Re: Agenda, 13 January 2011 SVG WG telcon

On Thu, 13 Jan 2011 20:51:32 +0100, Cameron McCormack <cam@mcc.id.au>  
wrote:

> Erik Dahlstrom:
>> If we're covering specific tests, here are some that I'd like to
>> change/discuss:
>>
>> - I'd like to add direction="rtl" to all tests that are clearly
>> dealing with RTL text, eg:
>>   * fonts-glyph-02-t.svg
>>   * text-intro-05-t.svg
>
> Maybe some of the text-text- ones too?

I went through and checked, and there are no RTL tests in text-text-*.

> smontagu (who recently
> implemented some level of bidi support for SVG text) mentioned something
> like that to me, and that that was the reason the text is going off in
> the wrong direction in Firefox, I think.  I’ll have to take another look
> at them in light of Chris’ wording backport.

Yes, having recently revisited the bidi code in Opera myself the  
conclusion I came to there was to always base the text-anchor alignment on  
the computed value for 'direction' (the initial value is 'ltr'). Another  
curiosity is that Tiny 1.2 doesn't apply 'text-anchor' on anything but  
<text> elements, unlike SVG 1.1 where 'text-anchor' applies to all 'text  
content elements'.

Anyhow, none of the tests are explicitly testing RTL text alignment and  
the direction property so I'd say it's better to add the direction anyway  
to make it explicit. Having heard no objections to this so far I'm going  
to make these minor edits.

>> - Would like to add a CSS :target rule to style the highlight in
>> linking-uri-01-b.svg.
>
> :target sounds good, but I think we should have some spec text to state
> that the SVG fragments that are meant to do highlighting should make the
> element match :target.

You're right, :target is not one of the required pseudoclasses in SVG 1.1  
according to the bulletlist at the bottom of  
http://www.w3.org/TR/SVG11/styling.html#StylingWithCSS. I've raised an  
issue on this for now.

...
>> - Passcriteria for linking-uri-02-b.svg are incorrect, the test
>> doesn't remain unchanged in the frame, it loads the -01 test in the
>> same frame.
>
> Yeah.

Ok, will update the passcriteria there.

>> - text-align-07-t.svg: Are missing-glyphs ok here? If so what's the
>> expected alignment?
>
> That is a good question.
>
>> - text-text-05-t.svg: The reference image seems to indicate that the
>> tightest fitting bbox around the text should be used for text-anchor
>> adjustments and not the sum of the advances (note that the biggest
>> square doesn't fully fill it's em-box). Can't find that requirement
>> in the spec.
>
> If there’s no requirement in the spec, then sum of the advances sounds
> more reasonable to me.

So I went back and checked again, and found for text-anchor="end" that  
"The rendered characters are aligned such that the end of the text string  
is at the initial current text position.". So, it depends on what "the end  
of the text string" is considered to be. You could probably also design a  
font such that text-anchor="start" would show the same "misalignment".

Cheers
/Erik

-- 
Erik Dahlstrom, Core Technology Developer, Opera Software
Co-Chair, W3C SVG Working Group
Personal blog: http://my.opera.com/macdev_ed

Received on Tuesday, 18 January 2011 14:21:07 UTC