Follow up on this... The only place the warnings need to be suppressed
is around the boost includes from qpid/Options.h.

The qpid-1673 branch has the committed files with boost dynamic
linking. I'm going to start merging trunk out to qpid-1673, test, then
merge back to trunk.

-Steve

--
Steve Huston, Riverace Corporation
Check out my networked programming blog at
http://stevehuston.wordpress.com/


> -----Original Message-----
> From: Steve Huston [mailto:[email protected]] 
> Sent: Tuesday, March 10, 2009 11:23 AM
> To: [email protected]
> Subject: RE: Experiences with qpid-1673 branch?
> 
> 
> Hi Danushka,
> 
> > > I think I understand the issues the compiler is warning 
> > against, but I
> > > don't have 100% confidence. What rationale did you use for
> disabling
> > > the warnings?
> > >   
> > Hi Steve,
> > 
> > I am sorry for the late reply in the first place.
> 
> No problem - thanks for your reply.
> 
> > Well ... we do not have any control what so ever on this as 
> > Boots itself 
> > possesses these issues. On the other hand as far I can see, this
> does 
> > not lead to any malfunctioning. When I checked on the Boost
mailing 
> > lists, the common knowledge is to disable the warning and proceed.
> 
> Ok, I see, and agree. I still don't feel safe blindly disabling the
> warning globally since we really need to know if Qpid has code that
> triggers one of these so it can be evaluated. I'm going to try
> selectively disabling the warnings around the usage.
> 
> To other devs... This selective disabling involves compiler-specific
> ifdefs in the header files... Like:
> 
> #ifdef MSVC
> #  pragma warning(push)
> #  pragma warning(disable : 4251 4231 4660)
> #endif
> 
> // code here
> 
> #ifdef MSVC
> #pragma warning(pop)
> #endif
> 
> Unfortunately, this is not the kind of thing that can be placed in a
> separate reusable header since the warning numbers may change on a
> case-by-case basis. The trade-off here is keeping compiler-specific
> things out of the headers vs. only disabling the warning in specific
> spots where we believe it to be safe.
> 
> Thoughts?
> 
> Thanks,
> -Steve
> 
> 
>
---------------------------------------------------------------------
> Apache Qpid - AMQP Messaging Implementation
> Project:      http://qpid.apache.org
> Use/Interact: mailto:[email protected]
> 
> 


---------------------------------------------------------------------
Apache Qpid - AMQP Messaging Implementation
Project:      http://qpid.apache.org
Use/Interact: mailto:[email protected]

Reply via email to