6947 2009-05-23 22:17:17 +0000 Inline SVG fragments test suite please 2009-05-24 18:14:57 +0000 1 1 1 Unclassified SVG Concepts SVG 1.1 Full Macintosh All NEW P2 normal Test Suite 1 Lars.Hellstrom schepers bzbarsky helder.magalhaes www-svg oldest_to_newest 25273 0 Lars.Hellstrom 2009-05-23 22:17:17 +0000 I'm new at SVG, so it's quite possible I'm missing something, but... It seems the test suite, as presented at http://www.w3.org/Graphics/SVG/Test/20061213/, only contains tests where the SVG is (i) a stand-alone SVG Web page or (ii) embedded by reference into a Web page. The Concepts chapter of the SVG 1.1 spec also lists (iii) embedding inline as a way of including SVG content, so it seems reasonable that this approach should be tested as well. Especially if browsers might process it differently. As far as I can tell, Safari (3.2.1) uses a plugin for rendering SVG embedded by reference, but renders inline SVG itself. Firefox (2.0.0) seems to not render text in inline SVG although it renders text in referenced SVG, which also suggests the two ways are handled differently. This latter case was what prompted me to seek the test suite out in the first case, but it could just as well be me that has gotten the code wrong. Either way I'd appreciate an official test covering inline SVG so that I can use it to determine if it is my fault or Firefox's. 25282 1 helder.magalhaes 2009-05-24 11:22:17 +0000 (In reply to comment #0) > I'm new at SVG, so it's quite possible I'm missing something, but... Well, you may double check that by crawling through the ("www-svg" and "public-html") mailing lists and/or creating a new thread... ;-) > It seems the test suite, as presented at > http://www.w3.org/Graphics/SVG/Test/20061213/, only contains tests where the > SVG is (i) a stand-alone SVG Web page or (ii) embedded by reference into a Web > page. The Concepts chapter of the SVG 1.1 spec also lists (iii) embedding > inline as a way of including SVG content, so it seems reasonable that this > approach should be tested as well. Especially if browsers might process it > differently. I share your opinion, but please note that SVG in HTML is still being discussed. A proposal [1] already exists, and there has been warm discussion [2] on HTML integration (more recently integration into HTML5). Naturally, somewhere in time, I hope a test suite will be derived... :-) > As far as I can tell, Safari (3.2.1) uses a plugin for rendering SVG embedded > by reference, but renders inline SVG itself. Firefox (2.0.0) seems to not > render text in inline SVG although it renders text in referenced SVG, which > also suggests the two ways are handled differently. This should only happen if you have an SVG viewer installed which registers itself for the SVG MIME type. As far as I know, when default settings are used, both implementations (as well as Opera and more) natively render SVG. Finally, I'm not sure if "Concepts" component is the most appropriate ("Compositing" and or "Extensibility" maybe?). Currently there's no specific "Integration" (with/into other technologies) component, so I take this opportunity to propose its creation (given that demand on this is raising as implementations get better). Hope this helps, Helder [1] http://dev.w3.org/SVG/proposals/svg-html/svg-html-proposal.html [2] http://www.nabble.com/forum/Search.jtp?query=SVG+HTML&local=y&forum=11732 25284 2 bzbarsky 2009-05-24 12:39:01 +0000 > please note that SVG in HTML is still being discussed. That has nothing to do with this issue, though. SVG in XHTML or other XML documents is in fact supported by the existing XML and SVG specifications, and SVG even has conformance requirements for the sizing of such SVG fragments... And those are not tested, at the very least. 25285 3 Lars.Hellstrom 2009-05-24 18:14:57 +0000 (In reply to comment #1) > (In reply to comment #0) > > I'm new at SVG, so it's quite possible I'm missing something, but... > > Well, you may double check that by crawling through the ("www-svg" and > "public-html") mailing lists and/or creating a new thread... ;-) Ah, so those have public archives? (I've too often been "burnt" by communities that keep their mailing list archives subscribers-only to have considered that as a place to start.) > > It seems the test suite, as presented at > > http://www.w3.org/Graphics/SVG/Test/20061213/, only contains tests where the > > SVG is (i) a stand-alone SVG Web page or (ii) embedded by reference into a Web > > page. The Concepts chapter of the SVG 1.1 spec also lists (iii) embedding > > inline as a way of including SVG content, so it seems reasonable that this > > approach should be tested as well. Especially if browsers might process it > > differently. > > I share your opinion, but please note that SVG in HTML is still being > discussed. A proposal [1] already exists, and there has been warm discussion > [2] on HTML integration (more recently integration into HTML5). Naturally, > somewhere in time, I hope a test suite will be derived... :-) As remarked in comment#2, I was thinking about SVG inline in XHTML, not HTML. The "Web page" wording is verbatim from the SVG specification. > > As far as I can tell, Safari (3.2.1) uses a plugin for rendering SVG embedded > > by reference, but renders inline SVG itself. Firefox (2.0.0) seems to not > > render text in inline SVG although it renders text in referenced SVG, which > > also suggests the two ways are handled differently. > > This should only happen if you have an SVG viewer installed which registers > itself for the SVG MIME type. As far as I know, when default settings are used, > both implementations (as well as Opera and more) natively render SVG. The plugin may be a leftover from Safari 2 days, or something. (Apple's Migration Assistant has a tendency to bring along all sorts of things one doesn't know about.) > Finally, I'm not sure if "Concepts" component is the most appropriate > ("Compositing" and or "Extensibility" maybe?). Currently there's no specific > "Integration" (with/into other technologies) component, so I take this > opportunity to propose its creation (given that demand on this is raising as > implementations get better). My initial impulse was that it should belong with "Test Suite", but that wasn't a component. Failing that, I (as per the Bug Writing Guidelines http://www.w3.org/Bugs/Public/page.cgi?id=bug-writing.html) looked for a "General" component, but that didn't exist either. The page of "component descriptions" (http://www.w3.org/Bugs/Public/describecomponents.cgi?product=SVG) didn't do much for describing the components, but at least it seems clear that the component division is merely according to the chapters of the specification, and as it happens it is in the "Concepts" chapter that it is claimed that SVG can be embedded inline in XHTML... Some reform of the component division seems appropriate.