238717 - Mac OS X: attachment file's Japanese name is not converted to composed Unicode

Status: RESOLVED DUPLICATE of bug 227547

(MailNews Core :: Internationalization, defect)

Description

[Katsuhiro MIHARA]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; ja-JP; rv:1.5) Gecko/20031007
Build Identifier: Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.7b) Gecko/20040316

In Mozilla 1.7 Alpha Release Notes,

> On Mac OS X, attachment file names are no longer displayed
> in decomposed Unicode but are converted to composed Unicode.

But Japanese file name is displayed and send in decomposed Unicode.

Reproducible: Always
Steps to Reproduce:
1.make a file "バージョンアップ.txt" ("versionup.txt" in Japanese Katakana)
2.compose a mail
3.attach that file
4.send the mail

Actual Results:  
Japanese file name is not converted to composed Unicode.
Mailers in Windows can't read attachment file names.


Expected Results:  
Japanese file name should be converted to composed Unicode.

Comment 1

[Katsuhiro MIHARA]

Attachment: message.eml (Japanese ISO-2022-JP)

attach a file "バージョンアップ.txt" ("versionup.txt" in Japanese Katakana)
and send a mail to Windows client.

Comment 2

[louis bennett]

see also: bug 172488, bug 147075, bug 191502

Comment 3

[Koike Kazuhiko]

Confirming bug.

Original report in Buzilla-jp:
http://bugzilla.mozilla.gr.jp/show_bug.cgi?id=3692

Comment 4

[WADA:World Anti-bad-Duping Agency]

Japanese filename in mail of Comment #0 becomes as follows (Decomposed Unicode,
UTF-8 encoded);
> filename="=?UTF-8?B?44OP44KZ44O844K344KZ44On44Oz44Ki44OD44OV44KaLnR4dA==?="
> (Hexa value of decoded UTF-8) 
> E3838F E38299 E383BC E382B7 E38299 E383A7 E383B3 E382A2 E38383 E38395 E3829A
2E 74 78 74

This is usually(by other than Mozilla/Thunderbird on Mac OS X 10.2 or after)
encoded as follows ;
> filename="=?UTF-8?B?44OQ44O844K444On44Oz44Ki44OD44OXLnR4dAo==?="
> (Hexa value of decoded UTF-8) 
> E38390 E383BC E382B8 E383A7 E383B3 E382A2 E38383 E38397 2E 74 78 74

E38299(U3099 COMBINING KATAKANA-HIRAGANA VOICED SOUND MARK) and E3829A(U309A
COMBINING KATAKANA-HIRAGANA SEMI-VOICED SOUND MARK) are not used as usual
Japanese character, and fonts usually does not have grifs for these code points.

If these symbols are used as an independent/stand-alone character like ones in
Half Width Katakana, E3829B(U309B KATAKANA-HIRAGANA VOICED SOUND MARK) and
E3829C(309C KATAKANA-HIRAGANA SEMIVOICED SOUND MARK) are used.

File name/Folder name in decomposed format should be composed if it is passed to
other system.

See "2.Combining-like characters" of "Annex 7: Programming Language Identifiers"
in Unicode Normalization Forms
 (http://www.unicode.org/reports/tr15/#Programming_Language_Identifiers)
for these special characters in Japanese.

See also http://developer.apple.com/qa/qa2001/qa1173.html and
http://developer.apple.com/qa/qa2001/qa1235.html for normalization related
descriptions on Mac OS X (10.2 or after).

Comment 5

[Jungshik Shin]

This is a dupe of bug 227547 which I thought I had fixed. 


*** This bug has been marked as a duplicate of 227547 ***