Re: [boost] [optional] Strict Aliasing Warnings on Trunk

Subject: Re: [boost] [optional] Strict Aliasing Warnings on Trunk
From: Dean Michael Berris (mikhailberis_at_[hidden])
Date: 2009-12-18 03:45:12

On Fri, Dec 18, 2009 at 4:11 PM, Emil Dotchevski
<emildotchevski_at_[hidden]> wrote:
> On Thu, Dec 17, 2009 at 11:17 PM, Dean Michael Berris
> <mikhailberis_at_[hidden]> wrote:
>>
>> Really? I thought this was implementation-defined?
>
> Sure, but it is specified that if you reinterpret_cast from T1* to
> T2*, then reinterpret_cast the resulting T2* back to T1*, you'll get
> the original value (as long as T2 doesn't have stricter alignment
> requirements.)
>

See, the strict alignment requirements are what the compiler is
warning about -- saying that it *may* break as a diagnostic is how it
does it. Because the issue raised has already been pointed out to be a
bug in that it's almost indiscriminate that even legitimate constructs
do get flagged as potentially breaking strict aliasing rules, I'd say
reinterpret_cast<> aside, the strict aliasing warnings do have merit
in situations that they may cause problems at runtime. It's not *only*
in cases that you have reinterpret_cast that it warns -- heck, it can
even warn for using C-style casts and not only in C++, but even in C
mode.

> Reinterpret_cast is tricky but it is part of the language and it's
> lame for the compiler to warn when it is used.

Which warning are you pointing at? The strict aliasing warnings?
Again, it's not only in cases where you have reinterpret_cast that it
warns, it also happens in cases where you do cast the types of the
pointers and try to dereference these type-punned or casted pointers.

What *is* lame is how after years of references and legitimate safe
pointer operations not involving reinterpret_cast, we're still having
a discussion about it -- reinterpret_cast. ;)

-- 
Dean Michael Berris
blog.cplusplus-soup.com | twitter.com/mikhailberis
linkedin.com/in/mikhailberis | facebook.com/dean.berris | deanberris.com