- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Thu, 25 Sep 2014 12:14:33 -0700
- To: Juergen Roethig <roethig@dhbw-karlsruhe.de>
- Cc: www-svg <www-svg@w3.org>
On Thu, Sep 25, 2014 at 12:04 PM, Juergen Roethig <roethig@dhbw-karlsruhe.de> wrote: > Robert Longson wrote: >>> >>> That's what I was talking about: In a similar way as you may reference an >>> external stylesheet in an HTML file (via the <link> tag), you may reference >>> the very same external stylesheet in an SVG document (not via a <link> tag >>> which does not exist in SVG, but via a <?xml-stylesheet ?> declaration, as >>> the usual way for XML files to reference CSS files). Your assumption "[...] >>> the SVG must then be complete in a single file" is simply wrong - your SVG >>> file may reference CSS files as well as >> >> >> No my assumption is right, otherwise you have a privacy leak. >> https://bugzilla.mozilla.org/show_bug.cgi?id=628747 contains a detailed >> discussion of this issue. >> >>> JavaScript files, it may even be generated by another (non-SVG) XML file >>> which references an XSLT file which might again reference other files ... >> >> >> javascript isn't allowed in svg-as-an-image either. You'll find all >> existing UAs enforce these rules. > > > So, some existing UAs do it like that (obviously even differently, some > allow parts, some disallow all), but there is not any statement in SVG 1.1 > (that's the very same specification you were referring to) saying what it > must be like in that case! I tend to argue that in such a case the browser > incompletely implements SVG 1.1. If this incompleteness (to not allow > whatever external resources in SVG files loaded via <img>) is a real must, > this might result in a statement in future standards (SVG 2.0 or whatever), > or in the relevant HTML5 "standard" giving such restrictions to SVG files > which are included via <img>. But why should SVG 1.1 be amended with such a > statement? SVG 1.1 is a standard since 2003, second edition since 2011, and > just because that, we should have a third edition? > > Once again: I do not see the need to modify the existing standard SVG 1.1 > (that's the one you were explicitly referring to) to reflect a restriction > which is obviously an individual browser implementation restriction. In the > same way, we should not remove some filters from SVG 1.1 just because some > browser vendors were not able to implement them up to now. Regardless of whether or not SVG1.1 is changed, the SVG Integration spec <https://svgwg.org/specs/integration/> covers the topic here. This is not a matter of "some browsers failing to do something", but all the browsers purposely making an unspecified behavior change, and the spec finally catching up to them. ~TJ
Received on Thursday, 25 September 2014 19:15:21 UTC