Requires clauses in pfn conversions for generic lambdas

← All discussions

Project / Subsystem

gcc / gcc

Date

2026-09-12

Proposer

Ivan Sokolov <sugrob9000aero@gmail.com>

Source type

public_inbox

Consensus

Proposed

Sentiment

6/10

Technical tradeoffs

• Enforcing constraints on lambda-to-function-pointer conversions improves type safety and consistency but might break existing code that relies on GCC's current lenient behavior.
• Aligning GCC's behavior with Clang's for this feature could improve cross-compiler compatibility.
• Potential standard defect investigation may lead to further language clarifications or modifications.

View thread →

All attributes

project: gcc
subsystem: gcc
patch_id: —
discussion_id: 20260912030100.325efc83@aabbin
source_type: public_inbox
title: Requires clauses in pfn conversions for generic lambdas
headline: Requires clauses in pfn conversions for generic lambdas
tldr: Proposes adding requires clause evaluation to generic lambda conversions to function pointers, matching Clang's behavior.
proposer: Ivan Sokolov <sugrob9000aero@gmail.com>
consensus: Proposed
outcome: proposed
sentiment_score: 6
technical_tradeoffs: • Enforcing constraints on lambda-to-function-pointer conversions improves type safety and consistency but might break existing code that relies on GCC's current lenient behavior.
• Aligning GCC's behavior with Clang's for this feature could improve cross-compiler compatibility.
• Potential standard defect investigation may lead to further language clarifications or modifications.
series_id: —
series_role: standalone
series_parts: []
tags: • c++
• lambda
• compiler
• standard
• gcc
url: https://inbox.sourceware.org/gcc/20260912030100.325efc83@aabbin
bugzilla_url: —
date: 2026-09-12T00:00:00.000Z

Requires clauses in pfn conversions for generic lambdas

This discussion proposes adding support for ‘requires’ clauses in C++ generic lambdas when converting them to function pointers. Currently, GCC does not apply these constraints, leading to ambiguities or internal compiler errors in certain cases, unlike Clang which appears to handle them correctly. The author offers to provide a patch if the change is deemed desirable for GCC and potentially the standard.