Re: Is it time to take the middle-end stringop/array warnings out of -Wall?

Project / Subsystem

gcc / gcc

Date

2026-05-22

Proposer

Richard Biener <richard.guenther@gmail.com>

Source type

public_inbox

Consensus

Proposed

Sentiment

/10

Technical tradeoffs

  • Enabling diagnostic-paths may increase compilation time.
  • May require changes to GCC's internal data structures to efficiently track and display preconditions.

All attributes

project
gcc
subsystem
gcc
patch_id
discussion_id
hgGNkR4qROW1AFjKLBH1JQ@gmx.net
source_type
public_inbox
title
Re: Is it time to take the middle-end stringop/array warnings out of -Wall?
headline
Is it time to take middle-end stringop/array warnings out of -Wall?
tldr
Richard suggests that middle-end diagnostics should enable diagnostic-paths to show currently visible preconditions in C++.
proposer
Richard Biener <richard.guenther@gmail.com>
consensus
Proposed
outcome
proposed
sentiment_score
technical_tradeoffs
  • Enabling diagnostic-paths may increase compilation time.
  • May require changes to GCC's internal data structures to efficiently track and display preconditions.
series_id
6C8E18B9-B946-4358-94A3-2979A435417C@gmail.com
series_role
reply
series_parts
[]
tags
  • gcc
  • warnings
  • middle-end
  • diagnostic-paths
  • C++
bugzilla_url
date
2026-05-22T00:00:00.000Z

Re: Is it time to take the middle-end stringop/array warnings out of -Wall?

Richard Biener comments on a previous proposal to categorize certain warnings as -Wmiddle-end. He suggests diagnostics should enable diagnostic-paths, to show currently visible preconditions. Biener notes that in C++ the set of visible preconditions depends heavily on inlining decisions, which affects how useful these warnings are.