| Issue |
Status |
Section |
Title |
Proposed Resolution |
Priority |
Duplicates |
| 2392(i) |
New |
3.36 [defns.ntcts] |
"character type" is used but not defined |
No |
3 |
|
| 3513(i) |
New |
3.43 [defns.prog.def.type] |
Fix definition of program-defined based on its uses |
Yes |
3 |
|
| 3669(i) |
New |
4.1.2 [intro.abstract] |
std::filesystem operations should be observable behaviour |
No |
3 |
|
| 2506(i) |
SG1 |
6.9.2 [intro.multithread] |
Underspecification of atomics |
No |
3 |
|
| 3105(i) |
New |
16 [library] |
T1 is convertible to T2 |
No |
3 |
|
| 2136(i) |
Open |
16.3.2 [structure] |
Postconditions vs. exceptions |
No |
3 |
|
| 3556(i) |
New |
16.3.2.3 [structure.requirements] |
Specification of when semantic constraints are imposed by use of concepts
is unclear |
No |
3 |
|
| 3401(i) |
New |
16.3.2.4 [structure.specifications] |
Is "as if by" equivalent to "equivalent to"? |
No |
3 |
|
| 3193(i) |
New |
16.3.2.4 [structure.specifications] |
Mandates: and Expects: elements are not defined for types |
Yes |
3 |
|
| 3092(i) |
Open |
16.3.3.3.3 [bitmask.types] |
Unclear semantics of enum class bitmask types |
Yes |
3 |
|
| 3620(i) |
New |
16.3.3.3.4.1 [character.seq.general] |
What are execution character sets and execution wide-character sets (after P2314R4)? |
No |
3 |
|
| 3690(i) |
New |
16.4.2.2 [contents] |
std::make_from_tuple etc. should find all tuple-like std::get overloads |
Yes |
3 |
|
| 3240(i) |
New |
16.4.3.2 [using.headers] |
Headers declare more than entities |
No |
3 |
|
| 3640(i) |
New |
16.4.4 [utility.requirements] |
Clarify which exceptions are propagated |
Yes |
3 |
|
| 2146(i) |
Open |
16.4.4.2 [utility.arg.requirements] |
Are reference types Copy/Move-Constructible/Assignable or Destructible? |
No |
3 |
|
| 2152(i) |
LEWG |
16.4.4.3 [swappable.requirements] |
Instances of standard container types are not swappable |
Yes |
3 |
|
| 3157(i) |
New |
16.4.4.6 [allocator.requirements] |
Allocator destroy and fancy pointer operations must be non-throwing |
Yes |
3 |
|
| 2461(i) |
New |
16.4.4.6 [allocator.requirements] |
Interaction between allocators and container exception safety guarantees |
No |
3 |
|
| 3044(i) |
New |
16.4.4.6 [allocator.requirements] |
Strange specification of max_size() for an allocator |
Yes |
3 |
|
| 3682(i) |
New |
16.4.4.6.1 [allocator.requirements.general] |
A Cpp17Allocator type can't silently ignore an unsupported alignment |
No |
3 |
|
| 3928(i) |
New |
16.4.5.2.2 [namespace.posix] |
Non-top-level namespace posix shouldn't be reserved |
Yes |
3 |
|
| 3550(i) |
New |
16.4.5.3 [reserved.names] |
Names reserved by C for standard library not reserved by C++ |
No |
3 |
|
| 3920(i) |
New |
16.4.5.3.4 [extern.names] |
Bad footnotes claiming external linkage for entities defined as macros |
No |
3 |
|
| 3142(i) |
New |
16.4.5.8 [res.on.functions] |
std::foo<incomplete> should be ill-formed NDR |
Yes |
3 |
|
| 3511(i) |
New |
16.4.5.9 [res.on.arguments] |
Clarify global permission to move |
Yes |
3 |
|
| 3429(i) |
New |
16.4.5.11 [res.on.requirements] |
"models" should subsume like "satisfies" |
Yes |
3 |
|
| 2695(i) |
New |
16.4.6.5 [member.functions] |
"As if" unclear in [member.functions] |
No |
3 |
|
| 2414(i) |
Open |
16.4.6.9 [reentrancy] |
Member function reentrancy should be implementation-defined |
Yes |
3 |
|
| 3229(i) |
New |
16.4.6.13 [res.on.exception.handling] |
§[res.on.exception.handling]#3 cannot apply to types with implicitly declared destructors |
Yes |
3 |
|
| 3854(i) |
New |
16.4.6.13 [res.on.exception.handling] |
§[res.on.exception.handling]/3 should not be applied to all standard library types |
No |
3 |
|
| 3217(i) |
New |
17.3.1 [support.limits.general] |
<memory> and <execution> should define __cpp_lib_parallel_algorithm |
Yes |
3 |
|
| 3931(i) |
New |
17.3.2 [version.syn] |
Too many paper bump __cpp_lib_ranges |
Yes |
3 |
|
| 2730(i) |
Open |
17.3.5 [numeric.limits] |
numeric_limits primary template definition |
No |
3 |
|
| 3922(i) |
New |
17.3.5.1 [numeric.limits.general] |
It's unclear whether numeric_limits can be specialized by users |
Yes |
3 |
|
| 3923(i) |
New |
17.3.5.1 [numeric.limits.general] |
The specification of numeric_limits doesn't clearly distinguish between implementation requirements
and user requirements |
Yes |
3 |
|
| 3370(i) |
New |
17.4.1 [cstdint.syn] |
§[cstdint.syn]p2 and §[headers]p5 are not sufficiently clear |
No |
3 |
|
| 2815(i) |
New |
17.5 [support.start.term] |
quick_exit can deadlock |
Yes |
3 |
|
| 3084(i) |
New |
17.5 [support.start.term] |
Termination in C++ is unclear |
No |
3 |
|
| 2737(i) |
New |
17.6.3.2 [new.delete.single] |
Consider relaxing object size restrictions for single-object allocation functions |
No |
3 |
|
| 3086(i) |
New |
17.6.3.2 [new.delete.single] |
Possible problem in §[new.delete.single] |
Yes |
3 |
|
| 2303(i) |
New |
17.6.3.4 [new.delete.placement] |
Explicit instantiation of std::vector<UserType> broken? |
No |
3 |
|
| 2508(i) |
New |
17.6.3.5 [new.delete.dataraces] |
§[new.delete.dataraces] wording needs to be updated |
No |
3 |
|
| 3624(i) |
New |
17.7 [support.rtti] |
Inconsistency of <typeinfo>, <initializer_list>, and
<compare> in the standard library |
Yes |
3 |
|
| 2398(i) |
Open |
17.7.3 [type.info] |
type_info's destructor shouldn't be required to be virtual |
Yes |
3 |
|
| 2088(i) |
Open |
17.9.5 [exception.terminate] |
std::terminate problem |
No |
3 |
|
| 2453(i) |
New |
17.10 [support.initlist] |
§[iterator.range] and now [iterator.container] aren't available via <initializer_list> |
No |
3 |
|
| 3584(i) |
New |
17.11.3 [cmp.common] |
Clarify common comparison category conversions |
Yes |
3 |
|
| 3587(i) |
New |
17.11.4 [cmp.concept] |
std::three_way_comparable_with<T, U, void> can be satisfied but can't be modeled |
No |
3 |
|
| 3491(i) |
New |
17.11.6 [cmp.alg] |
What is a "decayed type"? |
No |
3 |
|
| 3932(i) |
New |
17.11.6 [cmp.alg] |
Expression-equivalence is sometimes unimplementable when passing prvalue expressions to comparison CPOs |
No |
3 |
|
| 3653(i) |
New |
17.12.2 [coroutine.syn] |
<coroutine> is freestanding, but uses std::hash which is not |
No |
3 |
|
| 3484(i) |
New |
17.14.7 [support.c.headers.other] |
Should <stddef.h> declare ::nullptr_t? |
Yes |
3 |
|
| 3799(i) |
New |
17.14.7 [support.c.headers.other] |
Should <math.h> provide 3-argument ::hypot overloads? |
No |
3 |
|
| 3895(i) |
New |
18.3 [concepts.syn] |
Various relation concepts are missing default values of the second template parameters |
Yes |
3 |
|
| 3459(i) |
New |
18.4.4 [concept.convertible] |
Why doesn't std::convertible_to have semantic requirement when To is reference-to-function type? |
No |
3 |
|
| 3608(i) |
New |
18.4.4 [concept.convertible] |
convertible_to and temporary-bound references |
No |
3 |
|
| 3019(i) |
New |
19.5.3.4 [syserr.errcat.derived] |
Presentation of "program defined classes derived from error_category" [syserr.errcat.derived] unclear and contains mistakes |
No |
3 |
|
| 3053(i) |
New |
19.5.4.1 [syserr.errcode.overview] |
Prohibit error_code construction from rvalues of error_category |
Yes |
3 |
|
| 3162(i) |
New |
19.5.8.2 [syserr.syserr.members] |
system_error::system_error(error_code ec) not explicit |
Yes |
3 |
|
| 3625(i) |
New |
19.6.2 [stacktrace.syn] |
Should <stacktrace> provide range access function templates? |
Yes |
3 |
|
| 3626(i) |
New |
19.6.4.1 [stacktrace.basic.overview] |
Is std::basic_stacktrace required to use contiguous storage? |
Yes |
3 |
|
| 2421(i) |
New |
20.2.5 [ptr.align] |
Non-specification of handling zero size in std::align [ptr.align] |
No |
3 |
|
| 3192(i) |
New |
20.2.8.2 [allocator.uses.construction] |
§[allocator.uses.construction] functions misbehave for const types |
Yes |
3 |
|
| 3665(i) |
New |
20.2.9.2 [allocator.traits.types] |
Is std::allocator_traits<Alloc>::rebind_alloc SFINAE-friendly? |
No |
3 |
|
| 3916(i) |
New |
20.2.10 [default.allocator] |
allocator, polymorphic_allocator, and containers should forbid cv-qualified types |
No |
3 |
|
| 3917(i) |
New |
20.2.10 [default.allocator] |
Validity of allocator<void> and possibly polymorphic_allocator<void> should be clarified |
Yes |
3 |
|
| 3684(i) |
New |
20.2.10.2 [allocator.members] |
std::allocator<T>::allocate_at_least in constant evaluation |
Yes |
3 |
|
| 2262(i) |
Open |
20.3.1.3 [unique.ptr.single] |
Requirement for unique_ptr<T>::get_deleter()(p) to be able to destroy the unique_ptr |
Yes |
3 |
|
| 3159(i) |
New |
20.3.1.3 [unique.ptr.single] |
§[unique.ptr.single] requirements on deleter may be too strict |
No |
3 |
|
| 3911(i) |
New |
20.3.1.3.5 [unique.ptr.single.observers] |
unique_ptr's operator* is missing a mandate |
Yes |
3 |
|
| 2594(i) |
New |
20.3.2.2 [util.smartptr.shared] |
Contradicting definition of empty shared_ptr on shared_ptr(nullptr, d) |
Yes |
3 |
|
| 2906(i) |
New |
20.3.2.2.2 [util.smartptr.shared.const] |
There is no ability to supply an allocator for the control block when constructing a shared_ptr from a
unique_ptr |
No |
3 |
|
| 3210(i) |
New |
20.3.2.2.7 [util.smartptr.shared.create] |
allocate_shared is inconsistent about removing const from the pointer
passed to allocator construct and destroy |
Yes |
3 |
|
| 3216(i) |
New |
20.3.2.2.7 [util.smartptr.shared.create] |
Rebinding the allocator before calling construct/destroy in allocate_shared |
Yes |
3 |
|
| 3637(i) |
New |
20.4.2 [mem.res.class] |
pmr::memory_resource::do_allocate needs clarification |
No |
3 |
|
| 3634(i) |
New |
20.4.4 [mem.res.global] |
When are static-duration memory_resource objects destroyed? |
No |
3 |
|
| 2848(i) |
New |
20.4.5.2 [mem.res.pool.options] |
Pass-through threshold for pool allocator |
No |
3 |
|
| 2290(i) |
Open |
21 [meta] |
Top-level "SFINAE"-based constraints should get a separate definition in Clause 17 |
Yes |
3 |
|
| 2452(i) |
Core |
21 [meta] |
is_constructible, etc. and default arguments |
No |
3 |
|
| 2845(i) |
New |
21.3.2 [meta.rqmts] |
enable_if, result_of, common_type and aligned_storage do not meet the definition
of TransformationTrait |
No |
3 |
|
| 2077(i) |
Open |
21.3.5.4 [meta.unary.prop] |
Further incomplete constraints for type traits |
No |
3 |
|
| 2358(i) |
Open |
21.3.5.4 [meta.unary.prop] |
Apparently-bogus definition of is_empty type trait |
Yes |
3 |
|
| 2116(i) |
Open |
21.3.5.4 [meta.unary.prop] |
is_nothrow_constructible and destructors |
No |
3 |
|
| 2496(i) |
New |
21.3.5.4 [meta.unary.prop] |
Certain hard-to-avoid errors not in the immediate context are not allowed to be triggered by
the evaluation of type traits |
No |
3 |
|
| 3697(i) |
New |
21.3.5.4 [meta.unary.prop] |
Preconditions of reference_constructs_from_temporary/reference_converts_from_temporary seem wrong |
Yes |
3 |
|
| 3929(i) |
New |
21.3.5.4 [meta.unary.prop] |
Preconditions for type traits should be Mandates |
Yes |
3 |
|
| 2827(i) |
New |
21.3.5.4 [meta.unary.prop] |
is_trivially_constructible and non-trivial destructors |
No |
3 |
|
| 3174(i) |
New |
21.3.7 [meta.rel] |
Precondition on is_convertible is too strong |
Yes |
3 |
|
| 3400(i) |
New |
21.3.7 [meta.rel] |
Does is_nothrow_convertible consider destruction of the destination type? |
No |
3 |
|
| 3205(i) |
New |
21.3.8.7 [meta.trans.other] |
decay_t in the new common_type fallback should be remove_cvref_t |
Yes |
3 |
|
| 2599(i) |
New |
22.2.6 [declval] |
Library incomplete type permission phrase is unclear |
No |
3 |
|
| 2289(i) |
Open |
22.3.2 [pairs.pair] |
constexpr guarantees of defaulted functions still insufficient |
No |
3 |
|
| 3342(i) |
New |
22.3.2 [pairs.pair] |
Library wording uses "initializes x with y", which is underspecified |
No |
3 |
|
| 2766(i) |
New |
22.3.3 [pairs.spec] |
Swapping non-swappable types |
Yes |
3 |
|
| 3166(i) |
New |
22.3.4 [pair.astuple] |
No such descriptive element as Value: |
No |
3 |
|
| 3378(i) |
New |
22.4.2 [tuple.syn] |
tuple_size_v/tuple_element_t should be available when tuple_size/tuple_element are |
Yes |
3 |
|
| 2528(i) |
New |
22.4.4.1 [tuple.cnstr] |
Order of std::tuple construction unspecified |
No |
3 |
|
| 3583(i) |
New |
22.4.4.1 [tuple.cnstr] |
Clarify if/when short circuiting applies to conditions in Constraints: elements |
No |
3 |
|
| 2933(i) |
New |
22.4.5 [tuple.creation] |
PR for LWG 2773 could be clearer |
Yes |
3 |
|
| 3882(i) |
New |
22.4.9 [tuple.rel] |
tuple relational operators have confused friendships |
Yes |
3 |
|
| 2472(i) |
New |
22.4.9 [tuple.rel] |
Heterogeneous comparisons in the standard library can result in ambiguities |
No |
3 |
|
| 2990(i) |
Open |
22.5.3 [optional.optional] |
optional::value_type is not always a value type |
Yes |
3 |
|
| 3886(i) |
New |
22.5.3.1 [optional.optional.general] |
Monad mo' problems |
Yes |
3 |
|
| 2811(i) |
New |
22.5.3.2 [optional.ctor] |
"Selected constructor" wording is incorrect for optional/variant/any |
Yes |
3 |
|
| 2746(i) |
New |
22.5.3.4 [optional.assign] |
Inconsistency between requirements for emplace between optional and variant |
Yes |
3 |
|
| 3424(i) |
New |
22.5.3.6 [optional.observe] |
optional::value_or should never return a cv-qualified type |
Yes |
3 |
|
| 3613(i) |
New |
22.5.4 [optional.nullopt] |
Specify that nullopt_t is copyable |
Yes |
3 |
|
| 3627(i) |
New |
22.5.9 [optional.specalg] |
Inconsistent specifications for std::make_optional overloads |
Yes |
3 |
|
| 2881(i) |
New |
22.6.3 [variant.variant] |
Adopt section III of P0308R0 |
No |
3 |
|
| 3069(i) |
New |
22.6.3.4 [variant.assign] |
Move assigning variant's subobject corrupts data |
Yes |
3 |
|
| 3416(i) |
New |
22.7.4 [any.class] |
The Throws: specification of std::any does not mention allocation |
No |
3 |
|
| 3423(i) |
New |
22.7.5 [any.nonmembers] |
std::any_cast should never return a cv-qualified type |
Yes |
3 |
|
| 2348(i) |
Open |
22.9.2 [template.bitset] |
charT('1') is not the wide equivalent of '1' |
Yes |
3 |
|
| 3046(i) |
New |
22.10.6 [refwrap] |
Do not require reference_wrapper to support non-referenceable function types |
Yes |
3 |
|
| 2547(i) |
New |
22.10.8 [comparisons] |
Container requirements (and other library text) should say "strict total order", not just "total order" |
No |
3 |
|
| 2491(i) |
New |
22.10.8 [comparisons] |
std::less<T*> in constant expression |
Yes |
3 |
|
| 3493(i) |
New |
22.10.17.3.2 [func.wrap.func.con] |
The constructor of std::function taking an F is missing a constraint |
Yes |
3 |
|
| 3642(i) |
New |
22.10.17.4.3 [func.wrap.move.ctor] |
move_only_function assignment operators seem to be defined suboptimal |
Yes |
3 |
|
| 3512(i) |
New |
22.10.19 [unord.hash] |
Incorrect exception safety guarantee for unordered containers |
No |
3 |
|
| 3456(i) |
New |
22.13.3 [charconv.from.chars] |
Pattern used by std::from_chars is underspecified |
Yes |
3 |
|
| 3651(i) |
New |
22.14 [format] |
Unspecified lifetime guarantees for the format string |
No |
3 |
|
| 3641(i) |
New |
22.14.1 [format.syn] |
Add operator== to format_to_n_result |
Yes |
3 |
|
| 3939(i) |
New |
22.14.2.2 [format.string.std] |
§[format.string.std] char is not formatted as a character when charT is wchar_t |
No |
3 |
|
| 3776(i) |
LEWG |
22.14.6.1 [formatter.requirements] |
Avoid parsing format-spec if it is not present or empty |
Yes |
3 |
|
| 3943(i) |
New |
22.14.6.2 [format.formattable] |
Clarify lifetime requirements of BasicFormatter and Formatter |
Yes |
3 |
|
| 3706(i) |
New |
22.14.6.3 [format.formatter.spec] |
How does std::format work with character arrays of unknown bound? |
No |
3 |
|
| 3944(i) |
SG16 |
22.14.6.3 [format.formatter.spec] |
Formatters converting sequences of char to sequences of wchar_t |
No |
3 |
|
| 3942(i) |
New |
23.2.4 [char.traits.specializations] |
Inconsistent use of const char_type& in standard specializations of std::char_traits |
Yes |
3 |
|
| 2959(i) |
New |
23.2.4.4 [char.traits.specializations.char16.t] |
char_traits<char16_t>::eof is a valid UTF-16 code unit |
No |
3 |
|
| 3457(i) |
New |
23.3.3 [string.view.template] |
*this is not invalidated |
Yes |
3 |
|
| 3451(i) |
New |
23.4.3 [basic.string] |
Inconsistently explicit deduction guides |
Yes |
3 |
|
| 3339(i) |
New |
23.4.3 [basic.string] |
Move-constructed empty-container capacity |
No |
3 |
|
| 3663(i) |
New |
23.4.3.3 [string.cons] |
basic_string(const T&, const Alloc&) turns moves into copies |
Yes |
3 |
|
| 3662(i) |
New |
23.4.3.7.2 [string.append] |
basic_string::append/assign(NTBS, pos, n) suboptimal |
Yes |
3 |
|
| 3837(i) |
New |
23.4.4.5 [string.erasure] |
std::erase_if overloads for non-associative containers should move (and
not copy) their predicate object |
Yes |
3 |
|
| 2238(i) |
Open |
23.5 [c.strings] |
Problematic iterator-pair constructor of containers |
No |
3 |
|
| 3059(i) |
New |
24.2 [container.requirements] |
Wrong requirements for map-like associative container assignment? |
No |
3 |
|
| 2321(i) |
Open |
24.2.2.1 [container.requirements.general] |
Moving containers should (usually) be required to preserve iterators |
Yes |
3 |
|
| 1521(i) |
Open |
24.2.2.1 [container.requirements.general] |
Requirements on internal pointer representations in containers |
Yes |
3 |
|
| 3297(i) |
New |
24.2.4 [sequence.reqmts] |
Useless sequence container requirement |
Yes |
3 |
|
| 2705(i) |
New |
24.2.4 [sequence.reqmts] |
Questionable precondition on Sequence containers a.assign(n, t) |
Yes |
3 |
|
| 2206(i) |
Open |
24.2.4 [sequence.reqmts] |
Inaccuracy in initializer_list constructor requirements |
Yes |
3 |
|
| 3438(i) |
New |
24.2.5.1 [container.node.overview] |
§[container.node.overview] missing multiset/map cases |
No |
3 |
|
| 2215(i) |
Open |
24.2.7 [associative.reqmts] |
(unordered) associative container functors should be CopyConstructible |
Yes |
3 |
|
| 2227(i) |
Open |
24.2.7 [associative.reqmts] |
Stateful comparison objects in associative containers |
No |
3 |
|
| 2844(i) |
Open |
24.2.7 [associative.reqmts] |
Stability of a_uniq.insert(i, j) |
No |
3 |
|
| 3227(i) |
New |
24.2.7 [associative.reqmts] |
Ambiguity issue for extract in ordered and unordered associative containers |
Yes |
3 |
|
| 2362(i) |
New |
24.2.7 [associative.reqmts] |
unique, associative emplace() should not move/copy the mapped_type constructor
arguments when no insertion happens |
No |
3 |
|
| 3691(i) |
New |
24.2.7.1 [associative.reqmts.general] |
Replacement of keys in associative containers |
Yes |
3 |
|
| 3577(i) |
New |
24.2.7.1 [associative.reqmts.general] |
Merging an (unordered) associative container with itself |
No |
3 |
|
| 3578(i) |
New |
24.2.7.1 [associative.reqmts.general] |
Iterator SCARYness in the context of associative container merging |
No |
3 |
|
| 1175(i) |
Open |
24.2.8 [unord.req] |
unordered complexity |
Yes |
3 |
|
| 2198(i) |
Open |
24.2.8 [unord.req] |
max_load_factor(z) makes no strong guarantees, but bans useful behavior |
Yes |
3 |
|
| 2189(i) |
Open |
24.2.8.2 [unord.req.except] |
Throwing swap breaks unordered containers' state |
No |
3 |
|
| 617(i) |
Open |
24.3.7 [array] |
std::array is a sequence that doesn't satisfy the sequence requirements? |
No |
3 |
|
| 3219(i) |
New |
24.3.7.1 [array.overview] |
std::array overview container requirements are incorrect |
Yes |
3 |
|
| 2823(i) |
Open |
24.3.7.1 [array.overview] |
std::array initialization is still not permissive enough |
Yes |
3 |
|
| 3488(i) |
Open |
24.3.7.4 [array.special] |
Is array<const int, 0> swappable or not? |
Yes |
3 |
|
| 2157(i) |
Open |
24.3.7.5 [array.zero] |
How does std::array<T,0> initialization work when T is not default-constructible? |
Yes |
3 |
|
| 3308(i) |
New |
24.3.8.4 [deque.modifiers] |
vector and deque iterator erase invalidates elements even when no change occurs |
Yes |
3 |
|
| 3758(i) |
New |
24.3.11.3 [vector.capacity] |
Element-relocating operations of std::vector and std::deque
should conditionally require Cpp17CopyInsertable in their preconditions |
No |
3 |
|
| 2158(i) |
Open |
24.3.11.3 [vector.capacity] |
Conditional copy/move in std::vector |
Yes |
3 |
|
| 1102(i) |
Open |
24.3.11.3 [vector.capacity] |
std::vector's reallocation policy still unclear |
Yes |
3 |
|
| 3638(i) |
New |
24.3.12 [vector.bool] |
vector<bool>::swap(reference, reference) is useless |
Yes |
3 |
|
| 1422(i) |
Open |
24.3.12 [vector.bool] |
vector<bool> iterators are not random access |
No |
3 |
|
| 3531(i) |
New |
24.4.4.1 [map.overview] |
LWG 3025 broke previous valid code |
Yes |
3 |
|
| 2713(i) |
New |
24.5 [unord] |
More missing allocator-extended constructors for unordered containers |
Yes |
3 |
|
| 3189(i) |
New |
24.6.7 [priority.queue] |
Missing requirement for std::priority_queue |
No |
3 |
|
| 3161(i) |
Open |
24.6.8 [stack] |
Container adapters mandate use of emplace_back but don't require it |
Yes |
3 |
|
| 3615(i) |
New |
25.3.2.1 [incrementable.traits] |
The last specialization of incrementable_traits has wrong operand types |
Yes |
3 |
|
| 3838(i) |
New |
25.3.2.1 [incrementable.traits] |
The last specialization of incrementable_traits is under-constrained |
Yes |
3 |
|
| 3287(i) |
New |
25.3.2.3 [iterator.traits] |
Exposition-only cpp17-input-iterator concept is needlessly complex |
Yes |
3 |
|
| 3890(i) |
New |
25.3.4.4 [iterator.concept.winc] |
ABI issue for integer-class types |
Yes |
3 |
|
| 3716(i) |
New |
25.3.4.11 [iterator.concept.forward] |
§[iterator.concept.forward][forward.iterators] Two different definitions of multi-pass guarantee |
No |
3 |
|
| 2038(i) |
Open |
25.3.5.4 [output.iterators] |
Missing definition for incrementable iterator |
No |
3 |
|
| 2035(i) |
Open |
25.3.5.4 [output.iterators] |
Output iterator requirements are broken |
Yes |
3 |
|
| 3344(i) |
New |
25.4.3 [iterator.operations] |
advance(i, most-negative) and prev(i, most-negative) |
Yes |
3 |
|
| 3439(i) |
New |
25.4.3 [iterator.operations] |
"Distance" template parameter is underspecified |
No |
3 |
|
| 3197(i) |
New |
25.4.3 [iterator.operations] |
std::prev should not require BidirectionalIterator |
Yes |
3 |
|
| 2931(i) |
Open |
25.4.3 [iterator.operations] |
Missed optimization opportunity with single-argument std::next |
No |
3 |
|
| 3623(i) |
New |
25.5.1.1 [reverse.iterators.general] |
Uses of std::reverse_iterator with containers should not require manually including <iterator> |
Yes |
3 |
|
| 2595(i) |
New |
25.5.1.2 [reverse.iterator] |
reverse_iterator::operator[]'s return type revisited |
Yes |
3 |
|
| 3602(i) |
New |
25.5.1.4 [reverse.iter.cons] |
reverse_iterator's converting assignment is overconstrained |
Yes |
3 |
|
| 3725(i) |
New |
25.5.1.6 [reverse.iter.elem] |
reverse_iterator::operator-> should not use prev for non-pointer iterators |
Yes |
3 |
|
| 3783(i) |
New |
25.5.5.1 [common.iterator] |
views::common may not be a range adaptor object |
Yes |
3 |
|
| 3748(i) |
New |
25.5.5.6 [common.iter.cmp] |
common_iterator and counted_iterator' operator- are missing cast to return type |
Yes |
3 |
|
| 2366(i) |
New |
25.6.4 [istreambuf.iterator] |
istreambuf_iterator end-of-stream equality |
Yes |
3 |
|
| 3188(i) |
New |
25.6.4 [istreambuf.iterator] |
istreambuf_iterator::pointer should not be unspecified |
Yes |
3 |
|
| 3108(i) |
New |
25.6.4.2 [istreambuf.iterator.proxy] |
istreambuf_iterator::proxy::operator* should be const |
Yes |
3 |
|
| 3537(i) |
New |
25.7 [iterator.range] |
§[iterator.range] Missing noexcept for std::rbegin/rend for arrays and
initializer_list |
No |
3 |
|
| 3913(i) |
SG9 |
26.2 [ranges.syn] |
ranges::const_iterator_t<range R> fails to accept arrays of unknown bound |
Yes |
3 |
|
| 3907(i) |
New |
26.6 [range.factories] |
Can iterator types of range adaptors and range factories be SCARY? |
No |
3 |
|
| 3614(i) |
New |
26.6.4.2 [range.iota.view] |
iota_view::size and the most negative signed integer values |
Yes |
3 |
|
| 3846(i) |
New |
26.6.4.3 [range.iota.iterator] |
iota_view::iterator::operator- is overconstrained |
Yes |
3 |
|
| 3609(i) |
New |
26.6.4.4 [range.iota.sentinel] |
std::ranges::iota_view<int, long> has non-subtractable iterator and
sentinel types |
Yes |
3 |
|
| 3763(i) |
New |
26.6.5.3 [range.repeat.iterator] |
Should range adaptor iterators only provide iterator_category when its
difference_type is not an integer-class type? |
Yes |
3 |
|
| 3679(i) |
LEWG |
26.6.6 [range.istream] |
Is <ranges> sufficient for istream_view? |
No |
3 |
|
| 3489(i) |
New |
26.6.6.3 [range.istream.iterator] |
Improve istream_view wording |
Yes |
3 |
|
| 3829(i) |
New |
26.7.7.2 [range.as.rvalue.view] |
as_rvalue_view::end should improve non-common case |
Yes |
3 |
|
| 3730(i) |
New |
26.7.12.2 [range.drop.view] |
std::ranges::drop_view may have different size type from its underlying view |
Yes |
3 |
|
| 3873(i) |
New |
26.7.15.2 [range.join.with.view] |
join_with_view's const begin is underconstrained |
No |
3 |
|
| 3852(i) |
New |
26.7.15.3 [range.join.with.iterator] |
join_with_view::iterator's iter_move and iter_swap should be conditionally noexcept |
Yes |
3 |
|
| 3599(i) |
New |
26.7.16.2 [range.lazy.split.view] |
The const overload of lazy_split_view::begin should be constrained by const Pattern |
Yes |
3 |
|
| 3685(i) |
New |
26.7.16.2 [range.lazy.split.view] |
In lazy_split_view, CTAD doesn't work when given an input_range input and a tiny-range pattern |
Yes |
3 |
|
| 3686(i) |
New |
26.7.16.3 [range.lazy.split.outer] |
In lazy_split_view, comparing a default-constructed outer-iterator or
inner-iterator with std::default_sentinel results in null pointer dereference |
Yes |
3 |
|
| 3830(i) |
New |
26.7.20.2 [range.reverse.view] |
reverse_view should not cache when ranges::next has constant time complexity |
Yes |
3 |
|
| 3832(i) |
New |
26.7.22.3 [range.elements.iterator] |
Missing change for element_view::iterator in LWG 3798 |
Yes |
3 |
|
| 3919(i) |
New |
26.7.23 [range.enumerate] |
enumerate_view may invoke UB for sized common non-forward underlying ranges |
Yes |
3 |
|
| 3731(i) |
New |
26.7.24.2 [range.zip.view] |
zip_view and adjacent_view are underconstrained |
Yes |
3 |
|
| 3899(i) |
New |
26.8.5 [coro.generator.promise] |
co_yielding elements of an lvalue generator is unnecessarily inefficient |
Yes |
3 |
|
| 3900(i) |
New |
26.8.5 [coro.generator.promise] |
The allocator_arg_t overloads of generator::promise_type::operator new
should not be constrained |
Yes |
3 |
|
| 2963(i) |
New |
27 [algorithms] |
Algorithms with underspecified iterator requirements |
No |
3 |
|
| 1238(i) |
Open |
27 [algorithms] |
Defining algorithms taking iterator for range |
No |
3 |
|
| 3049(i) |
Open |
27.2 [algorithms.requirements] |
Missing wording allowing algorithms to use copies of function objects as substitutes for their parameters |
Yes |
3 |
|
| 3793(i) |
New |
27.6.5 [alg.foreach] |
Requirements for some algorithms' Size template parameters are unclear |
No |
3 |
|
| 3089(i) |
New |
27.7.1 [alg.copy] |
copy_n should require non-overlapping ranges |
Yes |
3 |
|
| 2471(i) |
Open |
27.7.1 [alg.copy] |
copy_n's number of InputIterator increments unspecified |
No |
3 |
|
| 2267(i) |
New |
27.8.2.4 [partial.sort.copy] |
partial_sort_copy underspecified for ranges of two different types |
No |
3 |
|
| 3534(i) |
LEWG |
27.8.7.4 [set.intersection] |
ranges::set_intersection and ranges::set_difference algorithm requirements are too strict |
Yes |
3 |
|
| 3029(i) |
Open |
27.8.8.3 [pop.heap] |
pop_heap over-constrains input |
Yes |
3 |
|
| 3487(i) |
New |
27.10 [numeric.ops] |
Missing precondition on input and output aliasing of [numeric.ops] |
No |
3 |
|
| 3463(i) |
New |
27.10.11 [transform.inclusive.scan] |
Incorrect requirements for transform_inclusive_scan without initial value |
Yes |
3 |
|
| 3063(i) |
New |
27.11 [specialized.algorithms] |
Parallel algorithms in <memory> are underspecified |
No |
3 |
|
| 3628(i) |
New |
27.11 [specialized.algorithms] |
"Effects: Equivalent to:" and uninitialized memory algorithms |
No |
3 |
|
| 3064(i) |
New |
27.11 [specialized.algorithms] |
How do uninitialized memory algorithms obtain pointer without undefined behavior? |
No |
3 |
|
| 3647(i) |
New |
27.11.2 [special.mem.concepts] |
nothrow-input-iterator constraints should not mention copying |
Yes |
3 |
|
| 3918(i) |
LEWG |
27.11.6 [uninitialized.move] |
std::uninitialized_move/_n and guaranteed copy elision |
Yes |
3 |
|
| 3888(i) |
New |
27.11.8 [specialized.construct] |
Most ranges uninitialized memory algorithms are underconstrained |
Yes |
3 |
|
| 3889(i) |
New |
27.11.9 [specialized.destroy] |
std::(ranges::)destroy_at should destroy array elements in the decreasing index order |
Yes |
3 |
|
| 3934(i) |
New |
28.4.3 [complex] |
std::complex<T>::operator=(const T&) has no specification |
Yes |
3 |
|
| 2714(i) |
New |
28.4.6 [complex.ops] |
complex stream extraction underspecified |
Yes |
3 |
|
| 2846(i) |
New |
28.4.9 [cmplx.over] |
Undefined phrase "effectively cast" |
Yes |
3 |
|
| 3809(i) |
New |
28.5.4.4 [rand.eng.sub] |
Is std::subtract_with_carry_engine<uint16_t> supposed to work? |
Yes |
3 |
|
| 3402(i) |
New |
28.5.9.3.4 [rand.dist.bern.negbin] |
Wording for negative_binomial_distribution is unclear as a consequence of LWG 2406 resolution |
No |
3 |
|
| 2847(i) |
New |
28.7.1 [cmath.syn] |
sin(float) should call sinf(float) |
No |
3 |
|
| 3093(i) |
New |
28.7.2 [c.math.abs] |
LWG 2294/2192 missed a std::abs overload |
No |
3 |
|
| 3172(i) |
New |
28.7.3 [c.math.hypot3] |
3-arg std::hypot is underspecified compared to the 2-arg overload |
Yes |
3 |
|
| 3066(i) |
New |
28.7.6 [sf.cmath] |
"report a domain error" in [sf.cmath]/1 is underspecified |
No |
3 |
|
| 3090(i) |
New |
29.5.2 [time.duration.cons] |
What is §[time.duration.cons]p4's "no overflow is induced in the conversion" intended to mean? |
Yes |
3 |
|
| 3503(i) |
New |
29.5.8 [time.duration.cast] |
chrono::ceil has surprising requirement |
No |
3 |
|
| 2383(i) |
Open |
29.5.9 [time.duration.literals] |
Overflow cannot be ill-formed for chrono::duration integer literals |
No |
3 |
|
| 3856(i) |
New |
29.12 [time.format] |
Unclear which conversion specifiers are valid for each chrono type |
Yes |
3 |
|
| 3921(i) |
New |
29.12 [time.format] |
Is std::chrono::duration<std::int64_t, std::ratio<INT64_MAX - 1, INT64_MAX>>{40}
required to be correctly formatted? |
No |
3 |
|
| 3831(i) |
New |
29.12 [time.format] |
Two-digit formatting of negative year is ambiguous |
Yes |
3 |
|
| 3844(i) |
Open |
29.12 [time.format] |
Non-numeric formats for negative durations |
Yes |
3 |
|
| 3353(i) |
New |
30.3.1 [locale] |
locale's copy assignment operator should return locale& |
Yes |
3 |
|
| 3767(i) |
Open |
30.3.1.2.1 [locale.category] |
codecvt<charN_t, char8_t, mbstate_t> incorrectly added to locale |
Yes |
3 |
|
| 3337(i) |
New |
30.4.2.5.3 [locale.codecvt.virtuals] |
What is "is initialized" supposed to mean? |
No |
3 |
|
| 3689(i) |
New |
30.4.3.2.3 [facet.num.get.virtuals] |
num_get overflow determination unclear and incorrect |
No |
3 |
|
| 2117(i) |
Open |
30.4.3.3.3 [facet.num.put.virtuals] |
ios_base manipulators should have showgrouping/noshowgrouping |
No |
3 |
|
| 2702(i) |
New |
30.4.3.3.3 [facet.num.put.virtuals] |
num_put::do_put(..., bool) performs ill-formed do_put call |
No |
3 |
|
| 2703(i) |
New |
30.4.3.3.3 [facet.num.put.virtuals] |
No provision for fill-padding when boolalpha is set |
No |
3 |
|
| 3275(i) |
New |
30.4.6.2.3 [locale.time.get.virtuals] |
Why does time_get::do_get require a valid pointer when none of the others do? |
Yes |
3 |
|
| 2983(i) |
New |
30.4.7.3.2 [locale.money.put.virtuals] |
money_put::do_put underspecified |
Yes |
3 |
|
| 2691(i) |
New |
30.4.7.4 [locale.moneypunct] |
money_base::space and do_put: U+0020 versus fill |
Yes |
3 |
|
| 423(i) |
Open |
31 [input.output] |
Effects of negative streamsize in iostreams |
Yes |
3 |
|
| 3696(i) |
New |
31.2.2 [stream.types] |
"Basic integral types" should not be used |
No |
3 |
|
| 2675(i) |
New |
31.5.2.7 [ios.base.callback] |
register_callback can fail |
No |
3 |
|
| 2504(i) |
New |
31.6.3 [streambuf] |
basic_streambuf is not an abstract class |
No |
3 |
|
| 3658(i) |
New |
31.6.3.3.5 [streambuf.pub.put] |
basic_streambuf::sputn is both overspecified and underspecified |
Yes |
3 |
|
| 2497(i) |
New |
31.7.6.2.4 [ostream.sentry] |
Use of uncaught_exception() |
Yes |
3 |
|
| 3501(i) |
New |
31.7.6.5 [ostream.manip] |
basic_syncbuf-related manipulators refer to some Allocator without defining it |
Yes |
3 |
|
| 3937(i) |
New |
31.7.7 [std.manip] |
I/O manipulators should be specified in terms of base classes |
No |
3 |
|
| 2984(i) |
New |
31.7.8 [ext.manip] |
put_money(99) is unnecessarily undefined |
Yes |
3 |
|
| 3309(i) |
New |
31.8 [string.streams] |
Is <ios> implicitly #included by <sstream>, <fstream> etc.? |
No |
3 |
|
| 3097(i) |
New |
31.8.2.5 [stringbuf.virtuals] |
basic_stringbuf seekoff effects trigger undefined behavior and have contradictory returns |
No |
3 |
|
| 3497(i) |
New |
31.11.2.4 [syncstream.syncbuf.members] |
Postconditions for basic_syncbuf::emit() |
No |
3 |
|
| 3496(i) |
New |
31.11.2.4 [syncstream.syncbuf.members] |
What does "uniquely associated" mean for basic_syncbuf::emit()? |
No |
3 |
|
| 3098(i) |
New |
31.12.6.5.9 [fs.path.decompose] |
Misleading example for filesystem::path::filename() |
Yes |
3 |
|
| 3699(i) |
New |
31.12.6.5.11 [fs.path.gen] |
lexically_relative on UNC drive paths (\\?\C:\...) results in a default-constructed value |
No |
3 |
|
| 3794(i) |
New |
31.12.6.6 [fs.path.itr] |
std::filesystem::path::iterator::reference should be allowed to be std::filesystem::path |
Yes |
3 |
|
| 2947(i) |
New |
31.12.8.1 [fs.enum.path.format] |
Clarify several filesystem terms |
No |
3 |
|
| 3078(i) |
New |
31.12.10 [fs.class.directory.entry] |
directory_entry, directory_iterator and recursive_directory_iterator perform needless path copies |
No |
3 |
|
| 3668(i) |
New |
31.12.11.2 [fs.dir.itr.members] |
[recursive_]directory_iterator constructors refer to undefined options |
Yes |
3 |
|
| 3056(i) |
New |
31.12.13.5 [fs.op.copy.file] |
copy_file() copies which attributes? |
Yes |
3 |
|
| 3744(i) |
New |
31.12.13.6 [fs.op.copy.symlink] |
copy_symlink(junction, new_symlink)'s behavior is unclear |
No |
3 |
|
| 2490(i) |
New |
32 [re] |
<regex> needs lots of noexcept |
No |
3 |
|
| 3606(i) |
New |
32.2 [re.req] |
Missing regex_traits::locale_type requirements |
No |
3 |
|
| 2431(i) |
New |
32.2 [re.req] |
Missing regular expression traits requirements |
No |
3 |
|
| 2331(i) |
Open |
32.4.2 [re.synopt] |
regex_constants::collate's effects are inaccurately summarized |
Yes |
3 |
|
| 3605(i) |
New |
32.4.3 [re.matchflag] |
regex_constants::match_prev_avail is underspecified |
No |
3 |
|
| 2338(i) |
Open |
32.6 [re.traits] |
§[re.traits]/7 expects of locale facets something not guaranteed by [locale.facet]/4 |
Yes |
3 |
|
| 3261(i) |
New |
32.7 [re.regex] |
regex components' noexcept annotations appear broken for POCMA or throwing
BidirectionalIterator |
No |
3 |
|
| 3341(i) |
New |
32.7.2 [re.regex.construct] |
basic_regex range constructor: Missing requirements for iterator types |
No |
3 |
|
| 3603(i) |
New |
32.7.2 [re.regex.construct] |
Matching of null characters by regular expressions is underspecified |
No |
3 |
|
| 3604(i) |
New |
32.7.2 [re.regex.construct] |
What is the effect of an invalid value of type syntax_option_type? |
No |
3 |
|
| 2137(i) |
Open |
32.7.3 [re.regex.assign] |
Misleadingly constrained post-condition in the presence of exceptions |
Yes |
3 |
|
| 3126(i) |
New |
32.8 [re.submatch] |
There's no std::sub_match::compare(string_view) overload |
Yes |
3 |
|
| 2216(i) |
New |
32.10.4 [re.alg.replace] |
regex_replace(basic_string) allocator handling |
No |
3 |
|
| 2220(i) |
Open |
32.11.2.3 [re.tokiter.comp] |
Under-specification of operator== for regex_token_iterator |
Yes |
3 |
|
| 2987(i) |
New |
32.12 [re.grammar] |
Relationship between traits_inst.lookup_collatename and the regex FSM is underspecified with
regards to ClassAtomCollatingElement |
No |
3 |
|
| 2819(i) |
New |
33.2.5 [thread.req.lockable] |
Unspecified Return type: elements |
Yes |
3 |
|
| 3499(i) |
New |
33.2.5.4 [thread.req.lockable.timed] |
Timed lockable and mutex requirements are imprecise about duration and time_point |
No |
3 |
|
| 3924(i) |
New |
33.3.1 [thread.stoptoken.intro] |
Stop token data race avoidance requirements unclear |
Yes |
3 |
|
| 3516(i) |
New |
33.4.3.2 [thread.thread.id] |
thread::id spaceship may be inconsistent with equality |
Yes |
3 |
|
| 3475(i) |
New |
33.4.3.3 [thread.thread.constr] |
std::thread's constructor needs to be able to report general memory allocation failures |
Yes |
3 |
|
| 3633(i) |
New |
33.5 [atomics] |
Atomics are copy constructible and copy assignable from volatile atomics |
Yes |
3 |
|
| 3220(i) |
New |
33.5.2 [atomics.syn] |
P0558 broke conforming C++14 uses of atomic shared_ptr |
Yes |
3 |
|
| 3941(i) |
SG1 |
33.5.4 [atomics.order] |
§[atomics.order] inadvertently prohibits widespread implementation techniques |
No |
3 |
|
| 3263(i) |
New |
33.5.6 [atomics.wait] |
Atomic waiting function calls should only be unblocked once |
Yes |
3 |
|
| 3485(i) |
SG1 |
33.5.7 [atomics.ref.generic] |
atomic_ref safety should be based on operations that "potentially conflict" rather than lifetime |
No |
3 |
|
| 3409(i) |
New |
33.5.7.2 [atomics.ref.ops] |
Too lax description of atomic_ref<T>::required_alignment |
Yes |
3 |
|
| 3417(i) |
SG1 |
33.5.8.2 [atomics.types.operations] |
Missing volatile atomic deprecations |
Yes |
3 |
|
| 3047(i) |
New |
33.5.8.3 [atomics.types.int] |
atomic compound assignment operators can cause undefined behavior when corresponding
fetch_meow members don't |
Yes |
3 |
|
| 3906(i) |
New |
33.5.8.5 [atomics.types.pointer] |
"Undefined address" is undefined |
No |
3 |
|
| 3418(i) |
New |
33.5.9 [atomics.nonmembers] |
Deprecated free functions in <atomic> |
Yes |
3 |
|
| 3343(i) |
New |
33.7.3 [thread.condition.nonmember] |
Ordering of calls to unlock() and notify_all() in Effects element of notify_all_at_thread_exit() should be reversed |
Yes |
3 |
|
| 3504(i) |
New |
33.7.4 [thread.condition.condvar] |
condition_variable::wait_for is overspecified |
No |
3 |
|
| 3898(i) |
New |
33.9.3.3 [thread.barrier.class] |
Possibly unintended preconditions for completion functions of std::barrier |
Yes |
3 |
|
| 2530(i) |
Open |
33.10.5 [futures.state] |
Clarify observable side effects of releasing a shared state |
No |
3 |
|
| 2532(i) |
Open |
33.10.6 [futures.promise] |
Satisfying a promise at thread exit |
Yes |
3 |
|
| 3582(i) |
New |
33.10.9 [futures.async] |
Unclear where std::async exceptions are handled |
Yes |
3 |
|
| 2507(i) |
New |
D.26 [depr.locale.stdcvt] |
codecvt_mode should be a bitmask type |
No |
3 |
|
| 3840(i) |
Open |
D.29 [depr.fs.path.factory] |
filesystem::u8path should be undeprecated |
Yes |
3 |
|
| 3812(i) |
New |
3.2.2.5 [fund.ts.v3::propagate_const.const_observers] |
[fund.ts.v3] Incorrect constraint on propagate_const conversion function |
Yes |
3 |
|
| 3167(i) |
Open |
5.2 [fund.ts.v3::memory.observer.ptr] |
[fund.ts.v3] Does observer_ptr support function types? |
No |
3 |
|
| 3357(i) |
Open |
99 [fund.ts.v3::rand.util.randint] |
[fund.ts.v3] default_random_engine is overspecified for per-thread engine |
Yes |
3 |
|
| 3124(i) |
New |
13.7.5 [networking.ts::async.exec.ctx.globals] |
[networking.ts] Unclear how execution_context is intended to store services |
Yes |
3 |
|
| 3021(i) |
New |
16.2.2 [networking.ts::buffer.reqmts.constbuffersequence] |
[networking.ts] Relax pointer equivalence requirement for ConstBufferSequence |
Yes |
3 |
|
| 3027(i) |
New |
16.2.4 [networking.ts::buffer.reqmts.dynamicbuffer] |
[networking.ts] DynamicBuffer prepare exception specification |
Yes |
3 |
|
| 3072(i) |
New |
16.2.4 [networking.ts::buffer.reqmts.dynamicbuffer] |
[networking.ts] DynamicBuffer object lifetimes underspecified |
Yes |
3 |
|
| 3073(i) |
New |
17 [networking.ts::buffer.stream] |
[networking.ts] (async_)read and (async_)write
don't support DynamicBuffer lvalues |
Yes |
3 |
|
| 3445(i) |
LEWG |
19.2.1 [networking.ts::socket.iostream.cons] |
[networking.ts] net::basic_socket_istream::connect should be constrained |
Yes |
3 |
|