[boost] Re: Re: Re: Intel 7.1 bug report for boost::checked_delete?

From: Daniel Frey (d.frey_at_[hidden])
Date: 2003-07-15 16:48:55

On Tue, 15 Jul 2003 23:20:06 +0200, Peter Dimov wrote:

> Daniel Frey wrote:
>> On Tue, 15 Jul 2003 22:22:03 +0200, Peter Dimov wrote:
>>
>>>> +#include <boost/config.hpp>
>>>> +#if defined(BOOST_INTEL)
>>>> +#include <boost/static_assert.hpp>
>>>> +#endif
>>>
>>> Dependencies. I hate dependencies.
>>
>> Do you have a better idea?
> 
> Several, like (1) using __INTEL_COMPILER for detection and (2) isolating
> the minimal example that results in the error we need. But I don't have
> access to Intel C++.

(1) Look into config/compilers/intel.hpp, it's not that easy. I do have
access to the Intel compiler, but of course not to all versions. And I
only use Linux, thus a single platform. I simply used all the knowledge
and experience compiled into the config-system as this is what it is
supposed to provide, right? But I can duplicate the config's code if you
want to detach checked_delete from the config system...

(2) This is only included for the Intel compiler and I prefer code clarity
here over duplicating strange work-around code. Actually
BOOST_STATIC_ASSERT nicely encapsulates this. Or do you know of any actual
problem by introducting the dependency? It's IMHO the smaller evil here...

Regards, Daniel

PS: While adding the comments Dave asked for, I also added a comment to my
local copy about the explicit namespace qualification in checked_deleter
to prevent ADL, as this is also not obvious from the code alone why it's
required. Should be OK, right? :)