[ 
https://issues.apache.org/jira/browse/PROTON-2237?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17176545#comment-17176545
 ] 

ASF GitHub Bot commented on PROTON-2237:
----------------------------------------

astitcher commented on a change in pull request #256:
URL: https://github.com/apache/qpid-proton/pull/256#discussion_r469470459



##########
File path: python/proton/_message.py
##########
@@ -90,13 +90,18 @@ def _check(self, err):
 
     def _check_property_keys(self):

Review comment:
       It would be really helpful here to explain what the AMQP limitations on 
property keys are. I can't figure out from reading this code what the actual 
point of the code is.

##########
File path: python/proton/_message.py
##########
@@ -90,13 +90,18 @@ def _check(self, err):
 
     def _check_property_keys(self):
         for k in self.properties.keys():
-            if isinstance(k, unicode):
-                # py2 unicode, py3 str (via hack definition)
+            # Check for string types. (py2: unicode, py3: str via type hack 
above)
+            # or string subclasses. Exclude string subclasses symbol and char.
+            if isinstance(k, unicode) and not (type(k) is symbol or type(k) is 
char):
+                # Convert string subclasses to string
+                if not type(k) is unicode:
+                    self.properties[unicode(k)] = self.properties.pop(k)
                 continue

Review comment:
       But if the above code is actually necessary for some reason, then this 
line is no longer necessary. As in context it is just another way of saying 
'pass'

##########
File path: python/proton/_message.py
##########
@@ -90,13 +90,18 @@ def _check(self, err):
 
     def _check_property_keys(self):
         for k in self.properties.keys():
-            if isinstance(k, unicode):
-                # py2 unicode, py3 str (via hack definition)
+            # Check for string types. (py2: unicode, py3: str via type hack 
above)
+            # or string subclasses. Exclude string subclasses symbol and char.
+            if isinstance(k, unicode) and not (type(k) is symbol or type(k) is 
char):
+                # Convert string subclasses to string
+                if not type(k) is unicode:
+                    self.properties[unicode(k)] = self.properties.pop(k)

Review comment:
       Why is this even necessary? It's already a (perhaps subclass of) 
unicode. won't encode just work directly?

##########
File path: python/proton/_message.py
##########
@@ -90,13 +90,18 @@ def _check(self, err):
 
     def _check_property_keys(self):
         for k in self.properties.keys():
-            if isinstance(k, unicode):
-                # py2 unicode, py3 str (via hack definition)
+            # Check for string types. (py2: unicode, py3: str via type hack 
above)
+            # or string subclasses. Exclude string subclasses symbol and char.
+            if isinstance(k, unicode) and not (type(k) is symbol or type(k) is 
char):
+                # Convert string subclasses to string
+                if not type(k) is unicode:
+                    self.properties[unicode(k)] = self.properties.pop(k)
                 continue
-            # If key is binary then change to string
-            elif isinstance(k, str):
-                # py2 str
-                self.properties[k.encode('utf-8')] = self.properties.pop(k)
+            # If key is binary then change to string. Exclude bytes subclass 
decimal128.
+            # Mostly for py2 users who encode strings without using the u'' 
prefix
+            # but py3 bytes() type will be converted also

Review comment:
       I don't think converting py3 bytes is acceptable as binary is not a good 
key (or is it?) it looks like in the previous code it would have triggered an 
error - only py2 binary would have been converted.

##########
File path: python/proton/_message.py
##########
@@ -90,13 +90,18 @@ def _check(self, err):
 
     def _check_property_keys(self):
         for k in self.properties.keys():
-            if isinstance(k, unicode):
-                # py2 unicode, py3 str (via hack definition)
+            # Check for string types. (py2: unicode, py3: str via type hack 
above)
+            # or string subclasses. Exclude string subclasses symbol and char.
+            if isinstance(k, unicode) and not (type(k) is symbol or type(k) is 
char):
+                # Convert string subclasses to string
+                if not type(k) is unicode:
+                    self.properties[unicode(k)] = self.properties.pop(k)
                 continue
-            # If key is binary then change to string
-            elif isinstance(k, str):
-                # py2 str
-                self.properties[k.encode('utf-8')] = self.properties.pop(k)
+            # If key is binary then change to string. Exclude bytes subclass 
decimal128.
+            # Mostly for py2 users who encode strings without using the u'' 
prefix
+            # but py3 bytes() type will be converted also
+            elif isinstance(k, bytes) and not (type(k) is decimal128):

Review comment:
       I think you can leave this test as 'str' as it was before - it only 
covers the py2 str (binary) case as py3 str is already covered above in 
unicode, I guess you still need the decimal128 case still though.




----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


> [python] Non-string message property keys not handled correctly
> ---------------------------------------------------------------
>
>                 Key: PROTON-2237
>                 URL: https://issues.apache.org/jira/browse/PROTON-2237
>             Project: Qpid Proton
>          Issue Type: Bug
>          Components: python-binding
>            Reporter: Kim van der Riet
>            Priority: Major
>
> The AMQP 1.0 spec allows only string keys for message properties.
> Proton's Python binding has a method (_message.py:91 _check_property_keys()) 
> for checking message property keys, but it does not handle all cases 
> correctly:
>  # Proton types Symbol and Char are derived from string, and are allowed in 
> the test. This results in an illegal encoding.
>  # Because in Python 2, many coders carelessly use string literals without 
> the required u'' prefix (and thus results in a bytes type), bytes types are 
> converted to unicode string types. However, the encode() function is being 
> used, which simply returns a binary type in Python 2 and raises an error in 
> Python 3. This should probably be the decode() method, which returns a string 
> and works for both Python 2 and 3.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@qpid.apache.org
For additional commands, e-mail: dev-h...@qpid.apache.org

Reply via email to