Updated behavior with oflag=check: 1.Probe only mounted, LVM PV, and MBR/GPT. No file-system check, no libblkid. 2.No risk detected, or probe failure: continue as before; probing never makes dd fail by itself. 3.Risk detected: prompt via /dev/tty. If declined, exit non-zero. 4.If /dev/tty is unavailable: print warning to stderr, treat as refusal, exit non-zero (fail closed). This revises my earlier warn-and-continue proposal. Is oflag=check the right name? Is fail-closed acceptable? Any wording/NEWS/manual preferences? If you have any suggestions, please feel free to discuss them with me.
Thanks， Jianing Weng On 2026-09-09 16:39, ii via GNU coreutils Bug Reports writes: >The current implementation addresses both concerns as follows: >1.Opt‑in only: oflag=check is a new, optional flag. Without it, dd behaves >exactly as before. The default behavior remains unchanged. >2.Non‑interactive safety: When /dev/tty is unavailable (e.g., in cron or >background jobs), dd prints the warning on stderr and proceeds automatically, >it never hangs or fails. The function ask_to_proceed() returns true if it >cannot open the controlling terminal. > >On 2026-09-09 13:50,"Paul Eggert"<[email protected]> writes： >Something like that might be OK as a new option, but it shouldn't be the >default; too many scripts routinely use dd to overwrite devices containing >file system>s, and there may not be anybody around to prompt (or the user may >not expect a prompt or know what to do with it). > >On 2026-09-08 19:26, ii via GNU coreutils Bug Reports wrote: > The pre-write check is safe and does not affect the write even if probing > fails. > It merely warns and prompts for confirmation > ii [email protected]
