Test failures on OpenBSD

[Topaz]

I tried installing bup (the 0.33.x branch, commit 3c33f2c2) on our OpenBSD VM, so we could back it up. It required a couple things not mentioned in the readme (like bash, and apparently rsync for the tests, which just fail if you don't have it installed) but even after installing those, a few of the tests are still failing.

Some of those look to be failing because the OpenBSD versions of commands don't necessarily take GNU-specific options, like "script -q" and "df -T". Others don't have a clear failure reason.

I've attached the full test log, in case anyone feels like having a poke at it.

(I know bup isn't supported on OpenBSD, I'm just thinking this might be useful if anyone wanted to add support.)

-- Topaz

[Greg Troxel]

"'Topaz' via bup-list" <bup-...@googlegroups.com> writes:

> I tried installing bup (the 0.33.x branch, commit 3c33f2c2) on our
> OpenBSD VM, so we could back it up. It required a couple things not
> mentioned in the readme (like bash, and apparently rsync for the
> tests, which just fail if you don't have it installed) but even after
> installing those, a few of the tests are still failing.

I've been using bup on NetBSD, and this has resulted in a stream of
portability fixes over the years, which have been handled in a a great
manner. I'm the "bup chief POSIX-says-foo" ranter :-)

> Some of those look to be failing because the OpenBSD versions of
> commands don't necessarily take GNU-specific options, like "script -q"
> and "df -T". Others don't have a clear failure reason.

NetBSD's df doens't have -T either. Not sure why that doesn't cause me
trouble, and in 5 minutes of code reading I couldn't figure out where df
is called.

NetBSD's script does have -q. But the test probaby should be adjusted
to not use it, and then have to sed out the non-wanted lines which
probably have non-constant content.

FWIW, NetBSD's man page says that script originated in 3.0BSD.

> (I know bup isn't supported on OpenBSD, I'm just thinking this might
> be useful if anyone wanted to add support.)

supported is a funny word. My take has always been that programs should
work on any reasonable mostly-POSIX system, and OpenBSD is certainly in
that category. Sounds like pretty minor changes, and I'm 99%+ sure if
you send a reasonable patch things would be adjusted.

I think in practice supported means that it's tested before release.
That's GNU/Linux by Rob and others, NetBSD by me, and macOS usually by
me and probably others. Perhaps more, probably FreeBSD.

[Rob Browning]

Greg Troxel <g...@lexort.com> writes:

> supported is a funny word. My take has always been that programs should
> work on any reasonable mostly-POSIX system, and OpenBSD is certainly in
> that category. Sounds like pretty minor changes, and I'm 99%+ sure if
> you send a reasonable patch things would be adjusted.

Indeed, be happy to consider relevant changes -- might even take a look
myself if I get time.

And regarding script, that's new, to let us test the "interactive"
output mess in 0.33.x, and what it's trying to test isn't very likely
platform dependent, so it should be fine to just skip those on some
platforms.

> I think in practice supported means that it's tested before release.
> That's GNU/Linux by Rob and others, NetBSD by me, and macOS usually by
> me and probably others. Perhaps more, probably FreeBSD.

And CI currently covers some test variants (via Debian), FreeBSD, and
macOS.

https://cirrus-ci.com/github/bup/bup

Thanks
--
Rob Browning
rlb @defaultvalue.org and @debian.org
GPG as of 2011-07-10 E6A9 DA3C C9FD 1FF8 C676 D2C4 C0F0 39E9 ED1B 597A
GPG as of 2002-11-03 14DD 432F AE39 534D B592 F9A0 25C8 D377 8C7E 73A4