[Webkit-unassigned] [Bug 16768] Position and thickness of underline as text size changes

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=16768





--- Comment #30 from Yusuke Sato <yusukes at chromium.org>  2009-10-06 06:15:27 PDT ---
Sorry for the very late reply, and thanks for the review.

> why we're looking for the largest position.  I guess thats' the lowest
> underline?  Because some characters might need a lower underline than others? 
> Seems strange that this could mean that if 'g' has a lower underline than all
> other characters that we'd end up with discontigous underlines for something
> like:

It searches for the lowest underline position because a single TextRun might
contain characters for two or more languages (e.g. "aあ") and in such case, the
fonts (e.g. for "a" and for "あ") might have different underline positions.

> <u>g<s>b</s>h</u>
>
> I would expect there to be a single underline along the whole bottom of that,
> but if 'g' has a different underline position than b and h it won't.  Maybe
> that's just an error in the font?

However, in the example above, the underline can't be discontiguous since "g",
"b", and "h" are all English and therefore are rendered with a single font.
(Please check the rebaselined version of
platform/mac/fast/dom/clone-node-dynamic-style-expected.png. The test contains
similar markups. You can see the underlines in the png are contiguous.)

If the example is something like this: 
    <u>a sentence in lang A <s>a sentence in lang B</s></u>
the underline could be discontiguous (depending on the font metrics and the
font size), but I'd guess it's rare.

> If this was an svn-create-patch instead of a git patch, we could actually see
> the image files for comparison. :(  Since I can't actually make intelligent
> comments about this one, r-.  rubber stamping is the only other option
> available to me. :)

I've prepared a new patch. Thanks.

>  Did you have any luck in tracking down mitz or hyatt via IRC?

I haven't had cycles to work on this patch these days... I'm going to ask them
tomorrow.

Thanks,
Yusuke

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.