Bug 51012

Summary: Kana should be part of CJK Ideographs in WebCore/platform/graphics/Font.cpp
Product: WebKit Reporter: Koan-Sin Tan <koansin.tan>
Component: TextAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: ap, commit-queue, eric, hamaji, mitz, yuzo
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: All   
OS: All   
Bug Depends on:    
Bug Blocks: 50365, 50619    
Attachments:
Description Flags
Kata not displayed right in vertical text
none
Proposed patch
none
Proposed patch
none
kana displayed correctly after the patch is applied
none
modified patch with expected result added
none
patch with expected results none

Description Koan-Sin Tan 2010-12-13 21:57:23 PST
Created attachment 76497 [details]
Kata not displayed right in vertical text

Katakana and hiragana should be considered part of CJK Ideographs, otherwise they are not displayed correctly in vertical writing
Comment 1 Koan-Sin Tan 2010-12-13 22:14:33 PST
Created attachment 76500 [details]
Proposed patch

Add the following to CJK ideographs:
hiragana: U+3040 - U+309F
katakana: U+30A0 - U+30FF
Halfwidth and Fullwidth Forms: U+FF00 - U+FFEF

Glyphs in Halfwidth and Fullwidth Forms are widely used in CJK. They are mostly used in CJK, I think.
Comment 2 Koan-Sin Tan 2010-12-13 23:54:39 PST
Created attachment 76508 [details]
Proposed patch

Same as 76500, updated to change WebCore::Font::isCJKIdeographOrSymbol introduced in http://trac.webkit.org/changeset/74005 instead of WebCore::Font::isCJKIdeograph
Comment 3 Koan-Sin Tan 2010-12-14 01:03:15 PST
Created attachment 76510 [details]
kana displayed correctly after the patch is applied

Yes, I know it's broken font, but sometimes people use broken font :-)
Comment 4 Alexey Proskuryakov 2010-12-14 12:04:31 PST
Comment on attachment 76508 [details]
Proposed patch

View in context: https://bugs.webkit.org/attachment.cgi?id=76508&action=review

All these checks look slow. I don't know if this is a practical problem.

> LayoutTests/ChangeLog:8
> +        Test if kana displayed correctly in vertical text with broken font.

The test should have expected results, see <http://webkit.org/coding/contributing.html>.
Comment 5 Koan-Sin Tan 2010-12-14 15:11:49 PST
(In reply to comment #4)
> (From update of attachment 76508 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=76508&action=review
> 
> All these checks look slow. I don't know if this is a practical problem.

you have to deal with it anyway, right?
how about combining checks of contiguous blocks, such as hiragana and katakana?
or try to come up some magic arithmetic?

> > LayoutTests/ChangeLog:8
> > +        Test if kana displayed correctly in vertical text with broken font.
> 
> The test should have expected results, see <http://webkit.org/coding/contributing.html>.

Will do that, thanks.
Comment 6 Alexey Proskuryakov 2010-12-14 15:15:40 PST
This was just a comment in passing, please don't implement any optimizations unless profiling show that this is hot code.
Comment 7 Koan-Sin Tan 2010-12-14 16:55:47 PST
Created attachment 76594 [details]
modified patch with expected result added

Same as 76508. Added expected results.
Comment 8 Koan-Sin Tan 2010-12-14 17:02:35 PST
Created attachment 76597 [details]
patch with expected results

Same as 76594. Deleted a stupid \n introduced by copy-and-paste
Comment 9 WebKit Commit Bot 2010-12-14 19:48:18 PST
Comment on attachment 76597 [details]
patch with expected results

Clearing flags on attachment: 76597

Committed r74096: <http://trac.webkit.org/changeset/74096>
Comment 10 WebKit Commit Bot 2010-12-14 19:48:25 PST
All reviewed patches have been landed.  Closing bug.