X-Git-Url: https://www.chiark.greenend.org.uk/ucgi/~mdw/git/disorder/blobdiff_plain/82c01b317cd1892f4376c68be88a74f971493428..8ab2aa9fd51a89e06d92a4f7c3792aaa4a08cc71:/lib/mime.c

diff --git a/lib/mime.c b/lib/mime.c
index 8776647..99dac39 100644
--- a/lib/mime.c
+++ b/lib/mime.c
@@ -1,40 +1,36 @@
 /*
  * This file is part of DisOrder
- * Copyright (C) 2005, 2007 Richard Kettlewell
+ * Copyright (C) 2005, 2007-10, 2013 Richard Kettlewell
  *
- * This program is free software; you can redistribute it and/or modify
+ * This program is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
- * the Free Software Foundation; either version 2 of the License, or
+ * the Free Software Foundation, either version 3 of the License, or
  * (at your option) any later version.
- *
- * This program is distributed in the hope that it will be useful, but
- * WITHOUT ANY WARRANTY; without even the implied warranty of
- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
- * General Public License for more details.
- *
+ * 
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ * 
  * You should have received a copy of the GNU General Public License
- * along with this program; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
- * USA
+ * along with this program.  If not, see <http://www.gnu.org/licenses/>.
  */
 /** @file lib/mime.c
  * @brief Support for MIME and allied protocols
  */
 
-#include <config.h>
-#include "types.h"
+#include "common.h"
 
-#include <string.h>
 #include <ctype.h>
 
-#include <stdio.h>
-
 #include "mem.h"
 #include "mime.h"
 #include "vector.h"
 #include "hex.h"
 #include "log.h"
 #include "base64.h"
+#include "kvp.h"
+#include "printf.h"
 
 /** @brief Match whitespace characters */
 static int whitespace(int c) {
@@ -50,7 +46,7 @@ static int whitespace(int c) {
 }
 
 /** @brief Match RFC2045 tspecial characters */
-static int tspecial(int c) {
+int mime_tspecial(int c) {
   switch(c) {
   case '(':
   case ')':
@@ -74,7 +70,7 @@ static int tspecial(int c) {
 }
 
 /** @brief Match RFC2616 separator characters */
-static int http_separator(int c) {
+int mime_http_separator(int c) {
   switch(c) {
   case '(':
   case ')':
@@ -115,7 +111,7 @@ static const char *skipwhite(const char *s, int rfc822_comments) {
   int c, depth;
   
   for(;;) {
-    switch(c = *s) {
+    switch(*s) {
     case ' ':
     case '\t':
     case '\r':
@@ -133,12 +129,14 @@ static const char *skipwhite(const char *s, int rfc822_comments) {
 	case '(': ++depth; break;
 	case ')': --depth; break;
 	case '\\':
-	  if(!*s) return 0;
+	  if(!*s)
+	    return 0;
 	  ++s;
 	  break;
 	}
       }
-      if(depth) return 0;
+      if(depth)
+	return 0;
       break;
     default:
       return s;
@@ -148,7 +146,7 @@ static const char *skipwhite(const char *s, int rfc822_comments) {
 
 /** @brief Test for a word character
  * @param c Character to test
- * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
+ * @param special mime_tspecial() (MIME/RFC2405) or mime_http_separator() (HTTP/RFC2616)
  * @return 1 if @p c is a word character, else 0
  */
 static int iswordchar(int c, int (*special)(int)) {
@@ -158,13 +156,13 @@ static int iswordchar(int c, int (*special)(int)) {
 /** @brief Parse an RFC1521/RFC2616 word
  * @param s Pointer to start of word
  * @param valuep Where to store value
- * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
+ * @param special mime_tspecial() (MIME/RFC2405) or mime_http_separator() (HTTP/RFC2616)
  * @return Pointer just after end of word or NULL if there's no word
  *
  * A word is a token or a quoted-string.
  */
-static const char *parseword(const char *s, char **valuep,
-			     int (*special)(int)) {
+const char *mime_parse_word(const char *s, char **valuep,
+			    int (*special)(int)) {
   struct dynstr value[1];
   int c;
 
@@ -174,13 +172,16 @@ static const char *parseword(const char *s, char **valuep,
     while((c = *s++) != '"') {
       switch(c) {
       case '\\':
-	if(!(c = *s++)) return 0;
+	if(!(c = *s++))
+	  return 0;
+	/* else fall through ... */
       default:
 	dynstr_append(value, c);
 	break;
       }
     }
-    if(!c) return 0;
+    if(!c)
+      return 0;
   } else {
     if(!iswordchar((unsigned char)*s, special))
       return NULL;
@@ -196,61 +197,75 @@ static const char *parseword(const char *s, char **valuep,
 /** @brief Parse an RFC1521/RFC2616 token
  * @param s Pointer to start of token
  * @param valuep Where to store value
- * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
+ * @param special mime_tspecial() (MIME/RFC2405) or mime_http_separator() (HTTP/RFC2616)
  * @return Pointer just after end of token or NULL if there's no token
  */
 static const char *parsetoken(const char *s, char **valuep,
 			      int (*special)(int)) {
-  if(*s == '"') return 0;
-  return parseword(s, valuep, special);
+  if(*s == '"')
+    return 0;
+  return mime_parse_word(s, valuep, special);
 }
 
 /** @brief Parse a MIME content-type field
  * @param s Start of field
  * @param typep Where to store type
- * @param parameternamep Where to store parameter name
- * @param parametervaluep Wher to store parameter value
+ * @param parametersp Where to store parameter list
  * @return 0 on success, non-0 on error
  *
  * See <a href="http://tools.ietf.org/html/rfc2045#section-5">RFC 2045 s5</a>.
  */
 int mime_content_type(const char *s,
 		      char **typep,
-		      char **parameternamep,
-		      char **parametervaluep) {
+		      struct kvp **parametersp) {
   struct dynstr type, parametername;
+  struct kvp *parameters = 0;
+  char *parametervalue;
 
   dynstr_init(&type);
-  if(!(s = skipwhite(s, 1))) return -1;
-  if(!*s) return -1;
-  while(*s && !tspecial(*s) && !whitespace(*s))
+  if(!(s = skipwhite(s, 1)))
+    return -1;
+  if(!*s)
+    return -1;
+  while(*s && !mime_tspecial(*s) && !whitespace(*s))
     dynstr_append(&type, tolower((unsigned char)*s++));
-  if(!(s = skipwhite(s, 1))) return -1;
-  if(*s++ != '/') return -1;
+  if(!(s = skipwhite(s, 1)))
+    return -1;
+  if(*s++ != '/')
+    return -1;
   dynstr_append(&type, '/');
-  if(!(s = skipwhite(s, 1))) return -1;
-  while(*s && !tspecial(*s) && !whitespace(*s))
+  if(!(s = skipwhite(s, 1)))
+    return -1;
+  while(*s && !mime_tspecial(*s) && !whitespace(*s))
     dynstr_append(&type, tolower((unsigned char)*s++));
-  if(!(s = skipwhite(s, 1))) return -1;
+  if(!(s = skipwhite(s, 1)))
+    return -1;
 
-  if(*s == ';') {
+  while(*s == ';') {
     dynstr_init(&parametername);
     ++s;
-    if(!(s = skipwhite(s, 1))) return -1;
-    if(!*s) return -1;
-    while(*s && !tspecial(*s) && !whitespace(*s))
+    if(!(s = skipwhite(s, 1)))
+      return -1;
+    if(!*s)
+      return -1;
+    while(*s && !mime_tspecial(*s) && !whitespace(*s))
       dynstr_append(&parametername, tolower((unsigned char)*s++));
-    if(!(s = skipwhite(s, 1))) return -1;
-    if(*s++ != '=') return -1;
-    if(!(s = skipwhite(s, 1))) return -1;
-    if(!(s = parseword(s, parametervaluep, tspecial))) return -1;
-    if(!(s = skipwhite(s, 1))) return -1;
+    if(!(s = skipwhite(s, 1)))
+      return -1;
+    if(*s++ != '=')
+      return -1;
+    if(!(s = skipwhite(s, 1)))
+      return -1;
+    if(!(s = mime_parse_word(s, &parametervalue, mime_tspecial)))
+      return -1;
+    if(!(s = skipwhite(s, 1)))
+      return -1;
     dynstr_terminate(&parametername);
-    *parameternamep = parametername.vec;
-  } else
-    *parametervaluep = *parameternamep = 0;
+    kvp_set(&parameters, parametername.vec, parametervalue);
+  }
   dynstr_terminate(&type);
   *typep = type.vec;
+  *parametersp = parameters;
   return 0;
 }
 
@@ -275,14 +290,27 @@ const char *mime_parse(const char *s,
   while(*s && !iscrlf(s)) {
     dynstr_init(&name);
     dynstr_init(&value);
-    while(*s && !tspecial(*s) && !whitespace(*s))
+    while(*s && !mime_tspecial(*s) && !whitespace(*s))
       dynstr_append(&name, tolower((unsigned char)*s++));
-    if(!(s = skipwhite(s, 1))) return 0;
-    if(*s != ':') return 0;
+    if(!(s = skipwhite(s, 1)))
+      return 0;
+    if(*s != ':')
+      return 0;
     ++s;
-    while(*s && !(*s == '\n' && !(s[1] == ' ' || s[1] == '\t')))
-      dynstr_append(&value, *s++);
-    if(*s) ++s;
+    while(*s && !(*s == '\n' && !(s[1] == ' ' || s[1] == '\t'))) {
+      const int c = *s++;
+      /* Strip leading whitespace */
+      if(value.nvec || !(c == ' ' || c == '\t' || c == '\n' || c == '\r'))
+	dynstr_append(&value, c);
+    }
+    /* Strip trailing whitespace */
+    while(value.nvec > 0 && (value.vec[value.nvec - 1] == ' '
+			     || value.vec[value.nvec - 1] == '\t'
+			     || value.vec[value.nvec - 1] == '\n'
+			     || value.vec[value.nvec - 1] == '\r'))
+      --value.nvec;
+    if(*s)
+      ++s;
     dynstr_terminate(&name);
     dynstr_terminate(&value);
     if(!strcmp(name.vec, "content-transfer-encoding")) {
@@ -290,12 +318,20 @@ const char *mime_parse(const char *s,
       for(p = cte; *p; p++)
 	*p = tolower((unsigned char)*p);
     }
-    if(callback(name.vec, value.vec, u)) return 0;
+    if(callback(name.vec, value.vec, u))
+      return 0;
   }
-  if(*s) s += 2;
+  if(*s)
+    s += 2;
   if(cte) {
-    if(!strcmp(cte, "base64")) return mime_base64(s, 0);
-    if(!strcmp(cte, "quoted-printable")) return mime_qp(s);
+    if(!strcmp(cte, "base64"))
+      return mime_base64(s, 0);
+    if(!strcmp(cte, "quoted-printable"))
+      return mime_qp(s);
+    if(!strcmp(cte, "7bit") || !strcmp(cte, "8bit"))
+      return s;
+    disorder_error(0, "unknown content-transfer-encoding '%s'", cte);
+    return 0;
   }
   return s;
 }
@@ -341,15 +377,19 @@ int mime_multipart(const char *s,
   int ret;
 
   /* We must start with a boundary string */
-  if(!isboundary(s, boundary, bl))
+  if(!isboundary(s, boundary, bl)) {
+    disorder_error(0, "mime_multipart: first line is not the boundary string");
     return -1;
+  }
   /* Keep going until we hit a final boundary */
   while(!isfinal(s, boundary, bl)) {
     s = strstr(s, "\r\n") + 2;
     start = s;
     while(!isboundary(s, boundary, bl)) {
-      if(!(e = strstr(s, "\r\n")))
+      if(!(e = strstr(s, "\r\n"))) {
+	disorder_error(0, "mime_multipart: line does not end CRLF");
 	return -1;
+      }
       s = e + 2;
     }
     if((ret = callback(xstrndup(start,
@@ -364,7 +404,7 @@ int mime_multipart(const char *s,
  * @param s Start of field
  * @param dispositionp Where to store disposition
  * @param parameternamep Where to store parameter name
- * @param parametervaluep Wher to store parameter value
+ * @param parametervaluep Where to store parameter value
  * @return 0 on success, non-0 on error
  *
  * See <a href="http://tools.ietf.org/html/rfc2388#section-3">RFC 2388 s3</a>
@@ -377,24 +417,34 @@ int mime_rfc2388_content_disposition(const char *s,
   struct dynstr disposition, parametername;
 
   dynstr_init(&disposition);
-  if(!(s = skipwhite(s, 1))) return -1;
-  if(!*s) return -1;
-  while(*s && !tspecial(*s) && !whitespace(*s))
+  if(!(s = skipwhite(s, 1)))
+    return -1;
+  if(!*s)
+    return -1;
+  while(*s && !mime_tspecial(*s) && !whitespace(*s))
     dynstr_append(&disposition, tolower((unsigned char)*s++));
-  if(!(s = skipwhite(s, 1))) return -1;
+  if(!(s = skipwhite(s, 1)))
+    return -1;
 
   if(*s == ';') {
     dynstr_init(&parametername);
     ++s;
-    if(!(s = skipwhite(s, 1))) return -1;
-    if(!*s) return -1;
-    while(*s && !tspecial(*s) && !whitespace(*s))
+    if(!(s = skipwhite(s, 1)))
+      return -1;
+    if(!*s)
+      return -1;
+    while(*s && !mime_tspecial(*s) && !whitespace(*s))
       dynstr_append(&parametername, tolower((unsigned char)*s++));
-    if(!(s = skipwhite(s, 1))) return -1;
-    if(*s++ != '=') return -1;
-    if(!(s = skipwhite(s, 1))) return -1;
-    if(!(s = parseword(s, parametervaluep, tspecial))) return -1;
-    if(!(s = skipwhite(s, 1))) return -1;
+    if(!(s = skipwhite(s, 1)))
+      return -1;
+    if(*s++ != '=')
+      return -1;
+    if(!(s = skipwhite(s, 1)))
+      return -1;
+    if(!(s = mime_parse_word(s, parametervaluep, mime_tspecial)))
+      return -1;
+    if(!(s = skipwhite(s, 1)))
+      return -1;
     dynstr_terminate(&parametername);
     *parameternamep = parametername.vec;
   } else
@@ -453,6 +503,50 @@ char *mime_qp(const char *s) {
   return d.vec;
 }
 
+/** @brief Match cookie separator characters
+ *
+ * This is a subset of the RFC2616 specials, and technically is in breach of
+ * the specification.  However rejecting (in particular) slashes is
+ * unreasonably strict and has broken at least one (admittedly somewhat
+ * obscure) browser, so we're more forgiving.
+ */
+static int cookie_separator(int c) {
+  switch(c) {
+  case '(':
+  case ')':
+  case ',':
+  case ';':
+  case '=':
+  case ' ':
+  case '"':
+  case '\t':
+    return 1;
+
+  default:
+    return 0;
+  }
+}
+
+/** @brief Match cookie value separator characters
+ *
+ * Same as cookie_separator() but allows for @c = in cookie values.
+ */
+static int cookie_value_separator(int c) {
+  switch(c) {
+  case '(':
+  case ')':
+  case ',':
+  case ';':
+  case ' ':
+  case '"':
+  case '\t':
+    return 1;
+
+  default:
+    return 0;
+  }
+}
+
 /** @brief Parse a RFC2109 Cookie: header
  * @param s Header field value
  * @param cd Where to store result
@@ -473,11 +567,20 @@ int parse_cookie(const char *s,
       s = skipwhite(s, 0);
       continue;
     }
-    if(!(s = parsetoken(s, &n, http_separator))) return -1;
+    if(!(s = parsetoken(s, &n, cookie_separator))) {
+      disorder_error(0, "parse_cookie: cannot parse attribute name");
+      return -1;
+    }
     s = skipwhite(s, 0);
-    if(*s++ != '=') return -1;
+    if(*s++ != '=') {
+      disorder_error(0, "parse_cookie: did not find expected '='");
+      return -1;
+    }
     s = skipwhite(s, 0);
-    if(!(s = parseword(s, &v, http_separator))) return -1;
+    if(!(s = mime_parse_word(s, &v, cookie_value_separator))) {
+      disorder_error(0, "parse_cookie: cannot parse value for '%s'", n);
+      return -1;
+    }
     if(n[0] == '$') {
       /* Some bit of meta-information */
       if(!strcmp(n, "$Version"))
@@ -486,14 +589,14 @@ int parse_cookie(const char *s,
 	if(cd->ncookies > 0 && cd->cookies[cd->ncookies-1].path == 0)
 	  cd->cookies[cd->ncookies-1].path = v;
 	else {
-	  error(0, "redundant $Path in Cookie: header");
+	  disorder_error(0, "redundant $Path in Cookie: header");
 	  return -1;
 	}
       } else if(!strcmp(n, "$Domain")) {
 	if(cd->ncookies > 0 && cd->cookies[cd->ncookies-1].domain == 0)
 	  cd->cookies[cd->ncookies-1].domain = v;
 	else {
-	  error(0, "redundant $Domain in Cookie: header");
+	  disorder_error(0, "redundant $Domain in Cookie: header");
 	  return -1;
 	}
       }
@@ -509,7 +612,7 @@ int parse_cookie(const char *s,
     }
     s = skipwhite(s, 0);
     if(*s && (*s != ',' && *s != ';')) {
-      error(0, "missing separator in Cookie: header");
+      disorder_error(0, "missing separator in Cookie: header");
       return -1;
     }
   }
@@ -544,7 +647,7 @@ char *quote822(const char *s, int force) {
   if(!force) {
     /* See if we need to quote */
     for(t = s; (c = (unsigned char)*t); ++t) {
-      if(tspecial(c) || http_separator(c) || whitespace(c))
+      if(mime_tspecial(c) || mime_http_separator(c) || whitespace(c))
 	break;
     }
     if(*t)
@@ -611,7 +714,7 @@ char *mime_to_qp(const char *text) {
       ++linelength;
     } else {
       /* Anything else that needs encoding */
-      snprintf(buffer, sizeof buffer, "=%02X", c);
+      byte_snprintf(buffer, sizeof buffer, "=%02X", c);
       dynstr_append_string(d, buffer);
       linelength += 3;
     }
@@ -635,7 +738,7 @@ char *mime_to_qp(const char *text) {
  * @param text Underlying UTF-8 text
  * @param charsetp Where to store charset string
  * @param encodingp Where to store encoding string
- * @return Encoded text (might be @ref text)
+ * @return Encoded text (might be @p text)
  */
 const char *mime_encode_text(const char *text,
 			     const char **charsetp,