Closed Bug 189480 Opened 22 years ago Closed 21 years ago

OS/2 only - some big5 characters show as "?"

Categories

(Core :: Internationalization, defect)

x86
OS/2
defect
Not set
normal

Tracking

()

RESOLVED FIXED

People

(Reporter: ccwu, Assigned: jhpedemonte)

References

()

Details

(Keywords: fixed1.7)

Attachments

(2 files, 4 obsolete files)

User-Agent: Mozilla/5.0 (OS/2; U; Warp 4.5; en-US; rv:1.3a) Gecko/20021210 Build Identifier: Mozilla/5.0 (OS/2; U; Warp 4.5; en-US; rv:1.3a) Gecko/20021210 some big5 characters show as "?" in os/2, we use cp950 as big5. this is the same in windows. there are many similar bug reports before, and i think it should be fixed. but it's not in os/2. it's strange that i can see these characters in windows 2000. Reproducible: Always Steps to Reproduce: 1. 2. 3. Expected Results: is it possible to use cp950 as big5 and big5hkscs like windows or add one character coding "cp950"? btw, i prefer the first one solution.
*** Bug 189488 has been marked as a duplicate of this bug. ***
I've been randomly seeing ? in place of ' for many weeks in OS/2 trunk builds using default En US charset.
OS: other → OS/2
Attached file Reduced testcase
The character in question is F9D6 which is in the user defined range of codepage 950. When Mozilla reads a page, it converts the contents to Unicode and then it is up to the OS to convert it back to the native codepage or display it in Unicode. In this case, Mozilla converts it to a Unicode char and when we ask OS/2 to convert it back, it doesn't know what to do with it because there is no standard for converting from Unicode to the user defined range. We have some plans to change our code to use Mozilla converters instead of OS/2 converters and that might help this problem.
Assignee: smontagu → mkaply
Status: UNCONFIRMED → NEW
Ever confirmed: true
Looks like OS/2 problem - no reproducible on Windows, Mac OS X and linux.
More information. At issue here is the ETen extension. Here is some info from http://seba.ulyssis.org/thesis/big5_2.php > A very important extension is the ETen extension. It contains two blocks, > the first block (C6A1 - C3D3) has 365 characters (numbers, kana, ...) the > second block (F9D6 - F9FE) only has 7 hanzi (碁 銹 裏 墻 恒 粧 嫺) and 34 line > drawing characters. So Mozilla maps F9D6 to the ETen extension characters in Unicode, but OS/2 does not. As mentioned in the previous post, when we convert our fonts to use Mozilla codepage conversions, this will be fixed.
Summary: some big5 characters show as "?" → OS/2 only - some big5 characters show as "?"
i think this is problem of os/2 not mozilla after reading your coments, and use Mozilla converters instead of OS/2 converters would fix this. thank you!! i've looked more information. in perl and linux locale they use big5 as alias of big5-eten now. http://www.mathematik.uni-ulm.de/help/perl5/doc/Encode/TW.html ps.big5 < cp950 < big5-eten < big5-hkscs ibm didn't implement big5-hkscs in os/2 :(
*** Bug 237506 has been marked as a duplicate of this bug. ***
Attached patch patch (obsolete) — Splinter Review
This patch changes WideCharToMultiByte and MultiByteToWideChar to use the nsIUnicodeEncoder and nsIUnicodeDecoder, instead of OS/2's unicode converters. With this patch, the attached testcase displays somewhat better, but the center character now displays as a square. Don't know if there is anything we can do about that, though. Also, this patch makes the testcase in bug 237506 work.
Assignee: mkaply → pedemont
Status: NEW → ASSIGNED
Attachment #144168 - Flags: review?(mkaply)
Blocks: 238039
Attached patch Change platform charset code (obsolete) — Splinter Review
Use nsPlatformCharset to get platform charset. remerge clipboard code.
Attachment #144168 - Attachment is obsolete: true
Comment on attachment 144168 [details] [diff] [review] patch r=mkaply for the pieces that are the same
Attachment #144168 - Flags: review?(mkaply) → review+
Attached patch The real slim shady (obsolete) — Splinter Review
oops
Attachment #145104 - Attachment is obsolete: true
Attached patch more fixes (obsolete) — Splinter Review
(1) Fix memory issue in GetUconvObject. (2) Set error behavior for nsIUnicodeEncoder in nsOSUni.cpp.
Attachment #145119 - Attachment is obsolete: true
Attached patch more fixes IISplinter Review
Posting correct patch.
Attachment #146026 - Attachment is obsolete: true
Attachment #146027 - Flags: review?(mkaply)
Depends on: 240596, 240597
Comment on attachment 146027 [details] [diff] [review] more fixes II r=mkaply sr=blizzard (platform specfic) once the dependent bugs are in, we need to update os2charset.properties with all OS/2 charsets.
Attachment #146027 - Flags: superreview+
Attachment #146027 - Flags: review?(mkaply)
Attachment #146027 - Flags: review+
Comment on attachment 146027 [details] [diff] [review] more fixes II a=mkaply for 1.7 - OS/2 only
Attachment #146027 - Flags: approval1.7+
Fix checked in.
Status: ASSIGNED → RESOLVED
Closed: 21 years ago
Keywords: fixed1.7
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: