X-Git-Url: http://www.chiark.greenend.org.uk/ucgi/~mdw/git/disorder/blobdiff_plain/39d4aa6b1c612305bf06760529e2a1532b9818a5..5891b0a8916232a54a4856e186b1d21a44b38a48:/lib/mime.c

diff --git a/lib/mime.c b/lib/mime.c
index c8ffe31..a97c4d1 100644
--- a/lib/mime.c
+++ b/lib/mime.c
@@ -27,11 +27,14 @@
 #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"
 
 /** @brief Match whitespace characters */
 static int whitespace(int c) {
@@ -70,7 +73,7 @@ static int tspecial(int c) {
   }
 }
 
-/** @brief Mathc RFC2616 seprator characters */
+/** @brief Match RFC2616 seprator characters */
 static int http_separator(int c) {
   switch(c) {
   case '(':
@@ -206,6 +209,8 @@ static const char *parsetoken(const char *s, char **valuep,
  * @param parameternamep Where to store parameter name
  * @param parameternvaluep Wher to store parameter value
  * @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,
@@ -251,7 +256,12 @@ int mime_content_type(const char *s,
  * @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)
+ * @return Pointer to decoded body (might be in original string), or NULL on error
+ *
+ * This does an RFC 822-style parse and honors Content-Transfer-Encoding as
+ * described in <a href="http://tools.ietf.org/html/rfc2045#section-6">RFC 2045
+ * s6</a>.  @p callback is called for each header field encountered, in order,
+ * with ASCII characters in the header name forced to lower case.
  */
 const char *mime_parse(const char *s,
 		       int (*callback)(const char *name, const char *value,
@@ -282,12 +292,13 @@ const char *mime_parse(const char *s,
   }
   if(*s) s += 2;
   if(cte) {
-    if(!strcmp(cte, "base64")) return mime_base64(s);
+    if(!strcmp(cte, "base64")) return mime_base64(s, 0);
     if(!strcmp(cte, "quoted-printable")) return mime_qp(s);
   }
   return s;
 }
 
+/** @brief Match the boundary string */
 static int isboundary(const char *ptr, const char *boundary, size_t bl) {
   return (ptr[0] == '-'
 	  && ptr[1] == '-'
@@ -295,24 +306,29 @@ static int isboundary(const char *ptr, const char *boundary, size_t bl) {
 	  && (iscrlf(ptr + bl + 2)
 	      || (ptr[bl + 2] == '-'
 		  && ptr[bl + 3] == '-'
-		  && iscrlf(ptr + bl + 4))));
+		  && (iscrlf(ptr + bl + 4) || *(ptr + bl + 4) == 0))));
 }
 
+/** @brief Match the final boundary string */
 static int isfinal(const char *ptr, const char *boundary, size_t bl) {
   return (ptr[0] == '-'
 	  && ptr[1] == '-'
 	  && !strncmp(ptr + 2, boundary, bl)
 	  && ptr[bl + 2] == '-'
 	  && ptr[bl + 3] == '-'
-	  && iscrlf(ptr + bl + 4));
+	  && (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 callback Callback for each part
  * @param boundary Boundary string
  * @param u Passed to callback
  * @return 0 on success, non-0 on error
+ * 
+ * See <a href="http://tools.ietf.org/html/rfc2046#section-5.1">RFC 2046
+ * s5.1</a>.  @p callback is called for each part (not yet decoded in any way)
+ * in succession; you should probably call mime_parse() for each part.
  */
 int mime_multipart(const char *s,
 		   int (*callback)(const char *s, void *u),
@@ -322,12 +338,16 @@ int mime_multipart(const char *s,
   const char *start, *e;
   int ret;
 
-  if(!isboundary(s, boundary, bl)) return -1;
+  /* We must start with a boundary string */
+  if(!isboundary(s, boundary, bl))
+    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"))) return -1;
+      if(!(e = strstr(s, "\r\n")))
+	return -1;
       s = e + 2;
     }
     if((ret = callback(xstrndup(start,
@@ -344,6 +364,9 @@ int mime_multipart(const char *s,
  * @param parameternamep Where to store parameter name
  * @param parameternvaluep Wher 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>
+ * and <a href="http://tools.ietf.org/html/rfc2183">RFC 2183</a>.
  */
 int mime_rfc2388_content_disposition(const char *s,
 				     char **dispositionp,
@@ -382,6 +405,9 @@ int mime_rfc2388_content_disposition(const char *s,
 /** @brief Convert MIME quoted-printable
  * @param s Quoted-printable data
  * @return Decoded data
+ *
+ * See <a href="http://tools.ietf.org/html/rfc2045#section-6.7">RFC 2045
+ * s6.7</a>.
  */
 char *mime_qp(const char *s) {
   struct dynstr d;
@@ -425,41 +451,6 @@ 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;
-  int b[4], n, c;
-  static const char table[] =
-    "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
-
-  dynstr_init(&d);
-  n = 0;
-  while((c = (unsigned char)*s++)) {
-    if((t = strchr(table, c))) {
-      b[n++] = t - table;
-      if(n == 4) {
-	dynstr_append(&d, (b[0] << 2) + (b[1] >> 4));
-	dynstr_append(&d, (b[1] << 4) + (b[2] >> 2));
-	dynstr_append(&d, (b[2] << 6) + b[3]);
-	n = 0;
-      }
-    } else if(c == '=') {
-      if(n >= 2) {
-	dynstr_append(&d, (b[0] << 2) + (b[1] >> 4));
-	if(n == 3)
-	  dynstr_append(&d, (b[1] << 4) + (b[2] >> 2));
-      }
-      break;
-    }
-  }
-  dynstr_terminate(&d);
-  return d.vec;
-}
-
 /** @brief Parse a RFC2109 Cookie: header
  * @param s Header field value
  * @param cd Where to store result