add url percent decoding

3 files changed, 262 insertions, 43 deletions

diff --git a/fhp.c b/fhp.c
index 7b95722..eb0a7ae 100644
--- a/fhp.c
+++ b/fhp.c
@@ -3,29 +3,7 @@
 
 #define UNUSED(a) ((void) (a))
 
-//
-// rfc7230, Appendix B
-// https://tools.ietf.org/html/rfc7230
-//
-//   tchar = "!" / "#" / "$" / "%" / "&" / "'" / "*" / "+" / "-" / "." /
-//           "^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA
-//   token = 1*tchar
-//
-#define CASE_TOKEN_CHARS \
-  case '!': \
-  case '#': \
-  case '$': \
-  case '%': \
-  case '&': \
-  case '\'': \
-  case '*': \
-  case '+': \
-  case '-': \
-  case '.': \
-  case '^': \
-  case '_': \
-  case '|': \
-  case '~': \
+#define CASE_ALNUM_CHARS \
   case '0': \
   case '1': \
   case '2': \
@@ -89,6 +67,99 @@
   case 'Y': \
   case 'Z':
 
+#define CASE_DIGIT_CHARS \
+  case '0': \
+  case '1': \
+  case '2': \
+  case '3': \
+  case '4': \
+  case '5': \
+  case '6': \
+  case '7': \
+  case '8': \
+  case '9':
+
+#define CASE_HEX_ALPHA_CHARS \
+  case 'a': \
+  case 'b': \
+  case 'c': \
+  case 'd': \
+  case 'e': \
+  case 'f': \
+  case 'A': \
+  case 'B': \
+  case 'C': \
+  case 'D': \
+  case 'E': \
+  case 'F':
+
+//
+// rfc7230, Appendix B
+// https://tools.ietf.org/html/rfc7230
+//
+//   tchar = "!" / "#" / "$" / "%" / "&" / "'" / "*" / "+" / "-" / "." /
+//           "^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA
+//   token = 1*tchar
+//
+#define CASE_TOKEN_CHARS \
+  CASE_ALNUM_CHARS \
+  case '!': \
+  case '#': \
+  case '$': \
+  case '%': \
+  case '&': \
+  case '\'': \
+  case '*': \
+  case '+': \
+  case '-': \
+  case '.': \
+  case '^': \
+  case '_': \
+  case '|': \
+  case '~': \
+
+#define CASE_URL_CHARS \
+  CASE_ALNUM_CHARS \
+  case '!': \
+  case '"': \
+  case '#': \
+  case '$': \
+  case '&': \
+  case '\'': \
+  case '(': \
+  case ')': \
+  case '*': \
+  case '+': \
+  case ',': \
+  case '-': \
+  case '.': \
+  case '/': \
+  case ':': \
+  case ';': \
+  case '<': \
+  case '=': \
+  case '>': \
+  case '?': \
+  case '@': \
+  case '[': \
+  case '\\': \
+  case ']': \
+  case '^': \
+  case '_': \
+  case '`': \
+  case '{': \
+  case '|': \
+  case '}': \
+  case '~':
+
+#define CASE_VISUAL_CHARS \
+  CASE_URL_CHARS \
+  case '%':
+
+#define CASE_VERSION_CHARS \
+  CASE_TOKEN_CHARS \
+  case '/':
+
 //
 // rfc7230, Appendix B
 // https://tools.ietf.org/html/rfc7230
@@ -105,6 +176,10 @@ fhp_errors[] = {
   "callback error",
   "bad state",
   "invalid character",
+  "invalid character in HTTP method",
+  "invalid character in HTTP URL",
+  "invalid percent-encoded character in HTTP URL",
+  "invalid character in HTTP version",
   "invalid error code",
   "buffer too small",
 };
@@ -265,8 +340,7 @@ retry:
 
       break;
     default:
-      // FIXME: invalid character
-      return FHP_ERR_INVALID_CHAR;
+      return FHP_ERR_INVALID_CHAR_IN_METHOD;
     }
 
     break;
@@ -293,8 +367,7 @@ retry:
 
       break;
     default:
-      // FIXME: invalid character
-      return FHP_ERR_INVALID_CHAR;
+      return FHP_ERR_INVALID_CHAR_IN_METHOD;
     }
 
     break;
@@ -304,12 +377,14 @@ retry:
       // FIXME: do we want to allow more than one whitespace?
       // ignore
       break;
-    default:
+    CASE_URL_CHARS
       if (!fhp->cb(fhp, FHP_TOKEN_URL_START, 0, 0))
         return FHP_ERR_CB;
 
       fhp->state = FHP_STATE_URL;
       goto retry;
+    default:
+      return FHP_ERR_INVALID_CHAR_IN_URL;
     }
 
     break;
@@ -329,13 +404,55 @@ retry:
       goto retry;
 
       break;
-    default:
+    case '%':
+      // set state
+      fhp->state = FHP_STATE_URL_PERCENT;
+
+      break;
+    CASE_URL_CHARS
       // add to buffer
       if (!fhp_buf_push(fhp, FHP_TOKEN_URL_FRAGMENT, byte))
         return FHP_ERR_CB;
     }
 
     break;
+  case FHP_STATE_URL_PERCENT:
+    switch (byte) {
+    CASE_DIGIT_CHARS
+      fhp->hex = byte - '0';
+      fhp->state = FHP_STATE_URL_PERCENT_LAST;
+
+      break;
+    CASE_HEX_ALPHA_CHARS
+      fhp->hex = 10 + byte - ((byte >= 'a') ? 'a' : 'A');
+      fhp->state = FHP_STATE_URL_PERCENT_LAST;
+
+      break;
+    default:
+      return FHP_ERR_INVALID_CHAR_IN_URL_PERCENT;
+    }
+
+    break;
+  case FHP_STATE_URL_PERCENT_LAST:
+    switch (byte) {
+    CASE_DIGIT_CHARS
+      fhp->hex = (fhp->hex << 4) + (byte - '0');
+      break;
+    CASE_HEX_ALPHA_CHARS
+      fhp->hex = (fhp->hex << 4) + (10 + byte - ((byte >= 'a') ? 'a' : 'A'));
+      break;
+    default:
+      return FHP_ERR_INVALID_CHAR_IN_URL_PERCENT;
+    }
+
+    // add to buffer
+    if (!fhp_buf_push(fhp, FHP_TOKEN_URL_FRAGMENT, fhp->hex))
+      return FHP_ERR_CB;
+
+    // set state
+    fhp->state = FHP_STATE_URL;
+
+    break;
   case FHP_STATE_URL_END:
     switch (byte) {
     case ' ':
@@ -367,15 +484,14 @@ retry:
       goto retry;
 
       break;
-    case ' ':
-      // FIXME: invalid character
-      return FHP_ERR_INVALID_CHAR;
-
-      break;
-    default:
+    CASE_VERSION_CHARS
       // add to buffer
       if (!fhp_buf_push(fhp, FHP_TOKEN_VERSION_FRAGMENT, byte))
         return FHP_ERR_CB;
+
+      break;
+    default:
+      return FHP_ERR_INVALID_CHAR_IN_VERSION;
     }
 
     break;
@@ -390,7 +506,7 @@ retry:
 
       break;
     default:
-      // FIXME: invalid character
+      // never reached
       return FHP_ERR_INVALID_CHAR;
     };
 
@@ -642,3 +758,8 @@ fhp_push(
   // return success
   return FHP_OK;
 }
+
+void *
+fhp_user_data(fhp_t * const fhp) {
+  return fhp->user_data;
+}
diff --git a/include/fhp/fhp.h b/include/fhp/fhp.h
index 3b7e280..72df8a5 100644
--- a/include/fhp/fhp.h
+++ b/include/fhp/fhp.h
@@ -10,6 +10,10 @@ typedef enum {
   FHP_ERR_CB,
   FHP_ERR_BAD_STATE,
   FHP_ERR_INVALID_CHAR,
+  FHP_ERR_INVALID_CHAR_IN_METHOD,
+  FHP_ERR_INVALID_CHAR_IN_URL,
+  FHP_ERR_INVALID_CHAR_IN_URL_PERCENT,
+  FHP_ERR_INVALID_CHAR_IN_VERSION,
   FHP_ERR_INVALID_ERROR,
   FHP_ERR_BUFFER_TOO_SMALL,
   FHP_ERR_LAST
@@ -60,6 +64,8 @@ typedef enum {
   FHP_STATE_METHOD,
   FHP_STATE_METHOD_END,
   FHP_STATE_URL,
+  FHP_STATE_URL_PERCENT,
+  FHP_STATE_URL_PERCENT_LAST,
   FHP_STATE_URL_END,
   FHP_STATE_VERSION,
   FHP_STATE_VERSION_END,
@@ -80,15 +86,27 @@ typedef enum {
 #define FHP_BUF_SIZE 1024
 
 struct fhp_t_ {
+  // current parser state
   fhp_state_t state;
+
+  // user callback
   fhp_cb_t cb;
-  uint64_t ofs;
-  fhp_err_t err;
 
+  // opaque user data
   void *user_data;
 
+  // last error
+  fhp_err_t err;
+
+  // number of bytes read
+  uint64_t ofs;
+
+  // fragment data buffer
   uint8_t buf[FHP_BUF_SIZE];
   size_t buf_len;
+
+  // state for url hex decoder
+  uint32_t hex;
 };
 
 fhp_err_t
@@ -97,4 +115,7 @@ fhp_init(fhp_t * const, fhp_cb_t, void * const);
 fhp_err_t
 fhp_push(fhp_t * const, uint8_t * const, size_t);
 
+void *
+fhp_user_data(fhp_t * const);
+
 #endif /* FHP_H */
diff --git a/test.c b/test.c
index 3749ce7..792316d 100644
--- a/test.c
+++ b/test.c
@@ -23,15 +23,19 @@ static void die(
   exit(EXIT_FAILURE);
 }
 
+/**************/
+/* basic test */
+/**************/
+
 static const char *
-str_test_basic = 
+basic_str =
   "GET / HTTP/1.1\r\n"
   "Host: pablotron.org\r\n"
   "Connection: close\r\n"
   "\r\n";
 
 static bool
-test_basic_cb(
+basic_cb(
   fhp_t *fhp,
   fhp_token_t token,
   uint8_t * const buf,
@@ -59,21 +63,94 @@ test_basic(void) {
   fhp_err_t err;
   fhp_t fhp;
 
-  if ((err = fhp_init(&fhp, test_basic_cb, NULL)) != FHP_OK) {
+  // init parser
+  if ((err = fhp_init(&fhp, basic_cb, NULL)) != FHP_OK) {
     die("test_basic", "fhp_init", err);
   }
 
-  size_t len = strlen(str_test_basic);
-  if ((err = fhp_push(&fhp, (uint8_t*) str_test_basic, len)) != FHP_OK) {
+  // parse data
+  size_t len = strlen(basic_str);
+  if ((err = fhp_push(&fhp, (uint8_t*) basic_str, len)) != FHP_OK) {
     die("test_basic", "fhp_push", err);
   }
 }
 
+/****************/
+/* percent test */
+/****************/
+
+typedef struct {
+  uint8_t buf[1024];
+  size_t buf_len;
+} percent_data;
+
+static const char *
+percent_str =
+  "GET /foo%20bar HTTP/1.1\r\n"
+  "Host: pablotron.org\r\n"
+  "Connection: close\r\n"
+  "\r\n";
+
+static bool
+percent_cb(
+  fhp_t *fhp,
+  fhp_token_t token,
+  uint8_t * const buf,
+  size_t len
+) {
+  percent_data *data = fhp_user_data(fhp);
+
+  switch (token) {
+  case FHP_TOKEN_URL_START:
+    // clear buffer
+    data->buf_len = 0;
+
+    break;
+  case FHP_TOKEN_URL_FRAGMENT:
+    // buffer overflow, do not use in real code!!!
+    memcpy(data->buf + data->buf_len, buf, len);
+    data->buf_len += len;
+
+    break;
+  case FHP_TOKEN_URL_END:
+    // terminate and print buffer
+    data->buf[data->buf_len] = '\0';
+    fprintf(stderr, "decoded URL: \"%s\"\n", data->buf);
+
+    break;
+  default:
+    // do nothing
+    NULL;
+  }
+
+  // return success
+  return true;
+}
+
+static void
+test_percent(void) {
+  fhp_err_t err;
+  fhp_t fhp;
+  percent_data data;
+
+  // init parser
+  if ((err = fhp_init(&fhp, percent_cb, &data)) != FHP_OK) {
+    die("test_percent", "fhp_init", err);
+  }
+
+  // parse data
+  size_t len = strlen(percent_str);
+  if ((err = fhp_push(&fhp, (uint8_t*) percent_str, len)) != FHP_OK) {
+    die("test_percent", "fhp_push", err);
+  }
+}
+
 int main(int argc, char *argv[]) {
   UNUSED(argc);
   UNUSED(argv);
 
   test_basic();
+  test_percent();
 
   return EXIT_SUCCESS;
 }