mbyte.c

Go to the documentation of this file.

 
#include "config.h"
#include <limits.h>
#include <stdbool.h>
#include <string.h>
#include <wchar.h>
#include <wctype.h>
#include "mbyte.h"
#include "buffer.h"
#include "charset.h"
#include "ctype2.h"
#include "memory.h"
#include "pool.h"
#include "string2.h"
 
bool OptLocales; 

int mutt_mb_charlen(const char *s, int *width)
{
  if (!s || (*s == '\0'))
    return 0;
 
  wchar_t wc = 0;
  mbstate_t mbstate = { 0 };
 
  size_t n = mutt_str_len(s);
  size_t k = mbrtowc(&wc, s, n, &mbstate);
  if (width)
    *width = wcwidth(wc);
  return ((k == ICONV_ILLEGAL_SEQ) || (k == ICONV_BUF_TOO_SMALL)) ? -1 : k;
}

bool mutt_mb_get_initials(const char *name, char *buf, size_t buflen)
{
  if (!name || !buf)
    return false;
 
  while (*name)
  {
    /* Char's length in bytes */
    int clen = mutt_mb_charlen(name, NULL);
    if (clen < 1)
      return false;
 
    /* Ignore punctuation at the beginning of a word */
    if ((clen == 1) && mutt_ispunct(*name))
    {
      name++;
      continue;
    }
 
    if (clen >= buflen)
      return false;
 
    /* Copy one multibyte character */
    buflen -= clen;
    while (clen--)
      *buf++ = *name++;
 
    /* Skip to end-of-word */
    for (; *name; name += clen)
    {
      clen = mutt_mb_charlen(name, NULL);
      if (clen < 1)
        return false;
      if ((clen == 1) && (mutt_isspace(*name) || (*name == '-')))
        break;
    }
 
    /* Skip any whitespace, or hyphens */
    while (*name && (mutt_isspace(*name) || (*name == '-')))
      name++;
  }
 
  *buf = '\0';
  return true;
}

int mutt_mb_width(const char *str, int col, bool indent)
{
  if (!str || !*str)
    return 0;
 
  bool nl = false;
  int total_width = 0;
  mbstate_t mbstate = { 0 };
 
  size_t str_len = mutt_str_len(str);
 
  while (*str && (str_len > 0))
  {
    wchar_t wc = L'\0';
    size_t consumed = mbrtowc(&wc, str, str_len, &mbstate);
    if (consumed == 0)
      break;
 
    if (consumed == ICONV_ILLEGAL_SEQ)
    {
      memset(&mbstate, 0, sizeof(mbstate));
      wc = ReplacementChar;
      consumed = 1;
    }
    else if (consumed == ICONV_BUF_TOO_SMALL)
    {
      wc = ReplacementChar;
      consumed = str_len;
    }
 
    int wchar_width = wcwidth(wc);
    if (wchar_width < 0)
      wchar_width = 1;
 
    if ((wc == L'\t') || (nl && (wc == L' ')))
    {
      /* correctly calc tab stop, even for sending as the line should look
       * pretty on the receiving end */
      nl = false;
      wchar_width = 8 - (col % 8);
    }
    else if (indent && (wc == '\n'))
    {
      /* track newlines for display-case: if we have a space after a newline,
       * assume 8 spaces as for display we always tab-fold */
      nl = true;
    }
 
    total_width += wchar_width;
    str += consumed;
    str_len -= consumed;
  }
 
  return total_width;
}

int mutt_mb_wcwidth(wchar_t wc)
{
  int n = wcwidth(wc);
  if (IsWPrint(wc) && (n > 0))
    return n;
  if (!(wc & ~0x7f))
    return 2;
  if (!(wc & ~0xffff))
    return 6;
  return 10;
}

int mutt_mb_wcswidth(const wchar_t *s, size_t n)
{
  if (!s)
    return 0;
 
  int w = 0;
  while (n--)
    w += mutt_mb_wcwidth(*s++);
  return w;
}

size_t mutt_mb_width_ceiling(const wchar_t *s, size_t n, int w1)
{
  if (!s)
    return 0;
 
  const wchar_t *s0 = s;
  int w = 0;
  for (; n; s++, n--)
    if ((w += mutt_mb_wcwidth(*s)) > w1)
      break;
  return s - s0;
}

void buf_mb_wcstombs(struct Buffer *dest, const wchar_t *wstr, size_t wlen)
{
  if (!dest || !wstr)
    return;
 
  // Give ourselves 4 utf-8 bytes per wide character
  buf_alloc(dest, 4 * wlen);
 
  mbstate_t mbstate = { 0 };
  size_t k = 0;
 
  char *buf = dest->data;
  size_t buflen = dest->dsize;
 
  for (; (wlen > 0) && (buflen >= MB_LEN_MAX); buf += k, buflen -= k, wstr++, wlen--)
  {
    k = wcrtomb(buf, *wstr, &mbstate);
    if (k == ICONV_ILLEGAL_SEQ)
      break;
    if (*wstr == L'\0')
      break;
  }
 
  *buf = '\0';
  buf_fix_dptr(dest);
}

size_t mutt_mb_mbstowcs(wchar_t **pwbuf, size_t *pwbuflen, size_t i, const char *buf)
{
  if (!pwbuf || !pwbuflen || !buf)
    return 0;
 
  wchar_t wc = 0;
  mbstate_t mbstate = { 0 };
  size_t k;
  wchar_t *wbuf = *pwbuf;
  size_t wbuflen = *pwbuflen;
 
  while (*buf != '\0')
  {
    memset(&mbstate, 0, sizeof(mbstate));
    for (; (k = mbrtowc(&wc, buf, MB_LEN_MAX, &mbstate)) &&
           (k != ICONV_ILLEGAL_SEQ) && (k != ICONV_BUF_TOO_SMALL);
         buf += k)
    {
      if (i >= wbuflen)
      {
        wbuflen = i + 20;
        MUTT_MEM_REALLOC(&wbuf, wbuflen, wchar_t);
      }
      wbuf[i++] = wc;
    }
    if ((*buf != '\0') && ((k == ICONV_ILLEGAL_SEQ) || (k == ICONV_BUF_TOO_SMALL)))
    {
      if (i >= wbuflen)
      {
        wbuflen = i + 20;
        MUTT_MEM_REALLOC(&wbuf, wbuflen, wchar_t);
      }
      wbuf[i++] = ReplacementChar;
      buf++;
    }
  }
  *pwbuf = wbuf;
  *pwbuflen = wbuflen;
  return i;
}

bool mutt_mb_is_shell_char(wchar_t ch)
{
  static const wchar_t shell_chars[] = L"<>&()$?*;{}| "; /* ! not included because it can be part of a pathname in NeoMutt */
  return wcschr(shell_chars, ch);
}

bool mutt_mb_is_lower(const char *s)
{
  if (!s)
    return false;
 
  wchar_t wc = 0;
  mbstate_t mbstate = { 0 };
  size_t l;
 
  memset(&mbstate, 0, sizeof(mbstate));
  size_t n = mutt_str_len(s);
 
  for (; (n > 0) && (*s != '\0') && (l = mbrtowc(&wc, s, n, &mbstate)) != 0; s += l, n -= l)
  {
    if ((l == ICONV_BUF_TOO_SMALL) || (l == ICONV_ILLEGAL_SEQ))
      return false; // error; assume upper-case
    if (iswalpha((wint_t) wc) && iswupper((wint_t) wc))
      return false; // upper-case
  }
 
  return true; // lower-case
}

bool mutt_mb_is_display_corrupting_utf8(wchar_t wc)
{
  if ((wc == (wchar_t) 0x00ad) || /* soft hyphen */
      (wc == (wchar_t) 0x200e) || /* left-to-right mark */
      (wc == (wchar_t) 0x200f) || /* right-to-left mark */
      (wc == (wchar_t) 0xfeff))   /* zero width no-break space */
  {
    return true;
  }
 
  /* left-to-right isolate, right-to-left isolate, first strong isolate,
   * pop directional isolate */
  if ((wc >= (wchar_t) 0x2066) && (wc <= (wchar_t) 0x2069))
    return true;
 
  /* left-to-right embedding, right-to-left embedding, pop directional formatting,
   * left-to-right override, right-to-left override */
  if ((wc >= (wchar_t) 0x202a) && (wc <= (wchar_t) 0x202e))
    return true;
 
  /* arabic letter mark */
  if (wc == (wchar_t) 0x061c)
    return true;
 
  return false;
}

int mutt_mb_filter_unprintable(char **s)
{
  if (!s || !*s)
    return -1;
 
  wchar_t wc = 0;
  size_t k, k2;
  char scratch[MB_LEN_MAX + 1];
  char *p = *s;
  mbstate_t mbstate1 = { 0 };
  mbstate_t mbstate2 = { 0 };
 
  struct Buffer *buf = buf_pool_get();
  for (; (k = mbrtowc(&wc, p, MB_LEN_MAX, &mbstate1)); p += k)
  {
    if ((k == ICONV_ILLEGAL_SEQ) || (k == ICONV_BUF_TOO_SMALL))
    {
      k = 1;
      memset(&mbstate1, 0, sizeof(mbstate1));
      wc = ReplacementChar;
    }
    if (CharsetIsUtf8 && IsBOM(wc))
    {
      continue;
    }
    if (!IsWPrint(wc))
      wc = '?';
    else if (CharsetIsUtf8 && mutt_mb_is_display_corrupting_utf8(wc))
      continue;
    k2 = wcrtomb(scratch, wc, &mbstate2);
    scratch[k2] = '\0';
    buf_addstr(buf, scratch);
  }
  FREE(s);
 
  if (buf_is_empty(buf))
    *s = MUTT_MEM_CALLOC(1, char); // Fake empty string
  else
    *s = buf_strdup(buf);
 
  buf_pool_release(&buf);
  return 0;
}