Re: Minor corrections for 20021102

[Thomas Dickey]

On Tue, Nov 05, 2002 at 07:55:03AM +0100, Philippe Blain wrote:
> >From Philippe Blain, Bordeaux, FRANCE.
> My old computer: P133 - 8,4 Go - 32 Mo Red Hat Linux 7.0
> 
> To maintainers of 'ncurses'.(and to Mr Dickey)
> Subject: Corrections for ncurses-5.3-20021102+
> 
> Here are some problems I found :
> 
> ----------------------------------------------------------------------------
> File : ncurses/base/lib_insch.c
> Function : _nc_insert_ch()
> 
> When an inserted string goes beyond the right edge of window, last inserted
> chars overwrites each other at win->_maxx, resulting in the last char of
> the string visible at that position.
> Minor modif possible, question of choice, as :
> 
>     default:
>     if (is7bits(ch) && iscntrl(ch)) {
>         _nc_insert_ch(win, '^');
>         _nc_insert_ch(win, '@' + (ch));
>     } else {
> ==>     if (win->_curx <= win->_maxx) { /* Insertion possible */

I did notice that, but (by the time I was done getting the test-case to work
properly), it was getting late.  But thanks for the reminder, since I'd
probably have forgotten to go back to this point.

>             struct ldat *line = &(win->_line[win->_cury]);
>             NCURSES_CH_T *end = &(line->text[win->_curx]);
>             NCURSES_CH_T *temp1 = &(line->text[win->_maxx]);
>             NCURSES_CH_T *temp2 = temp1 - 1;
> 
>             SetChar2(wch, ch);
> 
>             CHANGED_TO_EOL(line, win->_curx, win->_maxx);
>             while (temp1 > end)
>                 *temp1-- = *temp2--;
> 
>             *temp1 = _nc_render(win, wch);
> 
>             win->_curx++;
>         }
>     }
>     break;
> 
> ----------------------------------------------------------------------------
> File : ncurses/base/lib_instr.c
> Function : winnstr()
> 
> According to man pages, function should return ERR (-1) upon failure:
> 
> NCURSES_EXPORT(int) winnstr(WINDOW *win, char *str, int n)
> {
>     int i = 0, row, col;
> 
>     T((T_CALLED("winnstr(%p,%p,%d)"), win, str, n));
> 
> ==> if (!win || !str) returnCode(ERR);

ok

> ----------------------------------------------------------------------------
> File : ncurses/ncurses.priv.h
> 
> Wide chars. Macro PUTC seems incorrect :
> 
> #define PUTC(ch,b)\
>     do {\
>         if(!isnac(ch)) { \
>             memset (&PUT_st, '\0', sizeof (PUT_st));\
>             PUTC_i = 0;\
>             do {\
>                 PUTC_ch = (PUTC_i < CCHARW_MAX ? (ch).chars[PUTC_i]:L'\0');\
> ===>            PUTC_n = wcrtomb(PUTC_buf, (ch).chars[PUTC_i], &PUT_st);\
>                 if (PUTC_ch == L'\0') --PUTC_n;\
>                 if (PUTC_n <= 0) break;\
>                 fwrite (PUTC_buf, (unsigned) PUTC_n, 1, b);\
>                 ++PUTC_i;\
>             } while (PUTC_ch != L'\0');\
>         }\
>      } while (0)
> 
> Would it be    PUTC_n = wcrtomb(PUTC_buf, PUTC_ch, &PUT_st);\   instead ?

maybe (will have to read & see, myself - I don't know offhand)
 
> ----------------------------------------------------------------------------
> File : ncurses/curses.h.in
> 
> NCURSES FOR THE DUMMIES:
> Please, can you comment in source for all people, the definition of
> cchar_t (wide chars) ? Some indications will be welcome.

The reason for the array is that some locales may have characters that
are built up by overstriking.  The 0th entry of the array is (almost always)
a "spacing" character, and other entries are "nonspacing".
 
> #define CCHARW_MAX    5
> typedef struct
> {
>     attr_t    attr;
>     wchar_t    chars[CCHARW_MAX];
> }
> cchar_t;
> 
> I don't understand what that array of five wide chars is for ?
> Is it :
>     a UCS-4 character (wchar_t) followed with accents/diacritical marks ?
>     or the UTF-8 encoding for a Unicode char (6 bytes max) ?
> 
> 
> NOTES :
> In the GNU glibc 2.2, wchar_t is officially intended to be used only for
> storing the 32-bits ISO 10646 values (Universal Character Set UCS-4 codes),
> independent of the currently used locale.
> The ISO C99 multi-byte conversion functions are used to convert between
> wchar_t and any locale-dependent multibyte encoding (UTF-8, ISO 8859-1, ...)
> 
> ----------------------------------------------------------------------------
> - Philippe
> 
> 
> 
> 
> _______________________________________________
> Bug-ncurses mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/bug-ncurses

-- 
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net