incubator-stdcxx-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Sebor <se...@roguewave.com>
Subject Re: svn commit: r638369 - in /stdcxx/trunk/include: loc/_num_get.cc loc/_num_get.h rw/_iosfwd.h
Date Mon, 24 Mar 2008 14:40:37 GMT
Farid Zaripov wrote:
>> -----Original Message-----
>> From: Martin Sebor [mailto:sebor@roguewave.com] 
>> Sent: Wednesday, March 19, 2008 6:46 PM
>> To: dev@stdcxx.apache.org
>> Subject: Re: svn commit: r638369 - in /stdcxx/trunk/include: 
>> loc/_num_get.cc loc/_num_get.h rw/_iosfwd.h
>>
>>> URL: http://svn.apache.org/viewvc?rev=638369&view=rev
>>> Log:
>>> 2008-03-18  Farid Zaripov  <farid_zaripov@epam.com>
>>>
>>> 	* include/loc/_num_get.cc (num_get::get): Code checking 
>> for overflow moved from here...
>>> 	* include/loc/num_get.h 
>> (_rw_check_overflow_{short|int}): ... to here.
>>> 	* include/rw/_iosfwd.h: Added new define 
>> _RWSTD_FMTFLAGS (used in _rw_check_overflow_{short|int}).
>>
>> It occurs to me: is this forward binary compatible? We're 
>> adding a function in this patch, even though it's inline, I 
>> suspect that in debug builds (w/o inlining) this will break 
>> when we replace 4.2.1 with 4.2.0. Have you tested it?
> 
>   I have tested on MSVC 7.1 and found no problems. The
> __rw_check_overflow_{short|int}()
> are not exported from library and can't be called directly, only through
> num_get<>::get().

I assume you tested with both debug and optimized shared lib
builds.

> 
>   Also I have tested on gcc/Linux and found no problems in forward and
> backward compatibility.

That's interesting. If this is true on other platforms it would
mean a lot more flexibility in terms of what changes can go in
a patch release. Let's try to test this on other platforms so
we can make a policy decision for 4.2.1 and beyond.

Martin

Mime
View raw message