> On Sep 14, 2026, at 10:10 PM, Benjamin Kaduk <[email protected]> wrote: > > On Sun, Sep 13, 2026 at 5:32 AM Dag-Erling Smørgrav <[email protected] > <mailto:[email protected]>> wrote: >> The branch main has been updated by des: >> >> URL: >> https://cgit.FreeBSD.org/src/commit/?id=4cc85d1f00fa1938923fe8d6fc2934593d4303b4 >> >> <https://cgit.freebsd.org/src/commit/?id=4cc85d1f00fa1938923fe8d6fc2934593d4303b4> >> >> commit 4cc85d1f00fa1938923fe8d6fc2934593d4303b4 >> Author: Dag-Erling Smørgrav <[email protected]> >> AuthorDate: 2026-09-13 12:27:50 +0000 >> Commit: Dag-Erling Smørgrav <[email protected]> >> CommitDate: 2026-09-13 12:27:50 +0000 >> >> tests: Fix graid3 insert tests >> >> The script was passing the wrong device name to `graid3 insert` and >> didn't notice that the command was failing. > > > It looks like nothing changed w.r.t. being able to notice the command is > failing ... is this something where `set -e` is appropriate?
Using set -e in tests seems like it’s more likely to cause problems than solve them just based on past experience with integrating/maintaining tests written in other (non-ATF) paradigms. I agree that that would be the best default behavior for production code though. Cheers, -Enji
signature.asc
Description: Message signed with OpenPGP
