P0627R6: Function to mark unreachable code by AlexGuteniev · Pull Request #2526 · microsoft/STL

Alex Guteniev (AlexGuteniev) · 2022-02-07T18:26:11Z

Resolves #2531

Co-authored-by: Casey Carter <cartec69@gmail.com>

Casey Carter (CaseyCarter) · 2022-02-08T19:20:59Z

+
+[[noreturn]] inline void unreachable() noexcept /* strengthened */ {
+#ifdef _DEBUG
+    _CSTD abort();


I'm not sure this is a great idea. abort is well-defined, and therefore doesn't poison the code path leading to it with undefined behavior. That seems like a very subtle difference that folks may not be expecting when they flip the debug/release switch. (I'm going to call this "Comment" rather than "Request change" to get more reviewers' opinions.)

I observed that __assume(false) does provide UB in debug. Up to this

int main() { __assume(false); // does not necessarily return zero, even in /Od }

Normally I'd except debug mode to catch any UBs that can be caught.

Also look into ATLASSUME for a precedent.

Alex Guteniev (@AlexGuteniev) I really like your idea of using _STL_UNREACHABLE here as we're already using it in product code for this exact purpose (in <format> and <variant>).

Casey Carter (@CaseyCarter) I share your concern about well-defined behavior. How would you feel about:

_STL_UNREACHABLE; #ifdef _DEBUG _CSTD abort(); #endif // _DEBUG

I think this is the best of both worlds. +1 from me.

_STL_UNREACHABLE; #ifdef _DEBUG _CSTD abort(); #endif // _DEBUG

Is not really a good option.
By having abort(); after _STL_UNREACHABLE; you ask for UB, and compiler may implement the UB by not actually calling abort, and in /O2 it really does! _DEBUG may be used with /O2, and I don't think #pragma optimize is good here.

If you want to enjoy UB even in _DEBUG here, I'd rather leave just __assume(false)/__builtin_unreachable(), wrapped in _STL_UNREACHABLE;

A possible way to be able to debug the unreachable is to guard the original _STL_UNREACHABLE in #ifndef _STL_UNREACHABLE. This way a user may do #define _STL_UNREACHABLE abort(), #define _STL_UNREACHABLE __debugbreak() or #define _STL_UNREACHABLE __ud2(). Still it is arguable place for a customization; if we decide to embrace the UB, let's have it.

By having abort(); after _STL_UNREACHABLE; you ask for UB, and compiler may implement the UB by not actually calling abort, and in /O2 it really does!

I'm fine with this, it is working as intended. The goal of putting abort here is not to define the behavior, it's to increase the likelihood that the outcome of the UB is to crash near to the occurrence of unreachable in the user's code. That outcome isn't deterministic and can't be made deterministic without breaking the intended functionality. I don't agree that the lack of determinism makes the abort call useless.

A possible way to be able to debug the unreachable is to guard the original _STL_UNREACHABLE in #ifndef _STL_UNREACHABLE.

If people sometimes want to not use unreachable they can define their own macros.

Please review the comment I added next to abort().
(I don't think we can leave the code with deliberate UB without a comment).

Also what exactly the semantic of _STL_UNREACHABLE?
I mean if this is precondition and not internal invariant, maybe the abort() should rather be embedded there rather than here?

Alex Guteniev (AlexGuteniev) · 2022-02-08T19:48:52Z

Maybe should reuse:

STL/stl/inc/yvals_core.h

Lines 1500 to 1504 in ee6a79e

    
           #ifdef __clang__ 
        
           #define _STL_UNREACHABLE __builtin_unreachable() 
        
           #else // ^^^ clang ^^^ / vvv other vvv 
        
           #define _STL_UNREACHABLE __assume(false) 
        
           #endif // __clang__

Stephan T. Lavavej (StephanTLavavej) · 2022-02-08T23:20:18Z

+
+[[noreturn]] inline void unreachable() noexcept /* strengthened */ {
+#ifdef _DEBUG
+    _CSTD abort();


Alex Guteniev (@AlexGuteniev) I really like your idea of using _STL_UNREACHABLE here as we're already using it in product code for this exact purpose (in <format> and <variant>).

Casey Carter (@CaseyCarter) I share your concern about well-defined behavior. How would you feel about:

_STL_UNREACHABLE; #ifdef _DEBUG _CSTD abort(); #endif // _DEBUG

Stephan T. Lavavej (StephanTLavavej) · 2022-02-11T13:27:45Z

I'm mirroring this to the MSVC-internal repo - please notify me if any further changes are pushed.

Stephan T. Lavavej (StephanTLavavej) · 2022-02-12T02:22:59Z

Thanks for reaching this unreachable feature so quickly! 😹 🚀 🎉

unreachable code

ab32fa7

Alex Guteniev (AlexGuteniev) requested a review from a team as a code owner February 7, 2022 18:26

(c)

7b03bb9

Alex Guteniev (AlexGuteniev) commented Feb 7, 2022

View reviewed changes

Comment thread stl/inc/utility Outdated

inline

27d7a8c

Casey Carter (CaseyCarter) changed the title ~~unreachable code~~ P0672R6: Function to mark unreachable code Feb 7, 2022

Casey Carter (CaseyCarter) added the cxx23 C++23 feature label Feb 7, 2022

Casey Carter (CaseyCarter) suggested changes Feb 7, 2022

View reviewed changes

Comment thread stl/inc/utility Outdated

Comment thread stl/inc/yvals_core.h

Comment thread stl/inc/utility Outdated

Comment thread tests/std/tests/VSO_0157762_feature_test_macros/test.compile.pass.cpp Outdated

Comment thread tests/std/tests/P0627R6_unreachable/test.cpp Outdated

Alex Guteniev (AlexGuteniev) and others added 2 commits February 7, 2022 20:46

Apply suggestions from code review

6317604

Co-authored-by: Casey Carter <cartec69@gmail.com>

[[noreturn]] , noexcept /* strengthened */

b234d9b

Alex Guteniev (AlexGuteniev) requested a review from Casey Carter (CaseyCarter) February 7, 2022 18:52

Casey Carter (CaseyCarter) approved these changes Feb 7, 2022

View reviewed changes

This comment was marked as outdated.

Sign in to view

Casey Carter (CaseyCarter) linked an issue Feb 8, 2022 that may be closed by this pull request

P0627R6 unreachable() #2531

Closed

Stephan T. Lavavej (StephanTLavavej) changed the title ~~P0672R6: Function to mark unreachable code~~ P0627R6: Function to mark unreachable code Feb 8, 2022

__builtin_unreachable

d183d52

Daniel Marshall (SuperWig) reviewed Feb 8, 2022

View reviewed changes

Comment thread tests/std/tests/P0627R6_unreachable/test.cpp

debugging experience

04e36c8

Alex Guteniev (AlexGuteniev) requested a review from Casey Carter (CaseyCarter) February 8, 2022 10:49

it wouldn't hurt

6d9de6f

Casey Carter (CaseyCarter) reviewed Feb 8, 2022

View reviewed changes

Stephan T. Lavavej (StephanTLavavej) reviewed Feb 8, 2022

View reviewed changes

Casey Carter (CaseyCarter) self-requested a review February 9, 2022 00:12

Alex Guteniev (AlexGuteniev) added 3 commits February 9, 2022 08:52

Use _STL_UNREACHABLE and don't call abort()

b564b0d

unreachable()

e9f57df

char

ea8eb26

Alex Guteniev (AlexGuteniev) requested a review from Stephan T. Lavavej (StephanTLavavej) February 9, 2022 07:25

Alex Guteniev (AlexGuteniev) added 2 commits February 9, 2022 21:31

hail undefined behavior!

68dc5a7

to be (or not to be)

e64e511

Casey Carter (CaseyCarter) approved these changes Feb 9, 2022

View reviewed changes

Stephan T. Lavavej (StephanTLavavej) approved these changes Feb 10, 2022

View reviewed changes

Stephan T. Lavavej (StephanTLavavej) self-assigned this Feb 11, 2022

Stephan T. Lavavej (StephanTLavavej) merged commit b6b9778 into microsoft:main Feb 12, 2022

Alex Guteniev (AlexGuteniev) deleted the cantreach branch February 12, 2022 05:16

Agustín Bergé (K-ballo) mentioned this pull request Mar 10, 2026

<utility>: unreachable does not abort in debug builds #6151

Closed

Alex Guteniev (AlexGuteniev) mentioned this pull request Mar 14, 2026

<utility>: make unreachable() call the doom function in debug mode #6159

Merged

Uh oh!

Conversation

Alex Guteniev (AlexGuteniev) commented Feb 7, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

This comment was marked as outdated.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Alex Guteniev (AlexGuteniev) Feb 8, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Alex Guteniev (AlexGuteniev) commented Feb 8, 2022

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Stephan T. Lavavej (StephanTLavavej) commented Feb 11, 2022

Uh oh!

Stephan T. Lavavej (StephanTLavavej) commented Feb 12, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Alex Guteniev (AlexGuteniev) commented Feb 7, 2022 •

edited

Loading

Alex Guteniev (AlexGuteniev) Feb 8, 2022 •

edited

Loading