Loading ...
Sorry, an error occurred while loading the content.

527Re: win2k IME, multibyte/width characters, etc. (patch)

Expand Messages
  • Glenn F. Maynard
    Sep 23 12:31 AM
    • 0 Attachment
      On Sun, Sep 23, 2001 at 04:11:44PM +0900, Muraoka Taro wrote:
      > > I'll attach the un-reversed patch. Is it working for you? Also,
      > > does pasting work for you in UTF-8 mode? (or, for reference, CP932?)
      >
      > You use ImmGetCompositionStringW().
      > But ImmGetCompositionStringW() is only available for NT/2K, isn't it?.
      > How about 9x/Me?

      This code is for the IME2000, which is in Win2K and up. We must use
      the wide version; the ANSI version dumps question marks. Pretty useless.
      This code should never be touched if a user isn't using the IME.

      In theory, WM_IME_CHAR should be used for this; it's documented as being
      used for non-Unicode windows to retrieve wide characters from the IME.
      But ... it only gives question marks. ?! Since we trap WM_IME_COMPOSITION
      and don't pass completions to DefWindowProc, WM_IME_CHAR is suppressed.

      Oh, and the iconv stuff hasn't been tested, but vim only uses latin1,
      Unicode or DBCS internally (according to mbyte.c comments), so it'll never
      be used by this code.

      --
      Glenn Maynard
    • Show all 54 messages in this topic