用于解析资源 (.rc) 文件的正则表达式

Question

最终，我只想从 .rc 文件中提取字符串，以便我可以翻译它们，但任何与 .rc 文件相关的内容都对我有用。

Answer 1

I'd consider usage of gettext and .PO files , if your program fits GNU license

1) I'd suggest extracting from .rc files using state machine algorithm.

void ProcessLine(const char * str)
{
   if (strstr(str, " DIALOG"))
      state = Scan;
   else if (strstr(str, " MENU"))
      state = Scan;
   else if (strstr(str, " STRINGTABLE"))
      state = Scan;
   else if (strstr(str, "END"))
      state = DontScan;

   if (state == Scan)
   {
      const char * cur = sLine;
      string hdr = ...// for example "# file.rc:453"
      string msgid;
      string msgid = "";
      while (ExtractString(sLine, cur, msgid))
      {
         if (msgid.empty())
            continue;
         if (IsPredefined(msgid))
            continue;
         if (msgid.find("IDB_") == 0 || msgid.find("IDC_") == 0)
            continue;
         WritePoString(hdr, msgid, msgstr);
      }
   }
}

2) When extracting string inside ExtractString() you should consider that char " is represented as "", and there are also chars like \\t \\n \\r. So state machine is also a good option here.

The following string:

LTEXT           "Mother has washed ""Sony"", then \taquarium\\shelves\r\nand probably floors",IDC_TEXT1,24,14,224,19

represents such label on a dialog:

Mother has washed "Sony", then aquarium\shelves
and probably floors

3) Then on program startup you should load .po file via gettext and for each dialog translate its string on startup using a function like this:

int TranslateDialog(CWnd& wnd)
{
    int i = 0;
    CWnd *pChild;
    CString text;

    //Translate Title
wnd.GetWindowText(text);
LPCTSTR translation = Translate(text);
    window.SetWindowText(translation);

    //Translate child windows
    pChild=wnd.GetWindow(GW_CHILD);
    while(pChild)
    {
        i++;
    Child->GetWindowText(Text);//including NULL
        translation = Translate(Text);
        pChild->SetWindowText(translation);
        pChild = pChild->GetWindow(GW_HWNDNEXT);
    }
    return i;
}

Answer 2

Although rc files seems an obvious starting point for translation, it's not. The job of developers is to make sure the app is translatable. It's not to manage translations. Starting translations from the exe, although somewhat counter-intuitive, is a way better idea. Read more about it here: http://www.apptranslator.com/misconceptions.html

Answer 3

Maybe this helps? ( http://social.msdn.microsoft.com/forums/en-US/regexp/thread/5e87fce9-ec73-42eb-b2eb-c821e95e0d31/ )

They are using the following regex to find the stringtable in the rc source:

(?<=\\bSTRINGTABLE\\s+BEGIN\\s+).*?(?=\\s+END\\b)

Edit - And you can read the key values pairs with the following statement with the MultiLine option:

@"\\s+(.*?)\\s+""(.*)""";

Answer 4

如果是 rc，最好使用像http://www.soft-gems.net/index.php/java/windows-resource-file-parser-and-converter这样的高级解析器

Answer 5

This sounds like a job for a SED script.

By running this command line: sed.exe -n -f sed.txt test.rc

The following SED script will extract all the quoted strings from the input test.rc file:

# Run Script Using This Command Line
#
#   sed.exe -n -f sed.txt test.rc
#

# Check for lines that contain strings
/\".*\"/ {
    # print the string part of the line only
    s/\(.*\)\(\".*\"\)\(.*\)/\2/ p
}

Answer 6

ResxCrunch will be out sometimes soon. It will edit multiple resource files in multiple languages in one single table.