This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: RFA (libstdc++): C++/v3 PATCH for c++/24163 (lookup in dependent bases) and c++/29131


On 20 May 2011, at 20:30, Joe Buck wrote:

> On Fri, May 20, 2011 at 09:32:16AM -0700, Jason Merrill wrote:
>> G++ has had a long-standing bug with unqualified name resolution in 
>> templates: if we didn't find any declaration when looking up a name in 
>> the template definition, we would do an additional unqualified lookup at 
>> the point of instantiation.  This led to incorrectly finding 
>> namespace-scope functions declared later (29131) and member functions of 
>> dependent bases (24163).  This patch fixes that bug.
> 
> I get the impression that most competing C++ compilers (other than the
> old HP compiler) were (or are) very loose about that rule.
> 
>> To be friendly to users, the patch also allows affected code to compile 
>> with -fpermissive and provides suggestions about how to fix the code: 
>> either declaring the desired function earlier (29131) or explicitly 
>> qualifying the name with this-> or Class:: (24163).
> 
> I think that it's quite likely that there is a lot of C++ code out there
> that depends on this bug to compile.  So I'm glad that you've included
> user guidance in the error messages, and it would be interesting to see
> how much code is affected when, say, compiling a distro.

Fortunately clang already implements this rule correctly, so many code bases are already getting fixed. I personally (with a couple of other people) fixed dozens of places in boost where this was breaking code. I expect this to break huge amounts of code in g++-only code bases, but will also help improve compatibility with other compilers.

I could see the temptation to introduce this as a mandatory warning for a while, and only add it under -pedantic. However, it might be easier to just force people to fix their code.

Chris



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]