Re: [svg-developers] SVG Lint? Or a conformance test for content?
On Thursday, September 22, 2005, 11:30:40 PM, Mike wrote: MCF> Hi all, MCF> I'm working on a product that generates SVG graphics. We've been MCF> recommending use of the Adobe SVG Viewer where it's available, and MCF> Mozilla SVG elsewhere. Since development is done on an AMD64 Linux box, MCF> we use Mozilla SVG ourselves. However, we are getting reports from MCF> users that one of the graphics doesn't show up in Adobe SVG, MCF> particularly, the graphic appears, but is all-white (that is, the text MCF> and icons are present, can be clicked on, and generally seem to have MCF> been generated fine, but they are all white-on-white). The same graphic MCF> renders correctly (modulo the limitations of Mozilla SVG) on MozillaSVG MCF> 1.0.6 on Linux and Firefox DeerPark beta1 on Win32, as well as with KDE MCF> 3.4.1's KSVG 0.1 and Inkscape. MCF> My hypothesis is that the AdobeSVG viewer is picking up some unexpected MCF> feature/property and isn't able to deal with it, while the MCF> Mozilla/KDE/Inkscape engines are slightly more robust in dealing with MCF> that particular problem. To test that I'm wondering if there is a tool MCF> that is extremely anal about conformance that can be used to test MCF> content. An SVG Lint would be very useful (a SVG Tidy even more so) that would strip out Adobe Illustrator Inkscape etc stuff to produce a file for publication. However SVG has clear rules for what to do with content in other name spaces. I don't think that is the issue here, though. MCF> Basically just something that will parse an SVG file and MCF> complain about anything that could cause a problem on a browser. MCF> For reference, a failing file is here: MCF> http://www.vrplumber.com/programming/adoberenderswhiteonwhite.svg MCF> With thanks for any hints, MCF> Mike Change script type="text/javascript" to script type="text/ecmascript". SVG mandates support of ecmascript; other languages (even if in the case of javascript jscript etc they are nearly identical) is not required. -- Chris Lilleymailto:[EMAIL PROTECTED] Chair, W3C SVG Working Group W3C Graphics Activity Lead Co-Chair, W3C Hypertext CG Yahoo! Groups Sponsor ~--> Most low income households are not online. Help bridge the digital divide today! http://us.click.yahoo.com/cd_AJB/QnQLAA/TtwFAA/1U_rlB/TM ~-> - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
Re: [svg-developers] SVG Lint? Or a conformance test for content?
On Thursday, September 22, 2005, 11:30:40 PM, Mike wrote: MCF> Hi all, MCF> I'm working on a product that generates SVG graphics. We've been MCF> recommending use of the Adobe SVG Viewer where it's available, and MCF> Mozilla SVG elsewhere. Since development is done on an AMD64 Linux box, MCF> we use Mozilla SVG ourselves. However, we are getting reports from MCF> users that one of the graphics doesn't show up in Adobe SVG, MCF> particularly, the graphic appears, but is all-white (that is, the text MCF> and icons are present, can be clicked on, and generally seem to have MCF> been generated fine, but they are all white-on-white). The same graphic MCF> renders correctly (modulo the limitations of Mozilla SVG) on MozillaSVG MCF> 1.0.6 on Linux and Firefox DeerPark beta1 on Win32, as well as with KDE MCF> 3.4.1's KSVG 0.1 and Inkscape. MCF> My hypothesis is that the AdobeSVG viewer is picking up some unexpected MCF> feature/property and isn't able to deal with it, while the MCF> Mozilla/KDE/Inkscape engines are slightly more robust in dealing with MCF> that particular problem. To test that I'm wondering if there is a tool MCF> that is extremely anal about conformance that can be used to test MCF> content. Basically just something that will parse an SVG file and MCF> complain about anything that could cause a problem on a browser. MCF> For reference, a failing file is here: MCF> http://www.vrplumber.com/programming/adoberenderswhiteonwhite.svg You also have some attributes like object_name and object_counts on path. Those attributes should be in their own namespace if they are put on svg elements. cx="0.0" cy="0.0" on path?? -- Chris Lilleymailto:[EMAIL PROTECTED] Chair, W3C SVG Working Group W3C Graphics Activity Lead Co-Chair, W3C Hypertext CG Yahoo! Groups Sponsor ~--> Most low income households are not online. Help bridge the digital divide today! http://us.click.yahoo.com/cd_AJB/QnQLAA/TtwFAA/1U_rlB/TM ~-> - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
Re: [svg-developers] SVG Lint? Or a conformance test for content?
Chris Lilley wrote: >On Thursday, September 22, 2005, 11:30:40 PM, Mike wrote: > > ... >MCF> My hypothesis is that the AdobeSVG viewer is picking up some unexpected >MCF> feature/property and isn't able to deal with it, while the >MCF> Mozilla/KDE/Inkscape engines are slightly more robust in dealing with >MCF> that particular problem. To test that I'm wondering if there is a tool >MCF> that is extremely anal about conformance that can be used to test >MCF> content. Basically just something that will parse an SVG file and >MCF> complain about anything that could cause a problem on a browser. > >MCF> For reference, a failing file is here: > >MCF> http://www.vrplumber.com/programming/adoberenderswhiteonwhite.svg > >You also have some attributes like object_name and object_counts on >path. Those attributes should be in their own namespace if they are put >on svg elements. > >cx="0.0" cy="0.0" on path?? > > Okay, I have now defined a namespace for us (though not yet a DTD for it), and put the application specific attributes/annotations into that namespace. The cin:cx/cin:cy for the paths represents the center of the icon for scaling/positioning purposes, they are used solely by the script(s). I've also switched the script language to ecmascript as suggested previously. Thanks for pointing out all of those problems. Obviously I've been playing too loose with the XML rules. I've also prefixed the title attributes with xlink: and eliminated the duplicate IDs that the icons were declaring. With that, the W3C validator here: http://validator.w3.org/ is reporting only 20 errors, all of which are "this attribute doesn't exist" ones which appear to be something silly wrt the namespace declarations (e.g. sodipodi attributes added by inkscape are among those getting complaints), possibly just sodipodi not having a DTD published either?. However, I still seem to be breaking something, as Adobe SVG 6beta is still displaying the content as white-on-white. Newest version of the failing file is here in case anyone wants to look at it: http://www.vrplumber.com/programming/adoberenderswhiteonwhite3.svg With thanks, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com Yahoo! Groups Sponsor ~--> Fair play? Video games influencing politics. Click and talk back! http://us.click.yahoo.com/T8sf5C/tzNLAA/TtwFAA/1U_rlB/TM ~-> - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
Re: [svg-developers] SVG Lint? Or a conformance test for content?
On Friday, September 23, 2005, 1:26:57 AM, Mike wrote: MCF> Chris Lilley wrote: >>On Thursday, September 22, 2005, 11:30:40 PM, Mike wrote: >> >> MCF> ... >>MCF> My hypothesis is that the AdobeSVG viewer is picking up some unexpected >>MCF> feature/property and isn't able to deal with it, while the >>MCF> Mozilla/KDE/Inkscape engines are slightly more robust in dealing with >>MCF> that particular problem. To test that I'm wondering if there is a tool >>MCF> that is extremely anal about conformance that can be used to test >>MCF> content. Basically just something that will parse an SVG file and >>MCF> complain about anything that could cause a problem on a browser. >> >>MCF> For reference, a failing file is here: >> >>MCF> http://www.vrplumber.com/programming/adoberenderswhiteonwhite.svg >> >>You also have some attributes like object_name and object_counts on >>path. Those attributes should be in their own namespace if they are put >>on svg elements. >> >>cx="0.0" cy="0.0" on path?? >> >> MCF> Okay, I have now defined a namespace for us (though not yet a DTD for MCF> it), A DDTD wouldn't help - combining DTDs for different name spaces is non trivial hackery. MCF> and put the application specific attributes/annotations into that MCF> namespace. The cin:cx/cin:cy for the paths represents the center of the MCF> icon for scaling/positioning purposes, they are used solely by the MCF> script(s). I've also switched the script language to ecmascript as MCF> suggested previously. MCF> Thanks for pointing out all of those problems. Obviously I've been MCF> playing too loose with the XML rules. MCF> I've also prefixed the title attributes with xlink: and eliminated the MCF> duplicate IDs that the icons were declaring. With that, the W3C MCF> validator here: MCF> http://validator.w3.org/ MCF> is reporting only 20 errors, all of which are "this attribute doesn't MCF> exist" ones which appear to be something silly wrt the namespace MCF> declarations (e.g. sodipodi attributes added by inkscape are among those MCF> getting complaints), possibly just sodipodi not having a DTD published MCF> either?. Its that the DTD does not know about namespaces at all. Technically namespace declarations are not even attributes. Its not incorrect to have namespace declarations wherever they are needed. MCF> However, I still seem to be breaking something, as Adobe SVG 6beta is MCF> still displaying the content as white-on-white. Newest version of the MCF> failing file is here in case anyone wants to look at it: MCF> http://www.vrplumber.com/programming/adoberenderswhiteonwhite3.svg thanks - still not clear why it does that. I don't actually have ASV instaled right now - Firefox is at v1.5b1 using its internal renderer, Opera is at 8.5 with its internal renderer and I have IE using the Corel 2.1 plugin. All of those, plus eSVG, display the graphic in green and black. -- Chris Lilleymailto:[EMAIL PROTECTED] Chair, W3C SVG Working Group W3C Graphics Activity Lead Co-Chair, W3C Hypertext CG Yahoo! Groups Sponsor ~--> Most low income households are not online. Help bridge the digital divide today! http://us.click.yahoo.com/cd_AJB/QnQLAA/TtwFAA/1U_rlB/TM ~-> - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/
Re: [svg-developers] SVG Lint? Or a conformance test for content?
Chris Lilley wrote: >On Friday, September 23, 2005, 1:26:57 AM, Mike wrote: > > >MCF> Okay, I have now defined a namespace for us (though not yet a DTD for >MCF> it), > >A DDTD wouldn't help - combining DTDs for different name spaces is non >trivial hackery. > > I suppose, seems like overkill just to have annotations for a node that can be read by the javascript. Sigh. >Its that the DTD does not know about namespaces at all. Technically >namespace declarations are not even attributes. Its not incorrect to >have namespace declarations wherever they are needed. > > Fair enough. >MCF> However, I still seem to be breaking something, as Adobe SVG 6beta is >MCF> still displaying the content as white-on-white. Newest version of the >MCF> failing file is here in case anyone wants to look at it: > >MCF> http://www.vrplumber.com/programming/adoberenderswhiteonwhite3.svg > >thanks - still not clear why it does that. > > I believe I've figured it out. It's an unnecessary "feature" in the AVG renderer AFAICS if there is a reference to an undeclared linear gradient then the whole graphic turns white-on-white. was the culprit. The gradient got used in an icon without being added to the overall pool of gradients in the document. >I don't actually have ASV instaled right now - Firefox is at v1.5b1 >using its internal renderer, Opera is at 8.5 with its internal renderer >and I have IE using the Corel 2.1 plugin. All of those, plus eSVG, >display the graphic in green and black. > > They apparently just ignored the missing gradient reference, which IMO is probably the right thing to do (though a notification somewhere of the failure would be nice). Thanks for all your help, Mike -- Mike C. Fletcher Designer, VR Plumber, Coder http://www.vrplumber.com http://blog.vrplumber.com Yahoo! Groups Sponsor ~--> Fair play? Video games influencing politics. Click and talk back! http://us.click.yahoo.com/T8sf5C/tzNLAA/TtwFAA/1U_rlB/TM ~-> - To unsubscribe send a message to: [EMAIL PROTECTED] -or- visit http://groups.yahoo.com/group/svg-developers and click "edit my membership" Yahoo! Groups Links <*> To visit your group on the web, go to: http://groups.yahoo.com/group/svg-developers/ <*> To unsubscribe from this group, send an email to: [EMAIL PROTECTED] <*> Your use of Yahoo! Groups is subject to: http://docs.yahoo.com/info/terms/