W3C home > Mailing lists > Public > www-svg@w3.org > September 2012

Re: ACTION-3388: Investigating ArcThrough

From: Tab Atkins Jr. <jackalmage@gmail.com>
Date: Tue, 25 Sep 2012 10:38:29 -0700
Message-ID: <CAAWBYDC0V0MZ9WPVzaJ6i1-HuJ67yKSSAsYcVG94vK_qrdN7QA@mail.gmail.com>
To: Cameron McCormack <cam@mcc.id.au>
Cc: Robin Berjon <robin@w3.org>, www-svg <www-svg@w3.org>
On Tue, Sep 25, 2012 at 2:30 AM, Cameron McCormack <cam@mcc.id.au> wrote:
> Robin Berjon:
>> This appears to introduce as a side-effect the ability to draw an
>> infinite line, which we don't currently have. I just wanted to point
>> this out to make sure that it does not cause any issues.
>
> It makes it unclear how patterns and gradients would paint, if they are
> using objectBoundingBox units.  Various DOM methods that don't currently
> return Infinity can now do so (getComputedPathLength(), getBBox()
> properties).

Good points.  Hm.  I don't see a way around it, though, without
introducing discontinuities.  You can make the circle arbitrarily
large by getting the through point *nearly* colinear and far from the
source points.

> Also, what if you have a marker-end that references an overflow:visible
> <marker> that itself includes an infinite line?

Luckily, it's infinite in a convenient way - the start and end are
both at finite locations (the start and end points).  It wraps around
at infinity.


On Tue, Sep 25, 2012 at 10:33 AM, Jasper van de Gronde
<th.v.d.gronde@hccnet.nl> wrote:
> Also, what would you do with the fill? If you look at the filled disk, it
> becomes a half-plane. Apart from infinite shapes being a novelty in SVG,
> this also exposes an instability. If the point approaches the line through
> the start and end points from one side it becomes a half-plane on that side,
> if the approaches from the other it becomes a half-plane on the other
> side...

Ooh, fill is a very good point - it's indeterminate which side is
inside and which is outside.

Okay, then, I relent.  In that degenerate case, we should say that no
path is drawn, and it simply moves to the end point.

~TJ
Received on Tuesday, 25 September 2012 17:39:17 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:52 GMT