On 31 Aug 2016, at 11:46, Andy Wingo <wi...@pobox.com> wrote:
> On Wed 31 Aug 2016 11:28, Andy Wingo <wi...@pobox.com> writes:
>
>> On Thu 14 Jul 2016 20:20, Andy Wingo <wi...@pobox.com> writes:
>>
>>> I think the concerns are:
>>>
>>> (1) Do inlined definitions get inlined?
>>> (2) Are external definitions reified as well?
>>> (3) Do we avoid reifying definitions in each compilation unit?
>>> (4) Can you dlsym() an inline function?
>>>
>>> All these answers should be yes. No benchmarking needed, just
>>> inspection of the build artifacts under different configurations.
>>
>> I want to be able to use C11 atomics in Guile, internally. I think
>> externally as far as the interface goes we can probably upgrade to C99
>> given that even Emacs uses it these days. I will try to answer these
>> four questions with GCC with C89 and C99, and if the answers are good I
>> will upgrade to C99 first, for the build.
>
> Hah! Turns out we have been compiling in GCC's default mode the whole
> time, which is gnu11. I suspect many users have been doing that too.
> We might as well go ahead and require C99 internally plus the C11
> features we want. Again for the external interface (stdint.h et al) we
> can also probably require C99 for Guile 2.2, so that we can use uint32_t
> and not scm_t_uint32.
>
> Andy
Hi,
I've sent a patch to use C99 inline in a separate email. It's kind of obvious,
but let me know if it makes sense (or not) and I'll write a proper commit
message.
Regards
Daniel
