Converting Unicode wxString to UTF-8 Topic is solved

[geralds]

I'm using the Unicode build of wxW and I was delighted that wxW will convert Expat's UTF-8 text nodes to wxStrings as follows:

Code: Select all

// s is a std::string containing a UTF-8 string
wxString node = wxString(s.c_str(), wxConvUTF8, s.size());

I can then set the text of my Scintilla component to the wxString. My question is: how do I then convert the text back to UTF-8 for final output?

Sorry if this is a very lame question; I've just started using wxW!

[heda]

Hi,
AFAIK, you can do that using:
s = node.mbstr();

(s is a std::string variable and node is a wxString one)

Good Luck

[geralds]

Thanks!

[eco]

It's actually wxString::mb_str(), rather than wxString::mbstr(). Also, you should probably be specifying the target encoding like so:

Code: Select all

const char* s = node.mb_str(wxConvUTF8);

And std::string has no concept of UTF-8 so I'd recommend avoiding stuffing a UTF-8 string in one.

[heda]

It's actually wxString::mb_str(), rather than wxString::mbstr()

Oh yes, you're right. It was just a typing mistake; sorry.

Thanks for you comment.

[Ryan Norton]

And std::string has no concept of UTF-8 so I'd recommend avoiding stuffing a UTF-8 string in one.

It should be OK actually. std::string and wxString are just a bunch of bytes and don't check for zeros etc. so you should be ok with any encoding as long as you remember to explicitly specify the length.

[geralds]

Thanks heda, eco and Ryan! eco's suggestion works perfectly. I only wish I'd known about wxWidgets a year ago! This removes the last GPL function from my program, so I can finally switch to LGPL...

[leio]

And UTF-8 doesn't really contain \0's in strings, it is used for specifying the end of a string, as is common in C strings.

[eco]

It should be OK actually. std::string and wxString are just a bunch of bytes and don't check for zeros etc. so you should be ok with any encoding as long as you remember to explicitly specify the length.

And UTF-8 doesn't really contain \0's in strings, it is used for specifying the end of a string, as is common in C strings.

Both good points. I suppose the best advice would be to simply not use any character (byte) specific operations on the data. If you do that, either should work great.