[cfe-dev] 'const void' is considered to be an incomplete type by clang 3.5.0

[Richard Smith]

On Mon, Apr 21, 2014 at 2:27 PM, Stephen Canon <scanon at apple.com> wrote:

> On Apr 21, 2014, at 5:18 PM, Richard Smith <richard at metafoo.co.uk> wrote:
>
> On Mon, Apr 21, 2014 at 12:21 PM, Hal Finkel <hfinkel at anl.gov> wrote:
>
>> ----- Original Message -----
>> > From: "Yuri" <yuri at rawbw.com>
>> > To: "Hal Finkel" <hfinkel at anl.gov>
>> > Cc: cfe-dev at cs.uiuc.edu
>> > Sent: Monday, April 21, 2014 2:16:16 PM
>> > Subject: Re: [cfe-dev] 'const void' is considered to be an incomplete
>> type by clang 3.5.0
>> >
>> > On 04/21/2014 12:07, Hal Finkel wrote:
>> > > What do you mean? As I understand it, void is always incomplete.
>> > > What do you expect sizeof(const void) to return?
>> >
>> > 'incomplete' implies that it can be completed, and 'const void' can't
>> > be
>> > completed.
>>
>> Perhaps, but 'incomplete' here is the correct technical term. 3.9.1 says,
>> "The void type is an incomplete type that cannot be completed.".
>>
>> > So there are two different situations and this message
>> > doesn't differentiate between them.
>> > The message should say "can't apply 'sizeof' to 'const void'" to be
>> > more
>> > clear.
>> >
>> > Curiously, clang-3.3 thought that sizeof(const void) is 1.
>>
>> Interesting.
>>
>
> Hmm, maybe this was the GNU 'arithmetic on void pointers' extension
> leaking through?
>
>
> The gnu arithmetic on void pointers extension explicitly says sizeof(void)
> == 1, no leaking required.
>
> http://gcc.gnu.org/onlinedocs/gcc/Pointer-Arith.html
>

It also says "In GNU C". The leaking would be this extension leaking into
C++ mode.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20140421/81512ddb/attachment.html>