merge extra MIME parsing

[disorder] / lib / mime.c

diff --git a/lib/mime.c b/lib/mime.c
index 03aa94b9a03b3b3b2040aca4f7a1d78ca7999286..5562c7e5b8976d805cc76c5265335b2e42cf1d6a 100644 (file)
--- a/lib/mime.c
+++ b/lib/mime.c
@@ -1,6 +1,6 @@
 /*
  * This file is part of DisOrder
- * Copyright (C) 2005 Richard Kettlewell
+ * Copyright (C) 2005, 2007 Richard Kettlewell
  *
  * 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
@@ -17,7 +17,9 @@
  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
  * USA
  */
-
+/** @file lib/mime.c
+ * @brief Support for MIME and allied protocols
+ */
 
 #include <config.h>
 #include "types.h"
@@ -31,7 +33,9 @@
 #include "mime.h"
 #include "vector.h"
 #include "hex.h"
+#include "log.h"
 
+/** @brief Match whitespace characters */
 static int whitespace(int c) {
   switch(c) {
   case ' ':
@@ -44,6 +48,7 @@ static int whitespace(int c) {
   }
 }
 
+/** @brief Match RFC2045 tspecial characters */
 static int tspecial(int c) {
   switch(c) {
   case '(':
@@ -67,7 +72,43 @@ static int tspecial(int c) {
   }
 }
 
-static const char *skipwhite(const char *s) {
+/** @brief Mathc RFC2616 seprator characters */
+static int http_separator(int c) {
+  switch(c) {
+  case '(':
+  case ')':
+  case '<':
+  case '>':
+  case '@':
+  case ',':
+  case ';':
+  case ':':
+  case '\\':
+  case '"':
+  case '/':
+  case '[':
+  case ']':
+  case '?':
+  case '=':
+  case '{':
+  case '}':
+  case ' ':
+  case '\t':
+    return 1;
+  default:
+    return 0;
+  }
+}
+
+/** @brief Match CRLF */
+static int iscrlf(const char *ptr) {
+  return ptr[0] == '\r' && ptr[1] == '\n';
+}
+
+/** @brief Skip whitespace
+ * @param rfc822_comments If true, skip RFC822 nested comments
+ */
+static const char *skipwhite(const char *s, int rfc822_comments) {
   int c, depth;
   
   for(;;) {
@@ -79,6 +120,8 @@ static const char *skipwhite(const char *s) {
       ++s;
       break;
     case '(':
+      if(!rfc822_comments)
+       return s;
       ++s;
       depth = 1;
       while(*s && depth) {
@@ -100,66 +143,103 @@ static const char *skipwhite(const char *s) {
   }
 }
 
-static const char *parsestring(const char *s, char **valuep) {
-  struct dynstr value;
+/** @brief Test for a word character
+ * @param c Character to test
+ * @param special tspecial() (MIME/RFC2405) or http_separator() (HTTP/RFC2616)
+ * @return 1 if @p c is a word character, else 0
+ */
+static int iswordchar(int c, int (*special)(int)) {
+  return !(c <= ' ' || c > '~' || special(c));
+}
+
+/** @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)
+ * @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)) {
+  struct dynstr value[1];
   int c;
 
-  dynstr_init(&value);
-  ++s;
-  while((c = *s++) != '"') {
-    switch(c) {
-    case '\\':
-      if(!(c = *s++)) return 0;
-    default:
-      dynstr_append(&value, c);
-      break;
+  dynstr_init(value);
+  if(*s == '"') {
+    ++s;
+    while((c = *s++) != '"') {
+      switch(c) {
+      case '\\':
+       if(!(c = *s++)) return 0;
+      default:
+       dynstr_append(value, c);
+       break;
+      }
     }
+    if(!c) return 0;
+  } else {
+    if(!iswordchar((unsigned char)*s, special))
+      return NULL;
+    dynstr_init(value);
+    while(iswordchar((unsigned char)*s, special))
+      dynstr_append(value, *s++);
   }
-  if(!c) return 0;
-  dynstr_terminate(&value);
-  *valuep = value.vec;
+  dynstr_terminate(value);
+  *valuep = value->vec;
   return s;
 }
 
+/** @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)
+ * @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);
+}
+
+/** @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 parameternvaluep Wher to store parameter value
+ * @return 0 on success, non-0 on error
+ */
 int mime_content_type(const char *s,
                      char **typep,
                      char **parameternamep,
                      char **parametervaluep) {
-  struct dynstr type, parametername, parametervalue;
+  struct dynstr type, parametername;
 
   dynstr_init(&type);
-  if(!(s = skipwhite(s))) return -1;
+  if(!(s = skipwhite(s, 1))) return -1;
   if(!*s) return -1;
   while(*s && !tspecial(*s) && !whitespace(*s))
     dynstr_append(&type, tolower((unsigned char)*s++));
-  if(!(s = skipwhite(s))) return -1;
+  if(!(s = skipwhite(s, 1))) return -1;
   if(*s++ != '/') return -1;
   dynstr_append(&type, '/');
-  if(!(s = skipwhite(s))) return -1;
+  if(!(s = skipwhite(s, 1))) return -1;
   while(*s && !tspecial(*s) && !whitespace(*s))
     dynstr_append(&type, tolower((unsigned char)*s++));
-  if(!(s = skipwhite(s))) return -1;
+  if(!(s = skipwhite(s, 1))) return -1;
 
   if(*s == ';') {
     dynstr_init(&parametername);
     ++s;
-    if(!(s = skipwhite(s))) return -1;
+    if(!(s = skipwhite(s, 1))) return -1;
     if(!*s) return -1;
     while(*s && !tspecial(*s) && !whitespace(*s))
       dynstr_append(&parametername, tolower((unsigned char)*s++));
-    if(!(s = skipwhite(s))) return -1;
+    if(!(s = skipwhite(s, 1))) return -1;
     if(*s++ != '=') return -1;
-    if(!(s = skipwhite(s))) return -1;
-    if(*s == '"') {
-      if(!(s = parsestring(s, parametervaluep))) return -1;
-    } else {
-      dynstr_init(&parametervalue);
-      while(*s && !tspecial(*s) && !whitespace(*s))
-       dynstr_append(&parametervalue, *s++);
-      dynstr_terminate(&parametervalue);
-      *parametervaluep = parametervalue.vec;
-    }
-    if(!(s = skipwhite(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;
     dynstr_terminate(&parametername);
     *parameternamep = parametername.vec;
   } else
@@ -169,10 +249,12 @@ int mime_content_type(const char *s,
   return 0;
 }
 
-static int iscrlf(const char *ptr) {
-  return ptr[0] == '\r' && ptr[1] == '\n';
-}
-
+/** @brief Parse a MIME message
+ * @param s Start of message
+ * @param callback Called for each header field
+ * @param u Passed to callback
+ * @return Pointer to decoded body (might be in original string)
+ */
 const char *mime_parse(const char *s,
                       int (*callback)(const char *name, const char *value,
                                       void *u),
@@ -185,7 +267,7 @@ const char *mime_parse(const char *s,
     dynstr_init(&value);
     while(*s && !tspecial(*s) && !whitespace(*s))
       dynstr_append(&name, tolower((unsigned char)*s++));
-    if(!(s = skipwhite(s))) return 0;
+    if(!(s = skipwhite(s, 1))) return 0;
     if(*s != ':') return 0;
     ++s;
     while(*s && !(*s == '\n' && !(s[1] == ' ' || s[1] == '\t')))
@@ -227,6 +309,13 @@ static int isfinal(const char *ptr, const char *boundary, size_t bl) {
          && (iscrlf(ptr + bl + 4) || *(ptr + bl + 4) == 0));
 }
 
+/** @brief Parse a multipart MIME body
+ * @param s Start of message
+ * @param callback CAllback for each part
+ * @param boundary Boundary string
+ * @param u Passed to callback
+ * @return 0 on success, non-0 on error
+ */
 int mime_multipart(const char *s,
                   int (*callback)(const char *s, void *u),
                   const char *boundary,
@@ -255,39 +344,38 @@ int mime_multipart(const char *s,
   return 0;
 }
 
+/** @brief Parse an RFC2388-style content-disposition field
+ * @param s Start of field
+ * @param typep Where to store type
+ * @param parameternamep Where to store parameter name
+ * @param parameternvaluep Wher to store parameter value
+ * @return 0 on success, non-0 on error
+ */
 int mime_rfc2388_content_disposition(const char *s,
                                     char **dispositionp,
                                     char **parameternamep,
                                     char **parametervaluep) {
-  struct dynstr disposition, parametername, parametervalue;
+  struct dynstr disposition, parametername;
 
   dynstr_init(&disposition);
-  if(!(s = skipwhite(s))) return -1;
+  if(!(s = skipwhite(s, 1))) return -1;
   if(!*s) return -1;
   while(*s && !tspecial(*s) && !whitespace(*s))
     dynstr_append(&disposition, tolower((unsigned char)*s++));
-  if(!(s = skipwhite(s))) return -1;
+  if(!(s = skipwhite(s, 1))) return -1;
 
   if(*s == ';') {
     dynstr_init(&parametername);
     ++s;
-    if(!(s = skipwhite(s))) return -1;
+    if(!(s = skipwhite(s, 1))) return -1;
     if(!*s) return -1;
     while(*s && !tspecial(*s) && !whitespace(*s))
       dynstr_append(&parametername, tolower((unsigned char)*s++));
-    if(!(s = skipwhite(s))) return -1;
+    if(!(s = skipwhite(s, 1))) return -1;
     if(*s++ != '=') return -1;
-    if(!(s = skipwhite(s))) return -1;
-    if(*s == '"') {
-      if(!(s = parsestring(s, parametervaluep))) return -1;
-    } else {
-      dynstr_init(&parametervalue);
-      while(*s && !tspecial(*s) && !whitespace(*s))
-       dynstr_append(&parametervalue, *s++);
-      dynstr_terminate(&parametervalue);
-      *parametervaluep = parametervalue.vec;
-    }
-    if(!(s = skipwhite(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;
     dynstr_terminate(&parametername);
     *parameternamep = parametername.vec;
   } else
@@ -297,6 +385,10 @@ int mime_rfc2388_content_disposition(const char *s,
   return 0;
 }
 
+/** @brief Convert MIME quoted-printable
+ * @param s Quoted-printable data
+ * @return Decoded data
+ */
 char *mime_qp(const char *s) {
   struct dynstr d;
   int c, a, b;
@@ -339,6 +431,10 @@ char *mime_qp(const char *s) {
   return d.vec;
 }
 
+/** @brief Convert MIME base64
+ * @param s base64 data
+ * @return Decoded data
+ */
 char *mime_base64(const char *s) {
   struct dynstr d;
   const char *t;
@@ -370,6 +466,82 @@ char *mime_base64(const char *s) {
   return d.vec;
 }
 
+/** @brief Parse a RFC2109 Cookie: header
+ * @param s Header field value
+ * @param cd Where to store result
+ * @return 0 on success, non-0 on error
+ */
+int parse_cookie(const char *s,
+                struct cookiedata *cd) {
+  char *n = 0, *v = 0;
+
+  memset(cd, 0, sizeof *cd);
+  s = skipwhite(s, 0);
+  while(*s) {
+    /* Skip separators */
+    if(*s == ';' || *s == ',') {
+      ++s;
+      s = skipwhite(s, 0);
+      continue;
+    }
+    if(!(s = parsetoken(s, &n, http_separator))) return -1;
+    s = skipwhite(s, 0);
+    if(*s++ != '=') return -1;
+    s = skipwhite(s, 0);
+    if(!(s = parseword(s, &v, http_separator))) return -1;
+    if(n[0] == '$') {
+      /* Some bit of meta-information */
+      if(!strcmp(n, "$Version"))
+       cd->version = v;
+      else if(!strcmp(n, "$Path")) {
+       if(cd->ncookies > 0 && cd->cookies[cd->ncookies-1].path == 0)
+         cd->cookies[cd->ncookies-1].path = v;
+       else {
+         error(0, "redundant $Path in Cookie: header");
+         return -1;
+       }
+      } else if(!strcmp(n, "$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");
+         return -1;
+       }
+      }
+    } else {
+      /* It's a new cookie */
+      cd->cookies = xrealloc(cd->cookies,
+                            (cd->ncookies + 1) * sizeof (struct cookie));
+      cd->cookies[cd->ncookies].name = n;
+      cd->cookies[cd->ncookies].value = v;
+      cd->cookies[cd->ncookies].path = 0;
+      cd->cookies[cd->ncookies].domain = 0;
+      ++cd->ncookies;
+    }
+    s = skipwhite(s, 0);
+    if(*s && (*s != ',' && *s != ';')) {
+      error(0, "missing separator in Cookie: header");
+      return -1;
+    }
+  }
+  return 0;
+}
+
+/** @brief Find a named cookie
+ * @param cd Parse cookie data
+ * @param name Name of cookie
+ * @return Cookie structure or NULL if not found
+ */
+const struct cookie *find_cookie(const struct cookiedata *cd,
+                                const char *name) {
+  int n;
+
+  for(n = 0; n < cd->ncookies; ++n)
+    if(!strcmp(cd->cookies[n].name, name))
+      return &cd->cookies[n];
+  return 0;
+}
+
 /*
 Local Variables:
 c-basic-offset:2