[
https://issues.apache.org/activemq/browse/AMQCPP-56?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_38017
]
Albert Strasheim commented on AMQCPP-56:
----------------------------------------
GCC 4.1.1 doesn't seem to like the AMQCPP_UNUSED macro -- something about
__attribute__. You might consider simply leaving out the parameter name or
exception name. I think most compilers support that?
A few more warning I spotted:
{noformat}
activemq/io/DataInputStream.cpp:305: warning: comparison between signed and
unsigned integer expressions
activemq/io/BufferedOutputStream.cpp:92: warning: comparison between signed and
unsigned integer expressions
activemq/io/BufferedOutputStream.cpp:106: warning: comparison between signed
and unsigned integer expressions
activemq/connector/stomp/StompCommandReader.cpp:111: warning: comparison of
unsigned expression >= 0 is always true
activemq/network/ServerSocket.cpp:157: warning: dereferencing type-punned
pointer might break strict-aliasing rules
activemq/network/ServerSocket.cpp:203: warning: dereferencing type-punned
pointer might break strict-aliasing rules
activemq/network/TcpSocket.cpp:190: warning: dereferencing type-punned pointer
might break strict-aliasing rules
activemq/network/SocketOutputStream.cpp:84: warning: comparison between signed
and unsigned integer expressions
{noformat}
Some of these might be inevitable signed<->unsigned conversions, which probably
calls for some static_casts.
Not sure how to address those warnings in ServerSocket. I googled around a bit
but couldn't find a definite answer on how one should handle that cast.
> GCC warnings and possible SocketFactory bug on 64-bit Linux
> -----------------------------------------------------------
>
> Key: AMQCPP-56
> URL: https://issues.apache.org/activemq/browse/AMQCPP-56
> Project: ActiveMQ C++ Client
> Issue Type: Bug
> Affects Versions: 2.0
> Reporter: Albert Strasheim
> Assigned To: Nathan Mittler
> Priority: Minor
> Fix For: 2.0
>
>
> When compiling on 64-bit Linux with GCC 4.0.3, I see the following warnings:
> {noformat}
> activemq/concurrent/PooledThread.cpp:56: warning: 'task$first' may be used
> uninitialized in this function
> activemq/concurrent/PooledThread.cpp:56: warning: 'task$second' may be used
> uninitialized in this function
> activemq/network/SocketFactory.cpp:46: warning: comparison is always false
> due to limited range of data type
> activemq/connector/openwire/marshal/v2/ConnectionControlMarshaller.cpp:82:
> warning: unused variable 'info'
> activemq/connector/openwire/marshal/v2/IntegerResponseMarshaller.cpp:60:
> warning: unused variable 'info'
> activemq/connector/openwire/marshal/v2/ReplayCommandMarshaller.cpp:61:
> warning: unused variable 'info'
> activemq/connector/openwire/marshal/v2/ResponseMarshaller.cpp:60: warning:
> unused variable 'info'
> {noformat}
> The first three warnings probably need to be looked at, especially the one in
> SocketFactory.
> In my opinion, AMQCPP is using ints and unsigned ints in places where
> std::size_t might be more appropriate. MSVC's warning level 3 catches many of
> these issues, so we might consider upping the warning level and properly
> dealing with the type conversion issues: mostly use std::size_t and
> static_cast where we really intend to convert. I can prepare a few patches
> along these lines if you think this is worthwhile.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
