Re: rotated path segment proposal

So in this case, the second segment in the examples would NOT render in existing implementations – and (IMO) that would be BAD (as it's data loss).

Leonard

From: Vincent Hardy <vhardy@adobe.com<mailto:vhardy@adobe.com>>
Date: Mon, 25 Jul 2011 14:33:27 -0700
To: Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>>, Dirk Schulze <dirk@dschulze.com<mailto:dirk@dschulze.com>>
Cc: Cameron McCormack <cam@mcc.id.au<mailto:cam@mcc.id.au>>, SVG WG <public-svg-wg@w3.org<mailto:public-svg-wg@w3.org>>
Subject: Re: rotated path segment proposal

Hi Leonard,

I think that if an older implementation encountered the new syntax, it would behave as if there was another error in the path value:

http://www.w3.org/TR/SVG/implnote.html#PathElementImplementationNotes

"If a path data command contains an incorrect set of parameters, then the given path data command is rendered up to and including the last correctly defined path segment, even if that path segment is a sub-component of a compound path data command, such as a "lineto" with several pairs of coordinates. For example, for the path data string 'M 10,10 L 20,20,30', there is an odd number of parameters for the "L" command, which requires an even number of parameters. The user agent is required to draw the line from (10,10) to (20,20) and then perform error reporting since 'L 20 20' is the last correctly defined segment of the path data specification."

Vincent

From: Leonard Rosenthol <lrosenth@adobe.com<mailto:lrosenth@adobe.com>>
Date: Mon, 25 Jul 2011 11:12:23 -0700
To: Dirk Schulze <dirk@dschulze.com<mailto:dirk@dschulze.com>>
Cc: Cameron McCormack <cam@mcc.id.au<mailto:cam@mcc.id.au>>, SVG WG <public-svg-wg@w3.org<mailto:public-svg-wg@w3.org>>
Subject: Re: rotated path segment proposal

We are talking about changing the syntax of the path element, yes?  If so,
that means that if an SVG file using these new possible operators in the
value is read by an old/existing reader, the behavior is completely
undefined!  What is it going to do with it?!?!?

Separate segment
<path d="....">
<path d="...">
Instead of
<path d="... R ....">

Leonard

On 7/25/11 1:06 PM, "Dirk Schulze" <dirk@dschulze.com<mailto:dirk@dschulze.com>> wrote:


Am 25.07.2011 um 18:46 schrieb Leonard Rosenthol:

Wouldn't it be easier to just have two separate segments, and then just
change the rotation on the second??   Why would you need a single
segment?
What do you mean with two separate segments?

I simply don't see enough of a use case for this to SIGNIFICANTLY break
backwards compatibility by adding new operators to the path segment
format.  Do we even have any idea what such an addition to do to
existing
solutions when they encountered this??

I don't see how this breaks backward compatibility. If someone used f, r
or even ~ before, the path gets suddenly valid but definitely was invalid
before.

Dirk

Received on Monday, 25 July 2011 21:44:37 UTC