User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.3b) Gecko/20030210 Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.3b) Gecko/20030210 I have a file which has Chinese characters in its file name, though when I try to open the file, either from the open dialogue or by a drag and drop I get told that the file e:\?????\ can not be opened. Reproducible: Always Steps to Reproduce: 1. Take a file that has Chinese characters and that is displayed in Explorer with Chinese characters (requires the Chinese language pack) 2. Drag it to a Mozilla Window 3. Observe the error dialogue Actual Results: Error message Expected Results: File should have been opened - If I open the file by a drag-and-drop from Explorer then the file gets displayed in the error dialog as e:\?????.lrc - If I open the file via the open dialog then the file gets displayed in the error dialog as e:\_____.lrc - If I try viewing the directory containing the files with chinese file names then I only see the files that don't have chinese characters in them. - I can drag such a file to the attachments pane of the Mail Composition window though it will display the Chinese characters as question marks and then when I try sending the e-mail it will just act as if its doing something useful
To intl; I bet this is more "unicode" breakage.
This case happens with when you running a non-native Chinese OS. What about if you change your system locale to Chinese?
I don't have access to a chinese based WindowsOS. I am also having this same problem with Mozilla 1.4rc1 on Windows 2000 and trying to save a page in cyrillic. As an example take the page http://www.utro.ru/ and then try saving it as 'web-page, complete'. Where I should be seeing cyrillic characters I see underscores, so (example pasted with view set to ISO-8859-1, so I am not sure that it will come through right - bugzilla needs to be updated to support UTF-8): YTPO.ru - ежедневная e-газета appears as YTPO.ru - __________ e-_______ Not the same thing. This really needs to be resolved.
Created attachment 130167 [details] Arabic file that can't be handled with mozilla You might have to use a unicode clean browser to open this file. This file was saved on a Win98 Arabic system and then move on diskett to a WinXP from which I am submitting this. So I hope everything is alrigth
Confirmed on WinXP with mozila 1.5a, but with arabic filenames. This is a unicode issue, you need to add support to open unicode files. Try typing: file:///C:/%D8%A9%D8%B1.htm this displays a dialog saying that it coldn't find (two arabic characters).htm There are lots of programs that don't support this (like WinZIP, but IE andWord does it flawlessly) I supplied a file that bugzilla probably fscked the filename.
The code to fix this was checked-in a while back (#ifdef is not enabled though) when we convert the mozilla as an Unicode app (104934) NSPR needs to turn the #ifdef on to fix this and other unicode File I/O related bugs. (88103, 85836, 84384, 88292, 108000 and more) cc Wan-Teh for the status
See also bug 162361. With the patch(attachment 125603 [details] [diff] [review]) there(my Win32 build is always made with that patch :-)), I can mix Cyrillic, Devanagari, Chinese, Greek and all sort of characters in a single file name. See also a thread of articles beginning with news://firstname.lastname@example.org and news://email@example.com
> I don't have access to a chinese based WindowsOS. If you have Win2k/XP, then you have a Chinese(Korean, French, Arabic, Hebrew, Hindi, Thai, German, Japanese, Tamil, Russian ....) based Windows :-). Being a fully Unicodized OS, Win2k/XP can be made to run under any locale supported by them. Go to the control panel | language/country option. Set your default system locale to Chinese and reboot. This is by NO means to say that this bug is not to be fixed. This and a bunch of other bugs (to make Mozilla take the full advantage of Win2k/XP's Unicode support while still making it on non- Unicode based old Win9x/ME) have to be fixed soon because people like you and me want to use multiple scripts at a time. wtc, have you decided what to do with NSPR Unicode APIs (UTF-8 vs UTF-16)?
Just like IE in which some DLL's are now different for Win9x, ME, NT and 2K/XP, what about Mozilla making two builds for 9x/ME and NT/2K/XP? I know that might be annoying with respect to compiling, but that should build more efficient codes.
Dupe of bug 108000?
Sounds like a dupe. BTW I am adding this bug as depending on bug 239279. Once that bug has been dealt with this one should automagically be corrected.
(In reply to comment #10) > Dupe of bug 108000? I wouldn't mark this bug as a dupe because bug 108000 is a *meta bug*, ie a reminder of all relevant bugs. It talks just about Unicode filename support, but doesn't specify what. It could be eg opening an Html file in browser, opening the same file in composer, saving a page using unicode filename, attaching a unicode filename in mail, saving it back, etc etc. However, IMO, we could make this bug depend on that. OTOH, This bug depends on bug 162361 which depends on bug 239279, so I'm not sure if it's necessary to mark this bug as dependent on bug 239279.
> If you have Win2k/XP, then you have a Chinese(Korean, French, Arabic, Hebrew, > Hindi, Thai, German, Japanese, Tamil, Russian ....) based Windows :-). Being a > fully Unicodized OS, Win2k/XP can be made to run under any locale supported by > them. Go to the control panel | language/country option. Set your default > system locale to Chinese and reboot. This does not enable the opening of Chinese language files either, because the "Chinese" system only provides an overlay translation of menus, and some message window components. It still scrambles system messages and other elements, and is not a 'true Chinese' XP. Running many Chinese-only based softwares (such as early Dr. Eye versions and Ziran Shurufa 2003) results in scrambled text in window boxes &c. I use multiple language file names quite frequently, and many are HTML downloads from the web, and these remain unopenable in this latest version of Firefox. This bug seems to have been open for a year and a half - is it very hard to fix?
Michael, you're talking about MUI (Multi-language User Interface) for English based Win2k/XP only. It's different from changing system locale (eg in XP: Control Panel > Regional Setting > Advanced option tab > "Non Unicode program")
(In reply to comment #13) > This does not enable the opening of Chinese language files either, It should unless you try to open a file with SC name with the default system locale set to TC or vice versa. > is not a 'true Chinese' XP. There's nothing special about Chinese XP except for menus, help and etc translated into Chinese. > Running many Chinese-only based softwares (such as > early Dr. Eye versions and Ziran Shurufa 2003) results in scrambled text in > window boxes &c. If they're Unicode-based programs, they should work flawlessly regarldess of the default system locale on *any* language versions of Windows XP. If they're non-Unicode-based programs, they should run fine after switching the default system locale to TC or SC (depending on which non-Unicode code page they use, CP936 or CP950) > I use multiple language file names quite frequently, and many are HTML downloads > from the web, and these remain unopenable in this latest version of Firefox. See bug 162361 and bug 239279. We're working on it, but none of us has enough free time to fix it very soon (we know what to do, but don't have time)
*** Bug 320785 has been marked as a duplicate of this bug. ***
Now fixed on trunk thanks to patches for bug 162361 and bug 278161
*** Bug 88292 has been marked as a duplicate of this bug. ***
I've just downloaded Firefox 2 and tried to verify if this bug is resolved. Doing a drag 'n drop does indeed work. Bravo!! :D However, if I double-clicked on the same file, it doesn't work. "open files" as written in the summary can mean a lot of things, amongst them: 1. Drag 'n drop 2. Double-click (ie lauched by Windows Explorer) 3. Through File menu's "Open File ..." 4. In command line (or called by another program?) I've verified that point 1 and point 3 work, but point 2 failed (I can't check point 4 since it's impossible for me to type Unicode in console), so do anyone mind if I reopen this bug? Or do you prefer another bug report?
I can confirm point 2 in comment 19 is still an issue. My test: - create a filename with non-latin characters (used character map) for example '萔菾菰萩쫦쪹.html' - double-click that file The result: Firefox can't find the file at /C:/tmp/??????.html. This should be reopened, since it is still broken for this issue.
(In reply to comment #19) > 2. Double-click (ie lauched by Windows Explorer) > 3. Through File menu's "Open File ..." > 4. In command line (or called by another program?) Please, see bug 162361 comment #180 and a few more comments below it. See also bug 282285.
filed bug 359200 to track issue #2 explicitly.