mime: use percent-escaping for multipart form field and file names

Until now, form field and file names where escaped using the
backslash-escaping algorithm defined for multipart mails. This commit
replaces this with the percent-escaping method for URLs.

As this may introduce incompatibilities with server-side applications, a
new libcurl option CURLOPT_MIME_OPTIONS with bitmask
CURLMIMEOPT_FORMESCAPE is introduced to revert to legacy use of
backslash-escaping. This is controlled by new cli tool option
--form-escape.

New tests and documentation are provided for this feature.

Reported by: Ryan Sleevi
Fixes #7789
Closes #7805

lib/easyoptions.c

@@ -170,6 +170,7 @@ struct curl_easyoption Curl_easyopts[] = {
   {"MAX_RECV_SPEED_LARGE", CURLOPT_MAX_RECV_SPEED_LARGE, CURLOT_OFF_T, 0},
   {"MAX_SEND_SPEED_LARGE", CURLOPT_MAX_SEND_SPEED_LARGE, CURLOT_OFF_T, 0},
   {"MIMEPOST", CURLOPT_MIMEPOST, CURLOT_OBJECT, 0},
+  {"MIME_OPTIONS", CURLOPT_MIME_OPTIONS, CURLOT_LONG, 0},
   {"NETRC", CURLOPT_NETRC, CURLOT_VALUES, 0},
   {"NETRC_FILE", CURLOPT_NETRC_FILE, CURLOT_STRING, 0},
   {"NEW_DIRECTORY_PERMS", CURLOPT_NEW_DIRECTORY_PERMS, CURLOT_LONG, 0},
@@ -359,6 +360,6 @@ struct curl_easyoption Curl_easyopts[] = {
  */
 int Curl_easyopts_check(void)
 {
-  return ((CURLOPT_LASTENTRY%10000) != (314 + 1));
+  return ((CURLOPT_LASTENTRY%10000) != (315 + 1));
 }
 #endif

lib/mime.c

@@ -40,6 +40,7 @@
 #include "rand.h"
 #include "slist.h"
 #include "strcase.h"
+#include "dynbuf.h"
 /* The last 3 #include files should be in this order */
 #include "curl_printf.h"
 #include "curl_memory.h"
@@ -279,29 +280,52 @@ static void mimesetstate(struct mime_state *state,
 
 
 /* Escape header string into allocated memory. */
-static char *escape_string(const char *src)
+static char *escape_string(struct Curl_easy *data,
+                           const char *src, enum mimestrategy strategy)
 {
-  size_t bytecount = 0;
-  size_t i;
-  char *dst;
+  CURLcode result;
+  struct dynbuf db;
+  const char * const *table;
+  const char * const *p;
+  /* replace first character by rest of string. */
+  static const char * const mimetable[] = {
+    "\\\\\\",
+    "\"\\\"",
+    NULL
+  };
+  /* WHATWG HTML living standard 4.10.21.8 2 specifies:
+     For field names and filenames for file fields, the result of the
+     encoding in the previous bullet point must be escaped by replacing
+     any 0x0A (LF) bytes with the byte sequence `%0A`, 0x0D (CR) with `%0D`
+     and 0x22 (") with `%22`.
+     The user agent must not perform any other escapes. */
+  static const char * const formtable[] = {
+    "\"%22",
+    "\r%0D",
+    "\n%0A",
+    NULL
+  };
 
-  for(i = 0; src[i]; i++)
-    if(src[i] == '"' || src[i] == '\\')
-      bytecount++;
+  table = formtable;
+  /* data can be NULL when this function is called indirectly from
+     curl_formget(). */
+  if(strategy == MIMESTRATEGY_MAIL ||
+     (data && (data->set.mime_options & CURLMIMEOPT_FORMESCAPE)))
+    table = mimetable;
 
-  bytecount += i;
-  dst = malloc(bytecount + 1);
-  if(!dst)
-    return NULL;
+  Curl_dyn_init(&db, CURL_MAX_INPUT_LENGTH);
 
-  for(i = 0; *src; src++) {
-    if(*src == '"' || *src == '\\')
-      dst[i++] = '\\';
-    dst[i++] = *src;
+  for(result = Curl_dyn_add(&db, ""); !result && *src; src++) {
+    for(p = table; *p && **p != *src; p++)
+      ;
+
+    if(*p)
+      result = Curl_dyn_add(&db, *p + 1);
+    else
+      result = Curl_dyn_addn(&db, src, 1);
   }
 
-  dst[i] = '\0';
-  return dst;
+  return Curl_dyn_ptr(&db);
 }
 
 /* Check if header matches. */
@@ -1866,12 +1890,12 @@ CURLcode Curl_mime_prepare_headers(curl_mimepart *part,
       char *filename = NULL;
 
       if(part->name) {
-        name = escape_string(part->name);
+        name = escape_string(part->easy, part->name, strategy);
         if(!name)
           ret = CURLE_OUT_OF_MEMORY;
       }
       if(!ret && part->filename) {
-        filename = escape_string(part->filename);
+        filename = escape_string(part->easy, part->filename, strategy);
         if(!filename)
           ret = CURLE_OUT_OF_MEMORY;
       }

lib/setopt.c

@@ -2609,6 +2609,13 @@ CURLcode Curl_vsetopt(struct Curl_easy *data, CURLoption option, va_list param)
     break;
 #endif
 
+#if (!defined(CURL_DISABLE_HTTP) && !defined(CURL_DISABLE_MIME)) || \
+  !defined(CURL_DISABLE_SMTP) || !defined(CURL_DISABLE_IMAP)
+  case CURLOPT_MIME_OPTIONS:
+    data->set.mime_options = va_arg(param, long);
+    break;
+#endif
+
   case CURLOPT_SASL_AUTHZID:
     /* Authorisation identity (identity to act as) */
     result = Curl_setstropt(&data->set.str[STRING_SASL_AUTHZID],

lib/urldata.h

@@ -1749,6 +1749,7 @@ struct UserDefined {
   unsigned int scope_id;  /* Scope id for IPv6 */
   long allowed_protocols;
   long redir_protocols;
+  long mime_options;      /* Mime option flags. */
   struct curl_slist *mail_rcpt; /* linked list of mail recipients */
   /* Common RTSP header options */
   Curl_RtspReq rtspreq; /* RTSP request type */