73103 - Japanese line breaking and word wrapping rules are violated by some characters during mail composition.

Status: VERIFIED WORKSFORME

(MailNews Core :: Internationalization, defect)

Description

[Wesley George]

<summary>
Japanese line breaking and word wrapping rules are violated by some characters 
during mail composition.

<repro environment>
build 03-21-04 + Win2000-JA

Not repro on:
build 03-21-04 + Mac OS 9.04-JA

<scenario>
The following Unicode points violate line breaking and word wrapping rules:
0024, 0025, 002E, 00A2, 00A3, 00A5, 00B0, 2014, 2019, 2024, 2025, 2026, 2116, 
2030, 2031, 2032, 2033

<expected>
Line breaking and word wrapping rules to comply with the standard in the JIS X 
4051 document.

Comment 1

[nhottanscp]

Is that working on other places, HTML composer, browser?

Comment 2

[Wesley George]

Andreas, could you please ask the browser team to investigate and comment on
results for browser and composer with character coding ISO-2022-JP.  I see some
browser line breaking test cases for Shift-JIS and EUC-JP on http://babel, but
not ISO-2022-JP.  Thanks.

Comment 3

[nhottanscp]

Reassign to shanjian.

Comment 4

[Andreas Becker]

Adding [e-mail].  Yuying, do you have some information about these
test cases?

Comment 5

[Shanjian Li]

Please provide test case or detail description about what is broken. 

Comment 6

[Yuying Long]

I guess it's maybe not only a Mail?Windows issue, you can see how those problem
marks work in Browser here: 
http://babel/tests/Browser/kinsoku/linebreak2022jp.html
I added some lines at the end of document by using Composer(I added them on
Mac), those lines were not wrapped/line break well beside "?".

Comment 7

[Shanjian Li]

Except for Japanese, there is no formal standard for line/word break for other languages 
as I know of. Considering the fact that we are handling multi-lingual text, we did not 
follow 4051 strictly. If a CJK character is found before space when looking for a line 
break, 4051 will be followed. Otherwise, space character is the only place where line 
break is permitted. This is for the purpose to make ":-)" work. Some zenkaku punctuation
symbols are not recognizable after they have been converted to unicode, and we have this 
problem. If character before the punctuation is CJK character, you will see the correct 
behavior. The current work/line break behavior can be improved, and it will never be 
perfect. Mark mile stone to be future. 

If a typical situation is found in which current behavior is annoying, please provide 
more information and I will rethink this bug and make some improvement if possible. 

Comment 8

[ji]

Checked 06/26 branch build. With Japanese contents, the mail follows the
wrapping rules. The ascii strings in Japanese (iso-2022-jp) mail doesn't follow
the rule. It's the expected behaviour. Resolved it as WFM.

Comment 9

[ji]

Verified as such.