Review implementation support - Paths chapter

[ewilligers]

Changes:

Added new B and b "bearing" path commands. [At Risk]

Tests show no browsers support bearing commands.

Allow Z or z to fill in missing path coordinate data in the previous command with the coordinate of the initial point in the subpath. [At Risk]

Tests show no browsers support z filling in missing path coordinate data.
InkScape plans to implement.

Define ‘path’, ‘polygon’ and ‘polyline’ elements with no data set (empty or zero valid commands) to not render.

Tests show Chrome, Edge, Firefox, Safari support this functionality.

Removed the SVGPathSeg* and SVGAnimatedPathData interfaces and the related methods on SVGPathElement.

Tests show only Chrome has removed the interfaces and related methods.

Promoted the ‘d’ attribute to a property. [At Risk]

Tests show Chrome exposes 'd' as a property, with an implementation that differs from the current spec.

Removed the pathLength attribute, getTotalLength(), and getPointAtLength() methods from SVGPathElement, they are now on SVGGeometryElement. [At Risk]

Tests show lack of browser support.

Clarified that a value of zero for ‘pathLength’ is valid.

Tests exist. There are some test failures, these need investigation.

[css-meeting-bot]

The Working Group just discussed https://github.com/w3c/svgwg/issues/384.

The full IRC log of that discussion

<ericwilligers> Topic: https://github.com//issues/384
<BogdanBrinza> topic new
<BogdanBrinza> GitHub: https://github.com//issues/384
<BogdanBrinza> ericwilligers: We'll start with Extensibility chapter
<BogdanBrinza> two changes to the spec
<Chris> great, good to hear
<liam> awesome
<ericwilligers> GitHub: https://github.com//issues/385
<BogdanBrinza> all: sound great - no concerns to make width/height presentation attrbutes
<BogdanBrinza> GitHub: https://github.com//issues/385
<liam> [i think at TPAC we agreed to remove bearing commands, would be good to reaffirm that here]
<BogdanBrinza> ericwilligers: bearing commands - no current implementations, at risk
<BogdanBrinza> https://svgwg.org/specs/paths/
<AmeliaBR> https://www.w3.org/2017/04/svg-2017.html#scope " features which are in the reference draft of SVG2 and which do not meet the stability and interoperability requirements for a Proposed Recommendation may be moved to separate specification modules, work on which would remain in scope, but at a lower priority."
<BogdanBrinza> resolution: remove bearing commands from SVG 2.0 specification, optionally move to https://svgwg.org/specs/paths/
<Chris> Tav: Inkscape plans to implement this
<Chris> BogdanBrinza: OK, but no browser implementation yet
<Chris> AmeliaBR: much asked-for feature
<BogdanBrinza> resolution: remove Z and z from SVG 2.0 and move to https://svgwg.org/specs/paths/
<BogdanBrinza> no issues with path, polygon, etc *
<BogdanBrinza> "Removed the SVGPathSeg* and SVGAnimatedPathData interfaces and the related methods on SVGPathElement."
<BogdanBrinza> tests suggest that only Chrome removed it - shouldn't be a blocker to removal from the specification
<BogdanBrinza> krit: are browsers interested in implementing new features? specifically mesh gradients, new path commands
<BogdanBrinza> krit: if they're not implemented will we have just another revision of SVG 1.1?
<BogdanBrinza> AmeliaBR: the discussion is really about this year, in time for candidate recommendation
<BogdanBrinza> Chris: suggest we keep focused on the discussion on the agenda
<BogdanBrinza> BogdanBrinza: (getting back to GitHub discussion)
<BogdanBrinza> ericwilligers: Promoted the ‘d’ attribute to a property. [At Risk]
<Chris> yes, keep this one in the spec
<BogdanBrinza> Tav: considering implementation in Inkscape
<BogdanBrinza> ericwilligers: "Removed the pathLength attribute, getTotalLength(), and getPointAtLength() methods from SVGPathElement, they are now on SVGGeometryElement. [At Risk]"
<BogdanBrinza> ericwilligers: no browser implementations
<BogdanBrinza> AmeliaBR: Chrome should support this
<BogdanBrinza> ericwilligers: will double check the test
<BogdanBrinza> resolution: this is (same) interface change - keep this at risk
<BogdanBrinza> ericwilligers: "Clarified that a value of zero for ‘pathLength’ is valid."
<BogdanBrinza> resolution: not at risk - just a clarification to a spec
<liam> [big thank you yes]
<Chris> nice one, ericwilligers

[longsonr]

Firefox nightlies support pathLength attribute, getTotalLength(), and getPointAtLength() on SVGGeometryElement. All shapes are now SVGGeometryElement and Firefox nightly passes the SVGGeometryElement-rect.svg test.

Firefox is also removing the parts of the SVGPathSeg* and SVGAnimatedPathData interfaces that manipulate paths. The interface remains currently but only allows retrieval of path data.

[boggydigital]

Not blocking updated 2.0 CR publication - assigning 2.0 Recommendation milestone to clean this up before 2.0 REC

[Tavmjong]

Inkscape (trunk) exposes 'd' attribute as property as speced (rendering only for now). Doing differently introduces complications as we handle attributes and properties with the same code paths.