This page is a snapshot from the LWG issues list, see the Library Active Issues List for more information and the meaning of WP status.
Section: 20.3.4.3 [inout.ptr.t] Status: WP Submitter: JeanHeyd Meneide Opened: 2021-09-16 Last modified: 2022-11-17
Priority: 1
View all other issues in [inout.ptr.t].
View all issues with WP status.
Discussion:
More succinctly, the model for std::out_ptr_t and std::inout_ptr_t both have a conditional release in their destructors as part of their specification (20.3.4.1 [out.ptr.t] p8) (Option #2 below). But, if the wording is followed, they have an unconditional release in their constructor (Option #1 below). This is not exactly correct and can cause issues with double-free in inout_ptr in particular.
Consider a function MyFunc that sets rawPtr to nullptr when freeing an old value and deciding not to produce a new value, as shown below:// Option #1: auto uptr = std::make_unique<BYTE[]>(25); auto rawPtr = uptr.get(); uptr.release(); // UNCONDITIONAL MyFunc(&rawPtr); If (rawPtr) { uptr.reset(rawPtr); } // Option #2: auto uptr = std::make_unique<BYTE[]>(25); auto rawPtr = uptr.get(); MyFunc(&rawPtr); If (rawPtr) { uptr.release(); // CONDITIONAL uptr.reset(rawPtr); }
This is no problem if the implementation selected Option #1 (release in the constructor), but results in double-free if the implementation selected option #2 (release in the destructor).
As the paper author and after conferring with others, the intent was that the behavior was identical and whether a choice between the constructor or destructor is made. The reset should be unconditional, at least for inout_ptr_t. Suggested change for the ~inout_ptr_t destructor text is to remove the "if (p) { ... }" wrapper from around the code in 20.3.4.3 [inout.ptr.t] p11.[2021-09-24; Reflector poll]
Set priority to 1 after reflector poll.
Previous resolution [SUPERSEDED]:
This wording is relative to N4892.
Modify 20.3.4.3 [inout.ptr.t] as indicated:
~inout_ptr_t();-9- Let SP be POINTER_OF_OR(Smart, Pointer) (20.2.1 [memory.general]).
-10- Let release-statement be s.release(); if an implementation does not call s.release() in the constructor. Otherwise, it is empty. -11- Effects: Equivalent to:
(11.1) —
if (p) {apply([&](auto&&... args) { s = Smart( static_cast<SP>(p), std::forward<Args>(args)...); }, std::move(a));}if is_pointer_v<Smart> is true;
(11.2) — otherwise,
if (p) {apply([&](auto&&... args) { release-statement; s.reset(static_cast<SP>(p), std::forward<Args>(args)...); }, std::move(a));}if the expression s.reset(static_cast<SP>(p), std::forward<Args>(args)...) is well-formed;
(11.3) — otherwise,
if (p) {apply([&](auto&&... args) { release-statement; s = Smart(static_cast<SP>(p), std::forward<Args>(args)...); }, std::move(a));}if is_constructible_v<Smart, SP, Args...> is true;
(11.4) — otherwise, the program is ill-formed.
[2021-10-28; JeanHeyd Meneide provides improved wording]
[2022-08-24 Approved unanimously in LWG telecon.]
[2022-11-12 Approved at November 2022 meeting in Kona. Status changed: Voting → WP.]
Proposed resolution:
This wording is relative to N4901.
Modify 20.3.4.3 [inout.ptr.t] as indicated:
~inout_ptr_t();-9- Let SP be POINTER_OF_OR(Smart, Pointer) (20.2.1 [memory.general]).
-10- Let release-statement be s.release(); if an implementation does not call s.release() in the constructor. Otherwise, it is empty. -11- Effects: Equivalent to:
(11.1) —
if (p) { apply([&](auto&&... args) { s = Smart( static_cast<SP>(p), std::forward<Args>(args)...); }, std::move(a)); }if is_pointer_v<Smart> is true;
(11.2) — otherwise,
release-statement; if (p) { apply([&](auto&&... args) {release-statement;s.reset(static_cast<SP>(p), std::forward<Args>(args)...); }, std::move(a)); }if the expression s.reset(static_cast<SP>(p), std::forward<Args>(args)...) is well-formed;
(11.3) — otherwise,
release-statement; if (p) { apply([&](auto&&... args) {release-statement;s = Smart(static_cast<SP>(p), std::forward<Args>(args)...); }, std::move(a)); }if is_constructible_v<Smart, SP, Args...> is true;
(11.4) — otherwise, the program is ill-formed.