with CSS (which'll be tricky, because we don't define nesting well in the first place).

Tav: I think we'll have a hard time implementing but I think it's a good idea

<TabAtkins> Because they can cross a <tref>.

Tav: it doesn't work in FF for fill

<TabAtkins> Sorry, <tspan>.

<heycam> TabAtkins, tref has been removed

ed: that's right

<heycam> TabAtkins, ok

heycam: the other pseudo elements from css 2.1, the generated content ones, those FF doesn't support I don't think
... they were tricky to handle
... in principle I don't think we need to not support them

nikos: why were some of the pseudo classes left out?

ed: they're specific to HTML mostly

<TabAtkins> Until we define interop between SVG and CSS layout models, agree that the other pseudos like ::before aren't really includable.

ed: required the CSS layout engine which SVG didn't have
... it's hard to add generated content in SVG because you'd need to wrap it in a text element

<heycam> g::before { shape-inside: ... }

ed: so if we were to require the first-line and first-letter pseudo elements, how much would we need to define in the SVG spec? if anything?

<TabAtkins> I mean, we could treat it as a <text>.

<TabAtkins> ed: Need to define which SVG properties apply to them.

heycam: I haven't looked at the state of the text chapter recently. But if we have wording that says take this DOM subtree and treat text elements as blocks and tspans as inline and layout according to CSS box model, etc

<TabAtkins> Only those that don't expose that they may be broken up into multiple elements.

heycam: then you mention first-line and first-letter should be included as part of that

<ed> TabAtkins: yes... but without giving a y position too at the same time it would be outside the viewport in svg

<TabAtkins> Right, you'd support x/y props on them too in that case.

heycam: it should just be a small change
... not sure about Tab's concerns. Perhaps he can go into more detail in an email?

<heycam> concerns re nesting

<TabAtkins> Yup, will send.

ed: second half of my question was related to the outcome of this. Should we require CSS selectors in SVG or is that just going to fall out from requiring other CSS specs?

heycam: not sure if any of the CSS specs we've talked about have a strict requirement on a particular version of selectors
... so it might be good for us to have a requirement for a particular version

ED: currently we have an informal reference. We could make it normative

heycam: various parts of CSS that we're not relying on for SVG specific things. Then stuff that's in CSS 2 is probably good enough

Tav: will we be able to select 3rd rect in a group?

heycam: not sure which version of selectors that is

ed: think it's 3

<ed> http://www.w3.org/TR/css3-selectors/#structural-pseudos

heycam: maybe we should decide what the baseline version of features that SVG should support

<TabAtkins> Man, don't require a particular level. That's silly. Just point to Selectors and let people support the latest version.

heycam: don't know how far ahead we should normatively require when they're pretty orthogonal things

<TabAtkins> CSS doesn't require a particular version of Selectors.

<heycam> TabAtkins, how about a minimum version?

Tav: would be interesting to know if anyone supports 3rd rect in a group now

<TabAtkins> Sure.

heycam: think it should work

ed: I think so
... don't know if we have wording in the spec about not supporting style sheets

heycam: I think we had the discussion about requiring style sheets
... but not sure spec changes have been done

ed: think we've already resolved to have multi-line text. I think first-line makes sense once that's in

<TabAtkins> Agree.

ed: do we want to resolve on first-line and first-letter now or wait for more feedback?

heycam: Some details to finalise, but I think it makes sense to resolve now

RESOLUTION: SVG 2 will require ::first-line and ::first-letter from CSS selectors

<scribe> ACTION: Erik to add ::first-line and ::first-letter to SVG 2 [recorded in http://www.w3.org/2014/07/31-svg-minutes.html#action01]

<trackbot> Created ACTION-3635 - Add ::first-line and ::first-letter to svg 2 [on Erik Dahlström - due 2014-08-07].

Tav: can you do selectors inside the style attribute?

ed: no don't think so

Tav: So this would be the first case when you actually need a style sheet

heycam: also if you want to have :visited, :hover, etc

ed: selectors in general
... regarding the second part of the question, do we require CSS3 selectors
... do we want to resolve on that now too?

<TabAtkins> We tried selectors inside of style years ago, but didn't end up doing it.

ed: currently we reference a minimum version

heycam: don't think we have any wording about supporting later versions of CSS
... might be good to have something in the style chapter
... I think in general we should require the earliest specs that we actually need for features
... so if we don't require any CSS3 features just stick with the older version
... and say if you want to support later features that's fine

ed: I can take the action to state that

heycam: do you want to tackle that in a general fashion in the style chapter?

ed: yes just a paragraph in there

<scribe> ACTION: Erik to add wording to SVG 2 to allow later versions of CSS [recorded in http://www.w3.org/2014/07/31-svg-minutes.html#action02]

<trackbot> Created ACTION-3636 - Add wording to svg 2 to allow later versions of css [on Erik Dahlström - due 2014-08-07].

<heycam> https://www.w3.org/Graphics/SVG/WG/wiki/F2F/London_2014/Agenda_proposals

<ed> trackbot, end telcon

- DRAFT -

SVG Working Group Teleconference

31 Jul 2014

Attendees

Contents

Should css pseudo-elements (first-line, first-letter) be required in SVG2?

Summary of Action Items

Scribe.perl diagnostic output