Issue with Scalable Child test

[Jonathan Watt-3]

Hi,

I have a patch for the sizing of SVG in a CSS context (XHTML) going through
review at Mozilla right now. The test:

WICD Core Comb. #1
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdcore.xhtml#core-combined01

would seem to pass, except that the whitespace between the <object/> tags takes
up horizontal space. This extra space means there's not enough room for three
<object>s with a width of 33.33%, so it knocks the third embedded SVG over so
there are only two circles per line. If I remove the whitespace from between the
tags then I do indeed get three circles per line.

Since the children of the <div> containing the SVG are all inline I think the
behavior is correct, so perhaps the test needs to be rethought.

Regards,
Jonathan

[Timur Mehrvarz]

Hi Jonathan,

I think I got this fixed. Could you please verify?

Thanks,
Timur

http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-scalable-child-size-calculation.xhtml


On 09.11.2007, at 20:56, Jonathan Watt wrote:

>
> Hi,
>
> I have a patch for the sizing of SVG in a CSS context (XHTML) going  
> through
> review at Mozilla right now. The test:
>
> WICD Core Comb. #1
> http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdcore.xhtml#core-combined01
>
> would seem to pass, except that the whitespace between the <object/>  
> tags takes
> up horizontal space. This extra space means there's not enough room  
> for three
> <object>s with a width of 33.33%, so it knocks the third embedded  
> SVG over so
> there are only two circles per line. If I remove the whitespace from  
> between the
> tags then I do indeed get three circles per line.
>
> Since the children of the <div> containing the SVG are all inline I  
> think the
> behavior is correct, so perhaps the test needs to be rethought.
>
> Regards,
> Jonathan
>
>

[Jonathan Watt-3]

Hi Timur,

Perfect. With your fixes for the test, and with the patch, the development
version of Firefox now renders three rows of three circles as expected.

Thanks,
Jonathan

Timur Mehrvarz wrote:

> Hi Jonathan,
>
> I think I got this fixed. Could you please verify?
>
> Thanks,
> Timur
>
> http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-scalable-child-size-calculation.xhtml
>
>
> On 09.11.2007, at 20:56, Jonathan Watt wrote:
>
>> Hi,
>>
>> I have a patch for the sizing of SVG in a CSS context (XHTML) going  
>> through
>> review at Mozilla right now. The test:
>>
>> WICD Core Comb. #1
>> http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/wicdcore.xhtml#core-combined01
>>
>> would seem to pass, except that the whitespace between the <object/>  
>> tags takes
>> up horizontal space. This extra space means there's not enough room  
>> for three
>> <object>s with a width of 33.33%, so it knocks the third embedded  
>> SVG over so
>> there are only two circles per line. If I remove the whitespace from  
>> between the
>> tags then I do indeed get three circles per line.
>>
>> Since the children of the <div> containing the SVG are all inline I  
>> think the
>> behavior is correct, so perhaps the test needs to be rethought.
>>
>> Regards,
>> Jonathan
>>
>>
>
>
>

[Jonathan Watt-3]

Jonathan Watt wrote:
> Perfect. With your fixes for the test, and with the patch, the development
> version of Firefox now renders three rows of three circles as expected.
>>
>>
http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-scalable-child-size-calculation.xhtml

Actually, because the HTML <object> elements are display:inline they leave some
room for descenders underneath. This means that the bottom and top of the three
rows of circles are separated by a gap instead of meeting snuggly together. To
get rid of this gap the <object> elements need to have the additional rule:

  object { vertical-align: bottom; }

Regards,
Jonathan

[Timur Mehrvarz]

On 23.11.2007, at 19:51, Jonathan Watt wrote:

> http://www.w3.org/2004/CDF/TestSuite/WICD_CDR_WP1/test-scalable-child-size-calculation.xhtml
>
> Actually, because the HTML <object> elements are display:inline they  
> leave some
> room for descenders underneath. This means that the bottom and top  
> of the three
> rows of circles are separated by a gap instead of meeting snuggly  
> together. To
> get rid of this gap the <object> elements need to have the  
> additional rule:
>
>  object { vertical-align: bottom; }
>
> Regards,
> Jonathan

Btw, this was very helpful. No more gaps caused by our testcases :)

Thanks a lot.

Timur