[cfe-dev] Clang/C++ problem?

[Marshall Clow]

On Jan 24, 2011, at 1:33 PM, Steven Watanabe wrote:

> AMDG
> 
> On 1/24/2011 1:08 PM, Marshall Clow wrote:
>> Shouldn't this work?
> 
> My gut feel reaction is that it shouldn't,
> but I haven't actually checked the standard.
> 
>> If it should work, should I write up a bug?
>> 
>> $ cat junk.cpp
>>> typedef int Foo_t;
>>> 
>>> template<typename T>
>>> T DoSomething ( T t) { return t; }
>>> 
>>> int main ( int argc, char *argv [] ) {
>>>     int foo = 23;
>>>     DoSomething ( static_cast<unsigned int>(foo));
>>>     DoSomething ( static_cast<unsigned Foo_t>(foo));
>>>     return 0;
>>>     }
>>> 
>> 
> 
> What about something simpler?
> Does
> 
> typedef int Foo_t;
> typedef unsigned Foo_t uFoo_t;

No it does not - but gcc doesn't accept it either.
$ g++ junk.cpp
junk.cpp:2: error: invalid combination of multiple type-specifiers

On the other hand, clang doesn't like this (and gcc thinks that it's fine)
	DoSomething ((unsigned Foo_t)(foo));

I've opened bug #9040, but I'm still updating it.

Marshall Clow     Idio Software   <mailto:mclow.lists at gmail.com>

A.D. 1517: Martin Luther nails his 95 Theses to the church door and is promptly moderated down to (-1, Flamebait).
        -- Yu Suzuki