593601 - A gradient applied on a textPath should cover the bounding box

Status: UNCONFIRMED

(Core :: SVG, defect)

Description

[Sebastian Markbåge]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US) AppleWebKit/533.4 (KHTML, like Gecko) Chrome/5.0.375.127 Safari/533.4
Build Identifier: 

This was first described in bug 577850.

Reproducible: Always

Steps to Reproduce:
1. Create a textPath element within a text element.
2. Let the textPath refer to a path which doesn't start at origin 0,0.
3. Set the fill to a linear gradient.
Actual Results:  
The gradient looks like it covers a larger area than the rendered bounding box.

Expected Results:  
The gradient edges should touch the bounding box.

Comment 1

[Sebastian Markbåge]

Attachment: Test Case

The expected result is for the gradient to go from black to white to black within the rendered text area.

Comment 2

[Robert Longson [:longsonr]]

Attachment: with visible bounding box

Opera has the same bounding box as Firefox. Curious eh?

Batik has it tightly round the text. What does Webkit do?

Comment 3

[Robert Longson [:longsonr]]

Attachment: without spaces

So it's all about how you compress spaces then.

Comment 4

[Robert Longson [:longsonr]]

http://www.w3.org/TR/SVG/text.html#WhiteSpace

Comment 5

[Robert Longson [:longsonr]]

Attachment: in particular the whitespace between foo and bar

Comment 6

[Robert Longson [:longsonr]]

So it's all about where that middle space character is put that's not part of the textPaths. (The leading and trailing whitespace is removed anyway so it doesn't affect things).

Comment 7

[Robert Longson [:longsonr]]

And the position of the central space in Firefox is (128, 0). This seems right as it's the underlying position of the end of the text before the textPath transforms it. So I think this bug is invalid.

Comment 8

[Robert Longson [:longsonr]]

Attachment: M

Batik draws the M near the foo. Opera, the Adobe SVG Plug in for IE and Firefox draw it just off the top of the window.

Comment 9

[Boris Zbarsky [:bzbarsky]]

Actually, Opera changed its behavior here from 10.5 to 10.6.  It seems to now put the M near the foo (in 10.6), which is just wrong.  At least on Mac.

So am I missing something, or is this bug just invalid?

Comment 10

[Robert Longson [:longsonr]]

I'm still not totally convinced it's clearly specified in the SVG spec what to do. I plan to ask them.

Comment 11

[Sebastian Markbåge]

After some consideration, I think this bug is invalid according to the SVG spec.

Since the white space will be turned into a single space glyph, it cannot be removed following the existing white space guidelines.

Additionally, the bounding box calculation must consider the full size of the glyph. "The calculations assume that all glyphs occupy the full glyph cell."

Although it's counter intuitive, it's not less so than equivalent text flow in HTML.

Comment 12

[Robert Longson [:longsonr]]

http://lists.w3.org/Archives/Public/www-svg/2010Sep/0080.html

Comment 13

[Robert Longson [:longsonr]]

The whitespace can't be removed that's true. The outstanding question is where should it go?

Comment 14

[Robert Longson [:longsonr]]

Cam these tests no longer display much of anyhing at all any more, they used to (only the gradient was an issue). This seems to be a recent issue.

Comment 15

[Robert Longson [:longsonr]]

Could be bug 941585 I suppose.

Comment 16

[Robert Longson [:longsonr]]

Seems it is that bug.