From 1baa78dab630530cabe06ca0d48ac3ab2b30a0f7 Mon Sep 17 00:00:00 2001
From: Paul Duncan <pabs@pablotron.org>
Date: Fri, 26 Aug 2016 23:57:45 -0400
Subject: add url percent decoding

---
 fhp.c | 193 +++++++++++++++++++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 157 insertions(+), 36 deletions(-)

(limited to 'fhp.c')

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;
@@ -329,12 +404,54 @@ 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) {
@@ -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;
+}
-- 
cgit v1.2.3