C++ std::wofstream unicode issue.

Question

I wrote this code using VS 2012:

std::wofstream logout("my_log.txt", std::ios_base::out | std::ios_base::trunc);
std::locale utf8_locale(locale(), new codecvt_utf8<wchar_t>);
logout.imbue(utf8_locale);

if (!logout.is_open()) 
{
 printf("Cannot open file.\n"); 
 return 1; 
}
else printf("Log file created.\n");

logout << "Client IP            │"<< "Recv time                │"<< "Request               │"<< "Response     "<<endl;
logout << "─────────────────────┼"<< "─────────────────────────┼"<<endl;

In my_log.txt file all unicode symbols are replaced with "?????". I want to create something like a log file table. If i use standart ASCII symbols like "---"it will work, and they all are shown surrectly. Ive tried to change global locale, but i dont succeed.

Are you sure the editor you're using to look at your file supports UTF8/is in unicode mode? — Mat, Sep 12 '15 at 14:23
Hm, actually i dont sure. I had tried to open this file using MS Word and notepad but the problem stays. — N. Zazulin, Sep 12 '15 at 14:31
No MS product supports UTF8. All MS/Windows is UTF16. Try a hex viewer to see what's really in your file. — Richard Critten, Sep 12 '15 at 14:36
@RichardCritten All unicode symbols in my file are shown as 3F in hex format. — N. Zazulin, Sep 12 '15 at 15:31

score 0 · Accepted Answer · answered Sep 12 '15 at 15:33

This is why we don't use anything but ASCII in string literals.

Your code is fine, but your string literals evidently don't contain what you think they do. This is probably due to a misconfigured or underfeatured text editor.

Use escape sequences instead (\uXXXX) so you know what you're getting. In 2015 you can't really go wrong with source code formed from ASCII characters.

score 0 · Answer 2 · answered Sep 12 '15 at 16:33

0

Oh, this is actually so funny.

wchar_t *s1 = L"\u2502"; // "│"
wchar_t *s2 = L"\u2500"; // "─"
wchar_t *s3 = L"\u253C"; // "┼"
logout << s1<<s2<<s3<<endl;

In my .txt file now i see unicode numbers. But i can use the same struct without (\uXXXX)

logout << L" │ ─ ┼"<<endl;

This works aswell.

answered Sep 12 '15 at 16:33

N. Zazulin

35
3

Suggestion: `wchar_t const s1 = L'\u2502';`. – Ulrich Eckhardt Sep 12 '15 at 17:24
Well, those non-ASCII literals almost certainly won't work with GCC. It's non-portable. – Sep 12 '15 at 17:35

C++ std::wofstream unicode issue.

2 Answers2