win32 read java preference from c++ code

Question

One of our program writes program information(window title, memory etc) in Java Preferences. On windows this is available under registry. How can I read the values written by Java program using c (or c++).

Looks like API I should use is RegGetValue. Is this guaranteed to work on Windows XP 32 bit?

The String written by java is UTF-8 encoded. How do I read such strings in windows (win32 or vc++)

Cheers, Jayan

Answer 1

For accessing the registry in C++, you may be interested in the CRegKey ATL class.

Answer 2

I dug into this a bit. RegGetValue() is a new registry call that takes care of some underlying nastiness of the traditional way of querying the registry (RegQueryValueEx). There's some good info about the difference here: http://blogs.msdn.com/larryosterman/archive/2006/01/12/512115.aspx

If you need backwards compatibility, RegGetValue() isn't going to work, so you should be using RegQueryValueEx to read data from the registry.

And now on to what I think is the real question:

1. What do you get back when you use RegQueryValueEx() ?
2. How do you know that the values stored in the registry are in utf-8 encoding? Is it stored as a byte array in the registry, or as a REG_SZ?
3. Have you looked at the value using regedit? What do you see?

So if the question is how to convert a UTF-8 encoded string to an ascii null terminated string, then you should probably change the title of your question. For reference, I found this library that may be of use:

http://utfcpp.sourceforge.net

When I tried the link a few minutes ago, the server timed out - probably SF maintenance going on.

But I would suggest that you make real sure that the values in the registry aren't stored as REG_SZ entries already.

Answer 3

RegGetValue is a newer API. If you want your program to work on XP 32-bit (actually back before Windows 2000) use RegQueryValueEx

A UTF-8 string is NUL-terminated just like any single-byte ASCII string. However you might be interested in converting it to UTF-16. For that you'd want to use MultiByteToWideChar, setting CodePage to CP_UTF8. From UTF-16 you can convert to any other code page using WideCharToMultiByte.