Page MenuHomeFreeBSD

Convert tools/regression/sockets/socketpair to ATF
ClosedPublic

Authored by asomers on Feb 10 2018, 6:21 PM.
Tags
None
Referenced Files
Unknown Object (File)
Fri, May 17, 2:22 PM
Unknown Object (File)
Mar 11 2024, 3:47 PM
Unknown Object (File)
Mar 11 2024, 3:47 PM
Unknown Object (File)
Mar 11 2024, 3:43 PM
Unknown Object (File)
Mar 8 2024, 2:04 AM
Unknown Object (File)
Mar 7 2024, 10:57 AM
Unknown Object (File)
Feb 26 2024, 9:10 AM
Unknown Object (File)
Jan 13 2024, 3:40 AM
Subscribers

Details

Summary

Convert tools/regression/sockets/socketpair to ATF

Test Plan

ATF tests added

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
No Lint Coverage
Unit
No Test Coverage
Build Status
Buildable 14940
Build 15052: arc lint + arc unit

Event Timeline

contrib/netbsd-tests/lib/libc/sys/t_socketpair.c
72

Given how these are used, can type | flags be reduced into a single flags parameter (for run())?

104

Is this required by a standard, or just codifying current behavior?

131

This can't be done in a hypothetical capsicum mode; maybe dup stdin like the cloned test?

contrib/netbsd-tests/lib/libc/sys/t_socketpair.c
72

Technically yes, but I think it's more logical to do it this way.

104

Current behavior. The standard says "The socketpair() function is used primarily with UNIX domain sockets and need not be supported for other domains."

131

Do we need to run in capsicum mode? I choose to do it this way for consistency with the existing NetBSD-derived code.

cem added inline comments.
contrib/netbsd-tests/lib/libc/sys/t_socketpair.c
72

I find the other way more logical, but am fine with it as-is.

104

I'm not sure I like a test enforcing the lack of a feature, if the standard doesn't explicitly say that feature is disallowed. But I am no testing expert.

131

*shrug*. It doesn't matter for now.

This revision is now accepted and ready to land.Feb 10 2018, 7:31 PM
This revision was automatically updated to reflect the committed changes.