W3C home > Mailing lists > Public > www-svg@w3.org > November 2011

Text metrics with display:none

From: Brian Birtles <bbirtles@mozilla.com>
Date: Mon, 28 Nov 2011 10:59:39 +0900
Message-ID: <4ED2EB0B.3050701@mozilla.com>
To: 'www-svg' <www-svg@w3.org>
Dear all,

I'd like to clarify the expected behaviour for the SVGTextContentElement
interface[1] in conjunction with display:none.

Given the following markup:

  <text>a<tspan display="none">b</tspan>c</text>

What is the expected result of the following methods called on the
<text> element?


Looking at the spec, I would expected getNumberOfChars to return 3
(since "Effectively, this is equivalent to the length of the
Node::textContent attribute from DOM Level 3 Core"). However, the
results I get from a variety of browsers are as follows:[2]

 getNumberOfChars: 2 (Firefox 8, Chrome 15, IE9); 3 (Opera 11.60 beta)

Other metrics:

 Firefox 8 behaves as if there are just two characters.
 Chrome 15 behaves as if there are just two characters (but surprisingly
gives the same result for getSubStringLength(0, 1),
getSubStringLength(0, 2) and getSubStringLength(0, 3) which is half of
getComputedTextLength. It also does not throw INDEX_SIZE_ERR exceptions
but just returns 0 values of out of range character indices.)
 Opera 11.60 beta behaves as if there are three characters in most cases
(getSubStringLength(1,1) returns 0 but getExtentOfChar(1) returns a
non-zero glyph rect; substrings beginning at 0 equal
getComputedTextLength regardless of the length of the substring)
 IE 9 behaves as if there are just two characters (but fails to throw an
exception for getSubStringLength(0, 3)).

It seems that most of the major browsers are treating the display:none
characters as if they don't exist for the purpose of these metrics. Most
likely such characters are not present in the rendering tree and
calculating font metrics is tied to the rendering tree.

My questions:

1) Have I understood the spec correctly--that getNumberOfChars should
return 3?
2) If so, should we change the spec or the browsers?



[1] http://www.w3.org/TR/SVG11/text.html#InterfaceSVGTextContentElement
[2] Test case here: http://brian.sol1.net/svg/interop/text-test.svg
Received on Monday, 28 November 2011 02:00:12 GMT

This archive was generated by hypermail 2.3.1 : Friday, 8 March 2013 15:54:49 GMT