Third parameter of arc in path data should allow signed numbers

[dalboris]

Hi everyone,

As I'm currently writing a parser for SVG path data, I'm able to provide a new pair of eyes for the new grammar definition introduced in SVG 2. Disclaimer: I'm not a grammar specialist, but I did study those things at University a few years back, so hopefully I'm not too rusty.

I am now reasonably familiar with the original SVG 1.1 grammar:

https://www.w3.org/TR/SVG11/paths.html#PathDataBNF

The one currently drafted in:

https://svgwg.org/svg2-draft/paths.html#PathDataBNF

And I have read the following threads to get a sense of the history and rationale of some of the modifications:

#335
#331
#286
#325

Overall, I think the SVG 2 draft grammar reads much better than the SVG 1.1 grammar, thank you for that! I may suggest minor tweaks later to improve readability, but for now a more serious issue. I think the following:

elliptical_arc_argument::=
    number comma_wsp? number comma_wsp? number comma_wsp
    flag comma_wsp? flag comma_wsp? coordinate_pair

should be:

elliptical_arc_argument::=
    number comma_wsp? number comma_wsp? coordinate comma_wsp
    flag comma_wsp? flag comma_wsp? coordinate_pair

In other words, the third parameter of an arc should be coordinate rather than number (it can be negative). The original SVG 1.1 grammar was the following:

elliptical-arc-argument:
    nonnegative-number comma-wsp? nonnegative-number comma-wsp? 
        number comma-wsp flag comma-wsp? flag comma-wsp? coordinate-pair

This seems an unintentional mistake, as I haven't seen this discussed, but maybe I missed it.

Cheers,
Boris

[dalboris]

(Note that in SVG 1.1 grammar, "number" means "signed number", while in the SVG 2 draft it means "unsigned number", which may have contributed to the confusion)

[AmeliaBR]

The third value of arc represents a rotation angle, so there's no logical reason why it can't be negative. And a quick test in all the browsers I had easy access to shows that they accept it.

So I agree, this was probably an accidental oversight when renaming the grammar syntax tokens.

[dirkschulze]

The bot didn't add the discussion of the SVG WG. Here the link as reference: https://www.w3.org/2019/11/20-svg-minutes.html#item05

The SVG WG confirms that this has been an oversight and should be corrected.