bugzilla-daemon at mindrot.org
2024-Mar-31 00:29 UTC
[Bug 3676] New: Redirect conch stdin from /dev/zero rather than requiring a controlling terminal
https://bugzilla.mindrot.org/show_bug.cgi?id=3676 Bug ID: 3676 Summary: Redirect conch stdin from /dev/zero rather than requiring a controlling terminal Product: Portable OpenSSH Version: -current Hardware: Other OS: Linux Status: NEW Severity: enhancement Priority: P5 Component: Regression tests Assignee: unassigned-bugs at mindrot.org Reporter: cjwatson at debian.org Created attachment 3799 --> https://bugzilla.mindrot.org/attachment.cgi?id=3799&action=edit regress: Redirect conch stdin from /dev/zero The new controlling-terminal requirement in regress/conch-ciphers.sh is somewhat inconvenient. I know that in theory wrapping the regression tests in something like script(1) should be fine, but in practice I haven't been able to get this to work properly in Debian's autopkgtest framework where we run the regression tests automatically; when I tried the result was that all the output of the regression tests was hidden from stdout and only showed up in autopkgtest's separate log file, but only when running in CI jobs on salsa.debian.org and not locally, which was extremely weird. I could probably figure this out if I spent long enough debugging it, but a tiny patch to avoid the requirement in the first place seems like a more sensible use of time. Would you consider the attached patch? -- You are receiving this mail because: You are watching the assignee of the bug.
Apparently Analagous Threads
- Baffling regress/forwarding.sh failure, new in 7.4p1
- Call for testing: OpenSSH 8.4
- OpenSSH 7.7p1 interop failure with Twisted Conch 8.0.0-17.9.0
- OpenSSH 7.7p1 interop failure with Twisted Conch 8.0.0-17.9.0
- [Bug 2494] New: kex_protocol_error should send SSH2_MSG_UNIMPLEMENTED