1 files changed, 85 insertions, 0 deletions
diff --git a/src/ragel/rfc2047_parser.rl b/src/ragel/rfc2047_parser.rl
new file mode 100644
index 0000000..88e107b
--- /dev/null
+++ b/src/ragel/rfc2047_parser.rl
@@ -0,0 +1,85 @@
+%%{
+  # It actually implements rfc2047 + rfc2231 extension
+  machine rfc2047_parser;
+
+  action Start_Charset {
+    charset_start = p;
+  }
+
+  action End_Charset {
+    if (charset_start && p > charset_start) {
+      charset_end = p;
+    }
+  }
+
+  action End_Encoding {
+    if (p > in) {
+      switch (*(p - 1)) {
+      case 'B':
+      case 'b':
+        encoding = RSPAMD_RFC2047_BASE64;
+        break;
+      default:
+        encoding = RSPAMD_RFC2047_QP;
+        break;
+      }
+    }
+  }
+
+  action Start_Encoded {
+    encoded_start = p;
+  }
+
+  action End_Encoded {
+    if (encoded_start && p > encoded_start) {
+      encoded_end = p;
+    }
+  }
+
+  primary_tag = alpha{1,8};
+  subtag = alpha{1,8};
+  language = primary_tag ( "-" subtag )*;
+  especials = "(" | ")" | "<" | ">" | "@" | "," | ";" | ":" | "\"" | "/" | "[" | "]" | "?" | "." | "=" | "*";
+  token = (graph - especials)+;
+  charset = token;
+  encoding = "Q" | "q" | "B" | "b";
+  encoded_text = (print+ -- ("?="));
+  encoded_word = "=?" charset >Start_Charset %End_Charset
+    ("*" language)? "?"
+    encoding %End_Encoding "?"
+    encoded_text >Start_Encoded %End_Encoded
+    "?="?;
+  main := encoded_word;
+}%%
+
+#include "smtp_parsers.h"
+#include "mime_headers.h"
+
+%% write data;
+
+gboolean
+rspamd_rfc2047_parser (const gchar *in, gsize len, gint *pencoding,
+  const gchar **charset, gsize *charset_len,
+  const gchar **encoded, gsize *encoded_len)
+{
+  const char *p = in, *pe = in + len,
+    *encoded_start = NULL, *encoded_end = NULL,
+    *charset_start = NULL, *charset_end = NULL,
+    *eof = in + len;
+  gint encoding = RSPAMD_RFC2047_QP, cs = 0;
+
+  %% write init;
+  %% write exec;
+
+  if (encoded_end) {
+    *pencoding = encoding;
+    *charset = charset_start;
+    *charset_len = charset_end - charset_start;
+    *encoded = encoded_start;
+    *encoded_len = encoded_end - encoded_start;
+
+    return TRUE;
+  }
+
+  return FALSE;
+}