(just my personal notes, nothing official)

Clark Nelson, John Spicer, Jens Maurer, Walter Brown

2015-12-16 18:04 UTC

Clark: The evening session in Kona was not very productive.
Too much was going on elsewhere, Gaby didn't attend.
All the interesting stuff happened Friday afternoon, but that
did not include feature test macros in TSes. That question is
still unresolved by WG21 at large. It seems only a small number of people
think we're doing the wrong thing.

Clark: Wants to publish a new revision of SD-6 before Kona.

Discuss open issues in the current document:
 - About half of the passed proposals in Kona contained a macro name.
We're making progress.
 - P0007R1: use __cpp_lib_as_const (drop _function)
 - N4387 pair/tuple: This seems to be bug fix. No macro is necessary/useful.
 - P0074 std::owner_less more flexible: __cpp_lib_owner_less_void;
    see N3421 for the std::less<void> -> __cpp_lib_transparent_operators
    Jens: postpone
    John: We have lower hanging fish to fry.
 - P0006 feature test macro only in <type_traits> -> __cpp_lib_type_traits_variable_templates
 - P0092 polishing chrono -> __cpp_lib_chrono
 - N4258 -> none, ask paper author (was: __cpp_lib_allocator_traits_is_always_equal)
 - N4510 incomplete element types for containers -> __cpp_lib_incomplete_container_element_types
  Jens: argues in favor, others undecided
  John: emphasizes required support from WG chairs for feature test macros
 - N4379 -> __cpp_lib_try_emplace (not insertion)
 - N4508 __cpp_lib_shared_mutex was deleted in the previous version;
    do we want to consider this macro name burnt? No.
 - P0156R0 -> __cpp_lib_lock_guard_variadic
 - postpone: conditionally-support stuff

Walter: Today is Beethoven's birthday.

close at 19:18 UTC


Jens