W3C home > Mailing lists > Public > www-style@w3.org > March 2016

Re: [css-text] Shaping for break-all/word-break

From: Koji Ishii <kojiishi@gmail.com>
Date: Fri, 11 Mar 2016 17:26:04 +0900
Message-ID: <CAN9ydbXSYBL10EYQvwk0shQKrH9LWm6c87qQUMG7vFeSSga2jg@mail.gmail.com>
To: Amelia Bellamy-Royds <amelia.bellamy.royds@gmail.com>
Cc: fantasai <fantasai.lists@inkedblade.net>, Behdad Esfahbod <behdad@behdad.org>, "www-style@w3.org" <www-style@w3.org>
On Fri, Mar 11, 2016 at 3:36 AM, Amelia Bellamy-Royds <
amelia.bellamy.royds@gmail.com> wrote:

> From Koji:
>
>> But is the requirements for text-on-path the same as break-all? If you
>> place them on a path, they can keep the shape, but each shape are not
>> really connected unless the shape is a line, no? Does Gecko do something
>> special to connect them?
>>
>> Text-on-path is more complicated than break-all, yes.  In some cases, a
> path may have a complete break in it, so it would be the same. In other
> cases, it just has a bend. I should have clarified: SVG does allow
> ligatures in text-on-a-path (unlike manually positioned characters), and
> the entire glyph is drawn as a single block, even if it no longer nicely
> fits the curve of the path.  So we will still need special wording for that.
>
> It's true, the resulting string of glyphs may no longer connect up nicely
> like in straight-line text. There is an SVG property to stretch the glyphs
> into non-rectangles so they connect, but it is not implemented in the
> browsers (not even Gecko).  Gecko also turns off required ligatures when
> the letters are stroked.  When I said Firefox was the best for Arabic text
> on a path, it is only in comparison to the others, which use all-isolated
> Arabic letterforms, order the letters backwards, use incorrect alignment so
> the text gets pushed off the path, and/or otherwise fail completely!
>
> (You can play with some of my test cases here:
> http://oreillymedia.github.io/SVG_Text_Layout/ch09-textPath-files/
> I need to create more comprehensive tests of both script font and RTL test
> cases for the SVG 2 test suite.)
>

Thank you for the explanations and tests, yeah, Gecko looks nice. Blink has
a bug for this, just not happened yet.

So from your 1st paragraph, IIUC, it looks like the requirements for the
break are the same.

fantasai: what do you think? Making the text linkable is easy, but should
we have more detailed text to determine the breakable point? If yes, should
this be normative or non-normative?

/koji
Received on Friday, 11 March 2016 08:26:53 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:09:01 UTC