This bug report was migrated from our old Bugzilla tracker.

These attachments are available in the static archive:

• Proposed patch to handle Unicode... (SDL-win-SetWindowTextW.diff, text/plain, 2007-02-13 04:54:06 +0000, 1653 bytes)

Reported in version: 1.2.11
Reported for operating system, platform: Windows (All), x86

Comments on the original bug report:

On 2006-11-17 07:06:16 +0000, aki. wrote:

Hello,

When Japanese is given to SDL_WM_SetCaption() on Windows, it garbles.

WinAPI receives string corresponding to the code page of the system.
Therefore, I think that it should use such as following codes instead of SDL_iconv_utf8_latin1():

char *SDL_iconv_utf8_acp(char *in) {
Uint16 *lpsz = SDL_iconv_utf8_ucs2(in);
size_t len = WideCharToMultiByte(CP_ACP, 0, lpsz, -1, NULL, 0, NULL, NULL);
char *out = SDL_malloc(len + 1);
WideCharToMultiByte(CP_ACP, 0, lpsz, -1, out, len, NULL, NULL);
SDL_free(lpsz);
return out;
}

Regards,
aki. (c@tqzh.tk)

On 2007-02-13 04:54:06 +0000, Ryan C. Gordon wrote:

Created attachment 189
Proposed patch to handle Unicode...

Here's a proposed patch...it is untested and may not even compile, yet.

It will see if SetWindowTextW() is available in kernel32.dll (as it will be on WinNT, WinXP, Win2000 and Vista), and if so, it feeds the ucs-2 string to it. This will allow, for example, Japanese characters to be used in the titlebar on a WinXP box that uses a Russian codepage.

If SetWindowsTextW() isn't available (Win95, Win98, WinME), then it will use WideCharToMultiByte(CP_ACP, ...) as originally proposed. This will not necessarily work everywhere, but it will probably work in most real life scenarios, and definitely better than assuming a latin1 codepage as we currently do.

If there are no objections, I'll commit this patch once I've had a chance to compile and test it on Windows.

--ryan.

On 2007-02-16 01:21:47 +0000, Ryan C. Gordon wrote:

Fixed in svn revision # 2985 for the 1.2 branch.

--ryan.