content_info.c

Go to the documentation of this file.

 
#include "config.h"
#include <errno.h>
#include <stdbool.h>
#include <stdio.h>
#include <string.h>
#include <sys/stat.h>
#include "mutt/lib.h"
#include "config/lib.h"
#include "email/lib.h"
#include "core/lib.h"
#include "lib.h"

void mutt_update_content_info(struct Content *info, struct ContentState *s,
                              char *buf, size_t buflen)
{
  bool from = s->from;
  int whitespace = s->whitespace;
  bool dot = s->dot;
  int linelen = s->linelen;
  bool was_cr = s->was_cr;
 
  /* A NULL buffer signals end-of-file; finalize any pending state */
  if (!buf) /* This signals EOF */
  {
    if (was_cr)
      info->binary = true;
    if (linelen > info->linemax)
      info->linemax = linelen;
 
    return;
  }
 
  /* Scan each byte in the buffer and classify it for MIME encoding decisions */
  for (; buflen; buf++, buflen--)
  {
    char ch = *buf;
 
    /* A CR not followed by LF indicates binary content */
    if (was_cr)
    {
      was_cr = false;
      if (ch == '\n')
      {
        /* CR+LF: complete line ending; record line statistics */
        if (whitespace)
          info->space = true;
        if (dot)
          info->dot = true;
        if (linelen > info->linemax)
          info->linemax = linelen;
        whitespace = 0;
        dot = false;
        linelen = 0;
        continue;
      }
 
      info->binary = true;
    }
 
    linelen++;
    if (ch == '\n')
    {
      /* Bare LF line ending */
      info->crlf++;
      if (whitespace)
        info->space = true;
      if (dot)
        info->dot = true;
      if (linelen > info->linemax)
        info->linemax = linelen;
      whitespace = 0;
      linelen = 0;
      dot = false;
    }
    else if (ch == '\r')
    {
      info->crlf++;
      info->cr = true;
      was_cr = true;
      continue;
    }
    else if (ch & 0x80)
    {
      /* High-bit character: needs 8-bit or Base64 encoding */
      info->hibin++;
    }
    else if ((ch == '\t') || (ch == '\f'))
    {
      info->ascii++;
      whitespace++;
    }
    else if (ch == 0)
    {
      /* NUL byte: forces binary encoding */
      info->nulbin++;
      info->lobin++;
    }
    else if ((ch < 32) || (ch == 127))
    {
      info->lobin++;
    }
    else
    {
      /* Detect "From " at the start of a line (mbox from-quoting) */
      if (linelen == 1)
      {
        if ((ch == 'F') || (ch == 'f'))
          from = true;
        else
          from = false;
        if (ch == '.')
          dot = true;
        else
          dot = false;
      }
      else if (from)
      {
        /* Check chars 2-4 for "rom" to complete "From" detection */
        if ((linelen == 2) && (ch != 'r'))
        {
          from = false;
        }
        else if ((linelen == 3) && (ch != 'o'))
        {
          from = false;
        }
        else if (linelen == 4)
        {
          if (ch == 'm')
            info->from = true;
          from = false;
        }
      }
      if (ch == ' ')
        whitespace++;
      info->ascii++;
    }
 
    if (linelen > 1)
      dot = false;
    if ((ch != ' ') && (ch != '\t'))
      whitespace = 0;
  }
 
  s->from = from;
  s->whitespace = whitespace;
  s->dot = dot;
  s->linelen = linelen;
  s->was_cr = was_cr;
}

struct Content *mutt_get_content_info(const char *fname, struct Body *b,
                                      struct ConfigSubset *sub)
{
  struct Content *info = NULL;
  struct ContentState cstate = { 0 };
  FILE *fp = NULL;
  char *fromcode = NULL;
  char *tocode = NULL;
  char buf[100] = { 0 };
  size_t r;
 
  struct stat st = { 0 };
 
  if (b && !fname)
    fname = b->filename;
  if (!fname)
    return NULL;
 
  fp = mutt_file_fopen(fname, "r");
  if (!fp)
  {
    mutt_debug(LL_DEBUG1, "%s: %s (errno %d)\n", fname, strerror(errno), errno);
    return NULL;
  }
 
  if (fstat(fileno(fp), &st) == -1)
  {
    mutt_error(_("Can't stat %s: %s"), fname, strerror(errno));
    mutt_file_fclose(&fp);
    return NULL;
  }
 
  if (!S_ISREG(st.st_mode))
  {
    mutt_error(_("%s isn't a regular file"), fname);
    mutt_file_fclose(&fp);
    return NULL;
  }
 
  info = MUTT_MEM_CALLOC(1, struct Content);
 
  const char *const c_charset = cc_charset();
  if (b && (b->type == TYPE_TEXT) && (!b->noconv && !b->force_charset))
  {
    const struct Slist *const c_attach_charset = cs_subset_slist(sub, "attach_charset");
    const struct Slist *const c_send_charset = cs_subset_slist(sub, "send_charset");
    struct Slist *c_charset_slist = slist_parse(c_charset, D_SLIST_SEP_COLON);
 
    const struct Slist *fchs = b->use_disp ?
                                   (c_attach_charset ? c_attach_charset : c_charset_slist) :
                                   c_charset_slist;
 
    struct Slist *chs = slist_parse(mutt_param_get(&b->parameter, "charset"), D_SLIST_SEP_COLON);
 
    if (c_charset && (chs || c_send_charset) &&
        (mutt_convert_file_from_to(fp, fchs, chs ? chs : c_send_charset, &fromcode,
                                   &tocode, info) != ICONV_ILLEGAL_SEQ))
    {
      if (!chs)
      {
        char chsbuf[256] = { 0 };
        mutt_ch_canonical_charset(chsbuf, sizeof(chsbuf), tocode);
        mutt_param_set(&b->parameter, "charset", chsbuf);
      }
      FREE(&b->charset);
      b->charset = fromcode;
      fromcode = NULL;
      FREE(&tocode);
      mutt_file_fclose(&fp);
      slist_free(&c_charset_slist);
      slist_free(&chs);
      return info;
    }
 
    slist_free(&c_charset_slist);
    slist_free(&chs);
  }
 
  rewind(fp);
  while ((r = fread(buf, 1, sizeof(buf), fp)))
    mutt_update_content_info(info, &cstate, buf, r);
  mutt_update_content_info(info, &cstate, 0, 0);
 
  mutt_file_fclose(&fp);
 
  if (b && (b->type == TYPE_TEXT) && (!b->noconv && !b->force_charset))
  {
    mutt_param_set(&b->parameter, "charset",
                   (!info->hibin                                 ? "us-ascii" :
                    c_charset && !mutt_ch_is_us_ascii(c_charset) ? c_charset :
                                                                   "unknown-8bit"));
  }
 
  return info;
}