Uri with cyrillis characters, Windows and unicode

[davide ficano]

I've a local file containing cyrillic characters in name (dir or filename)and I need to pass the corresponding path string as parameter to nsIProcess.run().

After a great discussion on extension dev irc I'm able to pass string under linux, but under Windows doesn't work.

Below I show the code I use.

I convert the uri into a nsiFile using the code shown below

Code: Select all

    var ioService = Components.classes["@mozilla.org/network/io-service;1"]
                            .getService(Components.interfaces.nsIIOService);
    return ioService.newURI(url, null, null)
            .QueryInterface(Components.interfaces.nsIFileURL)
            .file;



Then I convert the path from unicode and under Linux works fine

Code: Select all

runArgument = fromUnicode(file.path, "UTF-8");

function fromUnicode(text, charset) {
    var unicodeCvt = Components.classes["@mozilla.org/intl/scriptableunicodeconverter"]
             .createInstance(Components.interfaces.nsIScriptableUnicodeConverter);
    unicodeCvt.charset = charset;
        return converter.ConvertFromUnicode(text) + converter.Finish();
}


I've tried different charset under Windows without success

Code: Select all

runArgument = fromUnicode(file.path, "xxxxxx");

May you help me?

[vladmir]

May you help me?

And authors of JSView and Firebug needs the same help.

[vladmir]

Davide

User Infocatcher from forum.mozilla-russia.org sends you this part of code from his new extension (not finished). The code based on extension IE Tab. May be it will give you direction.

Code: Select all

var myObj = {
   openUriWithApp: function( ... ) {
      // ...
      // var path = "c:\\Program Files\\Internet Explorer\\iexplore.exe";
      // var args = [];
      args.push(this.convertStrFromUnicode(uri));
      this.startProcess(path, args);
   },
   startProcess: function(path, args) {
      args = args || [];
      var file = Components.classes["@mozilla.org/file/local;1"]
         .createInstance(Components.interfaces.nsILocalFile);
      file.initWithPath(path);
      if(!file.exists()) {
         alert(path + "\nnot found!");
         return;
      }
      var process = Components.classes["@mozilla.org/process/util;1"]
         .getService(Components.interfaces.nsIProcess);
      process.init(file);
      process.run(false, args, args.length);
   },
   convertStrFromUnicode: function(str) {
      var charset = this.charset;
      if(!charset)
         return str;
      // this.ut._log("convertStrFromUnicode -> charset -> " + charset);
      var suc = Components.classes["@mozilla.org/intl/scriptableunicodeconverter"]
         .createInstance(Components.interfaces.nsIScriptableUnicodeConverter);
      suc.charset = charset;
      try {
         str = decodeURIComponent(str); // to UTF-8
      }
      catch(e) { // does not work in fx 1.5
         str = suc.ConvertToUnicode(unescape(str)); // Thanks to IE Tab!
         str = decodeURI(str);
      }
      return suc.ConvertFromUnicode(str);
   },
   get charset() {
      var charset = "";
      if(this.ut.pref("convertURIs")) {
         charset = this.ut.pref("convertURIsTo");
         if(!charset) {
            charset = this.ut.getPref("intl.charset.default");
            if(!charset || charset.indexOf("chrome://") == 0)
               charset = this.defaultCharset;
         }
      }
      return charset;
   },
   _defaultCharset: null,
   get defaultCharset() { // thanks to IE Tab!
      if(this._defaultCharset == null) {
         var strBundle = Components.classes["@mozilla.org/intl/stringbundle;1"]
            .getService(Components.interfaces.nsIStringBundleService);
         try {
            this._defaultCharset = strBundle.createBundle("chrome://global-platform/locale/intl.properties")
               .GetStringFromName("intl.charset.default");
         }
         catch(e) {
            this._defaultCharset = "";
         }
      }
      return this._defaultCharset;
   }
};
myObj.openUriWithApp( ... );

[davide ficano]

Davide

User Infocatcher from forum.mozilla-russia.org sends you this part of code from his new extension (not finished). The code based on extension IE Tab. May be it will give you direction.

Hi,

thank you very much but Infocatcher's code doesn't work for me, at least on my italian Windows with charset 8859-1

[Infocatcher]

doesn't work for me, at least on my italian Windows with charset 8859-1

Try choose charset manually:
this.ut.pref("convertURIs") = true
this.ut.pref("convertURIsTo") = "iso-8859-1" (?)

And what returns defaultCharset getter?

And IE Tab works for you? Set path to external application and try open local file (file://) with non-ASCII characters in name (ctrl+left-click on icon in status bar).

P.S. For me works conversion to windows-1251 (default system charset), but only for opening in IE (7.0) – it not work for Opera (9.52).

[davide ficano]

doesn't work for me, at least on my italian Windows with charset 8859-1

Try choose charset manually:
this.ut.pref("convertURIs") = true
this.ut.pref("convertURIsTo") = "iso-8859-1" (?)

And what returns defaultCharset getter?

The default charset is : ISO-8859-1
I've tried others charset, I obtain (obviously) different strings but all doesn't work (i.e. the passed path isn't recognized by editors)
I've tried the charset shown below

Windows-1250
Windows-1251
Windows-1252
Windows-1253
Windows-1254
Windows-1255
ISO 8859-1 (the default)
ISO 8859-2
ISO 8859-5

and UTF-8 ...

I've used different editors

Windows Notepad
Windows Wordpad
Komodo Edit 4.4.x
Notepad++
UltraEdit

And IE Tab works for you? Set path to external application and try open local file (file://) with non-ASCII characters in name (ctrl+left-click on icon in status bar).

IETab doesn't open local path with cyrillic characters, no error and no messages into Error Console, silently the file isn't open.

P.S. For me works conversion to windows-1251 (default system charset), but only for opening in IE (7.0) – it not work for Opera (9.52).

I've tried only on IE 6.0

thanks for your feedback

[davide ficano]

Hi,

I've fixed the problem but under Windows it works only if I change the "Language for non-Unicode programs" to Russian (*), my default value is Italian.
Under Linux seems to work simply using the UTF-8 charset.
I'm unable to test under Mac OSX so actually OSX uses the same technique used by Linux.
I've done a little improvement using the current browser document characterSet property.

I don't like the solution because I can't directly open russian files inside my italian-configured Windows but I think this is a little step ahead.

Thanks for your help

(*)
Control Panel -> Date, Time, Languages, and Regional Options -> Add other Languages
Select the Advanced tab and choose Russian as language.

[vladmir]

It works only for saved pages with win1251 in meta tag
For
<meta http-equiv="content-type" content="text/html; charset=UTF-8" />
doesn't work source and css-js
I,ll try more.

[davide ficano]

It works only for saved pages with win1251 in meta tag
For
<meta http-equiv="content-type" content="text/html; charset=UTF-8" />
doesn't work source and css-js
I,ll try more.

Does work with IETab?

I'm unable to find a original Russian WinXP

[vladmir]

IETab opens source of utf-8 pages

[vladmir]

New info by another user in forum.mozilla-russia.org
The same problem with the function in FF 3.0.1 itself

pref("view_source.editor.external", false);
set to true and set path to editor in
pref("view_source.editor.path", "");
Then
View -> Page Source

Error with pages with win-1251 and utf-8

===
In Netscape 9.0 (FF 2.0) the same.

[vladmir]

View source.editor.external - MozillaZine Knowledge Base
http://kb.mozillazine.org/View_source.editor.external

View source.editor.path - MozillaZine Knowledge Base
http://kb.mozillazine.org/View_source.editor.path

Bug 172817 – Allow external source viewer/editor
https://bugzilla.mozilla.org/show_bug.cgi?id=172817

[vladmir]

Bug 408923 – View source with an external editor doesn't works with a web page with a ' in title and UTF-8 encoding, editor is opened but file is not found
https://bugzilla.mozilla.org/show_bug.cgi?id=408923

[davide ficano]

IETab opens source of utf-8 pages

After many tests I'm unable to allow IETab to open UTF-8 local pages.

I use your test page and try to open using IETab and windows notepad inside a clean Firefox 3.0.3 but I receive error from notepad
"The filename, directory name, or volume label syntax is incorrect"

This simply means the IETab passed file path isn't recognized from Windows Operating System

[vladmir]

OK, I,ll see on new winXP.