Re: [PATCH 4/5] barrier: use int instead of unsigned char for the phase state

[Thiago Macieira <thiago.macieira@intel.com>]

On Monday, 1 March 2021 12:35:05 PST Ville Voutilainen wrote:
> I can't make the above code work, in any reasonable manner, because
> it's doing feature detection incorrectly. :)
> What I can imagine doing, however, is this:
> 
> 1) a published IS always bumps feature-macro values (this won't help
> now, it's a future fix, we don't currently do this in WG21)

This is mostly already the case. I haven't seen any need to bump the values.

> 2) an implementation uses the pre-IS draft values before it deems itself
> stable 

Before the IS is stable? From what I've seen so far, the macros are always 
around the month of the latest draft of a given feature. So if an 
implementation implemented an earlier draft, the macro version would increase 
in a new draft coming before the IS.

I think we just need to be careful of updates done after the latest draft is 
adopted, usually by NBs. Those have to bump the macro too. I think you're in 
good speaking terms with the Finland NB :-)

> 3) users who want stability should require the feature-testing macro
> to have at least the IS value

True. That's not the stability I was talking about, but it could be done.

> 4) adventurous users can either use the macro without checking its
> value, or use at least the value that gives them
> whatever shiny toy they're interested in

Fair enough.

Please note I am not talking about the stability of the feature as described 
in the standard or a proposal or draft. I am talking about the stability of 
the implementation. C++20 is out and the atomic-wait feature is stable, as 
defined by the standard. What isn't stable is GCC's implementation of it.

But your suggestion does work. We don't need to apply them to all macros, only 
those that are new in a given version, like __cpp_lib_atomic_wait or 
__cpp_lib_latch in this case. Alternatively, implementations can set the macro 
to a given value below the standard-required one (incl. 1) to indicate that 
they haven't achieved stability.

That would make my check:

#if __cplusplus >= 202002L && __has_include(<latch>)
#  include <latch>
#endif
#if __cpp_lib_latch < 201907L
#  error "Please upgrade your Standard Library"
#endif

-- 
Thiago Macieira - thiago.macieira (AT) intel.com
  Software Architect - Intel DPG Cloud Engineering