activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Martin Schlapfer (JIRA)" <j...@apache.org>
Subject [jira] Commented: (AMQCPP-232) OpenWire encode and decode UTF8 incorrect
Date Mon, 30 Mar 2009 23:27:34 GMT

    [ https://issues.apache.org/activemq/browse/AMQCPP-232?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=50906#action_50906
] 

Martin Schlapfer commented on AMQCPP-232:
-----------------------------------------

Tim/Peter, 

Having just worked in this area on another piece of software, I have a few code review comments
on this patch: 

(1) In the readString method, the transformation from UTF8 to Unicode should be limited to
1 byte (max value 255) since the UTF8 data is decoded in this method into a byte array. Decoding
values above 255 and stuffing the value into a byte value will only cause debugging problems
down then road when receiving UTF8 data with values above 255. Thus, as with the code before
patch was applied (with values above 127) , the readString method should throw an IO Exception
if a Unicode value greater than 255 is encountered indicating "Encoding is not supported".


(2) For performance reasons (although not much of a factor with 1 byte Unicode, however greater
factor in supporting 2 byte and 4 byte Unicode), bitwise operators should be used to decode
/ encode between UTF8 and Unicode rather than arithmetic. 

(3) The "null character", value 0, should not be skipped. It should be treated as a character
and decoded / endcoded along with the rest of the characters. The null character is a valid
value in UTF8 and Unicode (and c++ std::string's). The null character is a C style string
programming artifact. 

my two cents, thanks, 
Martin. 



> OpenWire encode and decode UTF8 incorrect
> -----------------------------------------
>
>                 Key: AMQCPP-232
>                 URL: https://issues.apache.org/activemq/browse/AMQCPP-232
>             Project: ActiveMQ C++ Client
>          Issue Type: Bug
>          Components: Openwire
>    Affects Versions: 2.2.5
>         Environment: Windows XP SP 3, Visual Studio 2008
>            Reporter: Peter Pfort
>            Assignee: Timothy Bish
>             Fix For: 2.2.6, 3.0
>
>         Attachments: OpenwireStringSupport.patch
>
>
> Hallo,
> we are using topic messages to sent messages from one user to another. Our program subscribe
a durable consumer with selector "UserName='<user>'" and send a message with the property
"UserName" and value "<user>".
> All works fine, when <user> contains only ASCII characters. When <user> contains
non ASCII characters like äöüßé, the message is not send to the
>  consumer.
> The problem ist that readString and writeString in OpenwireStringSupport.cpp have bugs
> Regards,
> Peter

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message