rfc822.c

广泛使用的邮件服务器！同时

/* ========================================================================
 * Copyright 1988-2008 University of Washington
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * 
 * ========================================================================
 */

/*
 * Program:	RFC 2822 and MIME routines
 *
 * Author:	Mark Crispin
 *		UW Technology
 *		University of Washington
 *		Seattle, WA  98195
 *		Internet: [e-mail]
 *
 * Date:	27 July 1988
 * Last Edited:	14 May 2008
 *
 * This original version of this file is
 * Copyright 1988 Stanford University
 * and was developed in the Symbolic Systems Resources Group of the Knowledge
 * Systems Laboratory at Stanford University in 1987-88, and was funded by the
 * Biomedical Research Technology Program of the NationalInstitutes of Health
 * under grant number RR-00785.
 */


#include <ctype.h>
#include <stdio.h>
#include <time.h>
#include "c-client.h"


/* Support for deprecated features in earlier specifications.  Note that this
 * module follows RFC 2822, and all use of "rfc822" in function names is
 * for compatibility.  Only the code identified by the conditionals below
 * follows the earlier documents.
 */

#define RFC733 1		/* parse "at" */
#define RFC822 0		/* generate A-D-L (MUST be 0 for 2822) */

/* RFC-822 static data */

#define RFC822CONT "    "	/* RFC 2822 continuation */

				/* should have been "Remailed-" */
#define RESENTPREFIX "ReSent-"
static char *resentprefix = RESENTPREFIX;
				/* syntax error host string */
static const char *errhst = ERRHOST;


/* Body formats constant strings, must match definitions in mail.h */

char *body_types[TYPEMAX+1] = {
  "TEXT", "MULTIPART", "MESSAGE", "APPLICATION", "AUDIO", "IMAGE", "VIDEO",
  "MODEL", "X-UNKNOWN"
};


char *body_encodings[ENCMAX+1] = {
  "7BIT", "8BIT", "BINARY", "BASE64", "QUOTED-PRINTABLE", "X-UNKNOWN"
};


/* Token delimiting special characters */

				/* RFC 2822 specials */
const char *specials = " ()<>@,;:\\\"[].\1\2\3\4\5\6\7\10\11\12\13\14\15\16\17\20\21\22\23\24\25\26\27\30\31\32\33\34\35\36\37\177";
				/* RFC 2822 phrase specials (no space) */
const char *rspecials = "()<>@,;:\\\"[].\1\2\3\4\5\6\7\10\11\12\13\14\15\16\17\20\21\22\23\24\25\26\27\30\31\32\33\34\35\36\37\177";
				/* RFC 2822 dot-atom specials (no dot) */
const char *wspecials = " ()<>@,;:\\\"[]\1\2\3\4\5\6\7\10\11\12\13\14\15\16\17\20\21\22\23\24\25\26\27\30\31\32\33\34\35\36\37\177";
				/* RFC 2045 MIME body token specials */
const char *tspecials = " ()<>@,;:\\\"[]/?=\1\2\3\4\5\6\7\10\11\12\13\14\15\16\17\20\21\22\23\24\25\26\27\30\31\32\33\34\35\36\37\177";

/* Subtype defaulting (a no-no, but regretably necessary...)
 * Accepts: type code
 * Returns: default subtype name
 */

char *rfc822_default_subtype (unsigned short type)
{
  switch (type) {
  case TYPETEXT:		/* default is TEXT/PLAIN */
    return "PLAIN";
  case TYPEMULTIPART:		/* default is MULTIPART/MIXED */
    return "MIXED";
  case TYPEMESSAGE:		/* default is MESSAGE/RFC822 */
    return "RFC822";
  case TYPEAPPLICATION:		/* default is APPLICATION/OCTET-STREAM */
    return "OCTET-STREAM";
  case TYPEAUDIO:		/* default is AUDIO/BASIC */
    return "BASIC";
  default:			/* others have no default subtype */
    return "UNKNOWN";
  }
}

/* RFC 2822 parsing routines */


/* Parse an RFC 2822 message
 * Accepts: pointer to return envelope
 *	    pointer to return body
 *	    pointer to header
 *	    header byte count
 *	    pointer to body stringstruct
 *	    pointer to local host name
 *	    recursion depth
 *	    source driver flags
 */

void rfc822_parse_msg_full (ENVELOPE **en,BODY **bdy,char *s,unsigned long i,
			    STRING *bs,char *host,unsigned long depth,
			    unsigned long flags)
{
  char c,*t,*d;
  char *tmp = (char *) fs_get ((size_t) i + 100);
  ENVELOPE *env = (*en = mail_newenvelope ());
  BODY *body = bdy ? (*bdy = mail_newbody ()) : NIL;
  long MIMEp = -1;		/* flag that MIME semantics are in effect */
  parseline_t pl = (parseline_t) mail_parameters (NIL,GET_PARSELINE,NIL);
  if (!host) host = BADHOST;	/* make sure that host is non-null */
  while (i && *s != '\n') {	/* until end of header */
    t = tmp;			/* initialize buffer pointer */
    c = ' ';			/* and previous character */
    while (i && c) {		/* collect text until logical end of line */
      switch (c = *s++) {	/* slurp a character */
      case '\015':		/* return, possible end of logical line */
	if (*s == '\n') break;	/* ignore if LF follows */
      case '\012':		/* LF, possible end of logical line */
				/* tie off unless next line starts with WS */
	if (*s != ' ' && *s != '\t') *t++ = c = '\0';
	break;
      case '\t':		/* tab */
	*t++ = ' ';		/* coerce to space */
	break;
      default:			/* all other characters */
	*t++ = c;		/* insert the character into the line */
	break;
      }
      if (!--i) *t++ = '\0';	/* see if end of header */
    }

				/* find header item type */
    if (t = d = strchr (tmp,':')) {
      *d++ = '\0';		/* tie off header item, point at its data */
      while (*d == ' ') d++;	/* flush whitespace */
      while ((tmp < t--) && (*t == ' ')) *t = '\0';
      ucase (tmp);		/* coerce to uppercase */
				/* external callback */
      if (pl) (*pl) (env,tmp,d,host);
      switch (*tmp) {		/* dispatch based on first character */
      case '>':			/* possible >From: */
	if (!strcmp (tmp+1,"FROM")) rfc822_parse_adrlist (&env->from,d,host);
	break;
      case 'B':			/* possible bcc: */
	if (!strcmp (tmp+1,"CC")) rfc822_parse_adrlist (&env->bcc,d,host);
	break;
      case 'C':			/* possible cc: or Content-<mumble>*/
	if (!strcmp (tmp+1,"C")) rfc822_parse_adrlist (&env->cc,d,host);
	else if ((tmp[1] == 'O') && (tmp[2] == 'N') && (tmp[3] == 'T') &&
		 (tmp[4] == 'E') && (tmp[5] == 'N') && (tmp[6] == 'T') &&
		 (tmp[7] == '-') && body)
	  switch (MIMEp) {
	  case -1:		/* unknown if MIME or not */
	    if (!(MIMEp =	/* see if MIME-Version header exists */
		  search ((unsigned char *) s-1,i,
			  (unsigned char *)"\012MIME-Version",(long) 13))) {
#if 1
	      /* This is a disgusting kludge, and most of the messages which
	       * benefit from it are spam.
	       */
	      if (!strcmp (tmp+8,"TRANSFER-ENCODING") ||
		  (!strcmp (tmp+8,"TYPE") && strchr (d,'/'))) {
		MM_LOG ("Warning: MIME header encountered in non-MIME message",
			PARSE);
		MIMEp = 1;	/* declare MIME now */
	      }
	      else
#endif
		break;		/* non-MIME message */
	    }
	  case T:		/* definitely MIME */
	    rfc822_parse_content_header (body,tmp+8,d);
	  }
	break;
      case 'D':			/* possible Date: */
	if (!env->date && !strcmp (tmp+1,"ATE")) env->date = cpystr (d);
	break;
      case 'F':			/* possible From: */
	if (!strcmp (tmp+1,"ROM")) rfc822_parse_adrlist (&env->from,d,host);
	else if (!strcmp (tmp+1,"OLLOWUP-TO")) {
	  t = env->followup_to = (char *) fs_get (1 + strlen (d));
	  while (c = *d++) if (c != ' ') *t++ = c;
	  *t++ = '\0';
	}
	break;
      case 'I':			/* possible In-Reply-To: */
	if (!env->in_reply_to && !strcmp (tmp+1,"N-REPLY-TO"))
	  env->in_reply_to = cpystr (d);
	break;

      case 'M':			/* possible Message-ID: or MIME-Version: */
	if (!env->message_id && !strcmp (tmp+1,"ESSAGE-ID"))
	  env->message_id = cpystr (d);
	else if (!strcmp (tmp+1,"IME-VERSION")) {
				/* tie off at end of phrase */
	  if (t = rfc822_parse_phrase (d)) *t = '\0';
	  rfc822_skipws (&d);	/* skip whitespace */
				/* known version? */
	  if (strcmp (d,"1.0") && strcmp (d,"RFC-XXXX"))
	    MM_LOG ("Warning: message has unknown MIME version",PARSE);
	  MIMEp = T;		/* note that we are MIME */
	}
	break;
      case 'N':			/* possible Newsgroups: */
	if (!env->newsgroups && !strcmp (tmp+1,"EWSGROUPS")) {
	  t = env->newsgroups = (char *) fs_get (1 + strlen (d));
	  while (c = *d++) if (c != ' ') *t++ = c;
	  *t++ = '\0';
	}
	break;
      case 'R':			/* possible Reply-To: */
	if (!strcmp (tmp+1,"EPLY-TO"))
	  rfc822_parse_adrlist (&env->reply_to,d,host);
	else if (!env->references && !strcmp (tmp+1,"EFERENCES"))
	  env->references = cpystr (d);
	break;
      case 'S':			/* possible Subject: or Sender: */
	if (!env->subject && !strcmp (tmp+1,"UBJECT"))
	  env->subject = cpystr (d);
	else if (!strcmp (tmp+1,"ENDER"))
	  rfc822_parse_adrlist (&env->sender,d,host);
	break;
      case 'T':			/* possible To: */
	if (!strcmp (tmp+1,"O")) rfc822_parse_adrlist (&env->to,d,host);
	break;
      default:
	break;
      }
    }
  }
  fs_give ((void **) &tmp);	/* done with scratch buffer */
				/* default Sender: and Reply-To: to From: */
  if (!env->sender) env->sender = rfc822_cpy_adr (env->from);
  if (!env->reply_to) env->reply_to = rfc822_cpy_adr (env->from);
				/* now parse the body */
  if (body) rfc822_parse_content (body,bs,host,depth,flags);
}

/* Parse a message body content
 * Accepts: pointer to body structure
 *	    body string
 *	    pointer to local host name
 *	    recursion depth
 *	    source driver flags
 */

void rfc822_parse_content (BODY *body,STRING *bs,char *h,unsigned long depth,
			   unsigned long flags)
{
  char c,c1,*s,*s1;
  int f;
  unsigned long i,j,k,m;
  PARAMETER *param;
  PART *part = NIL;
  if (depth > MAXMIMEDEPTH) {	/* excessively deep recursion? */
    body->type = TYPETEXT;	/* yes, probably a malicious MIMEgram */
    MM_LOG ("Ignoring excessively deep MIME recursion",PARSE);
  }
  if (!body->subtype)		/* default subtype if still unknown */
    body->subtype = cpystr (rfc822_default_subtype (body->type));
				/* note offset and sizes */
  body->contents.offset = GETPOS (bs);
				/* note internal body size in all cases */
  body->size.bytes = body->contents.text.size = i = SIZE (bs);
  if (!(flags & DR_CRLF)) body->size.bytes = strcrlflen (bs);
  switch (body->type) {		/* see if anything else special to do */
  case TYPETEXT:		/* text content */
    if (!body->parameter) {	/* no parameters set */
      body->parameter = mail_newbody_parameter ();
      body->parameter->attribute = cpystr ("CHARSET");
      while (i--) {		/* count lines and guess charset */
	c = SNX (bs);		/* get current character */
				/* charset still unknown? */
	if (!body->parameter->value) {
	  if ((c == I2C_ESC) && (i && i--) && ((c = SNX (bs)) == I2C_MULTI) &&
	      (i && i--) && (((c = SNX (bs)) == I2CS_94x94_JIS_NEW) ||
			     (c == I2CS_94x94_JIS_OLD)))
	    body->parameter->value = cpystr ("ISO-2022-JP");
	  else if (c & 0x80) body->parameter->value = cpystr ("X-UNKNOWN");
	}
	if (c == '\n') body->size.lines++;
      }
				/* 7-bit content */
      if (!body->parameter->value) switch (body->encoding) {
      case ENC7BIT:		/* unadorned 7-bit */
      case ENC8BIT:		/* unadorned 8-bit (but 7-bit content) */
      case ENCBINARY:		/* binary (but 7-bit content( */
	body->parameter->value = cpystr ("US-ASCII");
	break;
      default:			/* QUOTED-PRINTABLE, BASE64, etc. */
	body->parameter->value = cpystr ("X-UNKNOWN");
	break;
      }
    }
				/* just count lines */
    else while (i--) if ((SNX (bs)) == '\n') body->size.lines++;
    break;

  case TYPEMESSAGE:		/* encapsulated message */
				/* encapsulated RFC-822 message? */
    if (!strcmp (body->subtype,"RFC822")) {
      body->nested.msg = mail_newmsg ();
      switch (body->encoding) {	/* make sure valid encoding */
      case ENC7BIT:		/* these are valid nested encodings */
      case ENC8BIT:
      case ENCBINARY:
	break;
      default:
	MM_LOG ("Ignoring nested encoding of message contents",PARSE);
      }
				/* hunt for blank line */
      for (c = '\012',j = 0; (i > j) && ((c != '\012') || (CHR(bs) != '\012'));
	   j++) if ((c1 = SNX (bs)) != '\015') c = c1;
      if (i > j) {		/* unless no more text */
	c1 = SNX (bs);		/* body starts here */
	j++;			/* advance count */
      }
				/* note body text offset and header size */
      body->nested.msg->header.text.size = j;
      body->nested.msg->text.text.size = body->contents.text.size - j;
      body->nested.msg->text.offset = GETPOS (bs);
      body->nested.msg->full.offset = body->nested.msg->header.offset =
	body->contents.offset;
      body->nested.msg->full.text.size = body->contents.text.size;
				/* copy header string */
      SETPOS (bs,body->contents.offset);
      s = (char *) fs_get ((size_t) j + 1);
      for (s1 = s,k = j; k--; *s1++ = SNX (bs));
      s[j] = '\0';		/* tie off string (not really necessary) */
				/* now parse the body */
      rfc822_parse_msg_full (&body->nested.msg->env,&body->nested.msg->body,s,
			     j,bs,h,depth+1,flags);
      fs_give ((void **) &s);	/* free header string */
				/* restore position */
      SETPOS (bs,body->contents.offset);
    }
				/* count number of lines */
    while (i--) if (SNX (bs) == '\n') body->size.lines++;
    break;
  case TYPEMULTIPART:		/* multiple parts */
    switch (body->encoding) {	/* make sure valid encoding */
    case ENC7BIT:		/* these are valid nested encodings */
    case ENC8BIT:
    case ENCBINARY:
      break;
    default:
      MM_LOG ("Ignoring nested encoding of multipart contents",PARSE);
    }
				/* remember if digest */
    f = !strcmp (body->subtype,"DIGEST");
				/* find cookie */
    for (s1 = NIL,param = body->parameter; param && !s1; param = param->next)
      if (!strcmp (param->attribute,"BOUNDARY")) s1 = param->value;
    if (!s1) s1 = "-";		/* yucky default */
    j = strlen (s1) + 2;	/* length of cookie and header */
    c = '\012';			/* initially at beginning of line */

    while (i > j) {		/* examine data */
      if (m = GETPOS (bs)) m--;	/* get position in front of character */
      switch (c) {		/* examine each line */