Handle byte type on sockets

[mudpy.git] / lib / mudpy / misc.py

diff --git a/lib/mudpy/misc.py b/lib/mudpy/misc.py
index 3193e11..8f6630f 100644 (file)
--- a/lib/mudpy/misc.py
+++ b/lib/mudpy/misc.py
@@ -1,7 +1,7 @@
 # -*- coding: utf-8 -*-
 """Miscellaneous functions for the mudpy engine."""
 
-# Copyright (c) 2004-2012 Jeremy Stanley <fungi@yuggoth.org>. Permission
+# Copyright (c) 2004-2013 Jeremy Stanley <fungi@yuggoth.org>. Permission
 # to use, copy, modify, and distribute this software is granted under
 # terms provided in the LICENSE file distributed with this software.
 
@@ -589,7 +589,7 @@ class User:
         self.menu_seen = False
         self.negotiation_pause = 0
         self.output_queue = []
-        self.partial_input = ""
+        self.partial_input = b""
         self.password_tries = 0
         self.state = "initial"
         self.telopts = {}
@@ -924,7 +924,7 @@ class User:
         try:
             raw_input = self.connection.recv(1024)
         except:
-            raw_input = ""
+            raw_input = b""
 
         # we got something
         if raw_input:
@@ -936,18 +936,18 @@ class User:
             mudpy.telnet.negotiate_telnet_options(self)
 
             # separate multiple input lines
-            new_input_lines = self.partial_input.split("\n")
+            new_input_lines = self.partial_input.split(b"\n")
 
             # if input doesn't end in a newline, replace the
             # held partial input with the last line of it
-            if not self.partial_input.endswith("\n"):
+            if not self.partial_input.endswith(b"\n"):
                 self.partial_input = new_input_lines.pop()
 
             # otherwise, chop off the extra null input and reset
             # the held partial input
             else:
                 new_input_lines.pop()
-                self.partial_input = ""
+                self.partial_input = b""
 
             # iterate over the remaining lines
             for line in new_input_lines:
@@ -958,8 +958,8 @@ class User:
                 # log non-printable characters remaining
                 if mudpy.telnet.is_enabled(self, mudpy.telnet.TELOPT_BINARY,
                                            mudpy.telnet.HIM):
-                    asciiline = "".join(
-                        filter(lambda x: " " <= x <= "~", line))
+                    asciiline = b"".join(
+                        filter(lambda x: b" " <= x <= b"~", line))
                     if line != asciiline:
                         logline = "Non-ASCII characters from "
                         if self.account and self.account.get("name"):
@@ -970,10 +970,22 @@ class User:
                         log(logline, 4)
                         line = asciiline
 
+                try:
+                    line = line.decode("utf-8")
+                except UnicodeDecodeError:
+                    logline = "Non-UTF-8 characters from "
+                    if self.account and self.account.get("name"):
+                        logline += self.account.get("name") + ": "
+                    else:
+                        logline += "unknown user: "
+                    logline += repr(line)
+                    log(logline, 4)
+                    return
+
+                line = unicodedata.normalize("NFKC", line)
+
                 # put on the end of the queue
-                self.input_queue.append(
-                    unicodedata.normalize("NFKC", line.decode("utf-8"))
-                )
+                self.input_queue.append(line)
 
     def new_avatar(self):
         """Instantiate a new, unconfigured avatar for this user."""