Should MOC Continue To Support OpenBSD?

I've been working recently on bringing MOC up to the currently common C99 and POSIX.1-2001 standards. However, the road block I keep running into is with OpenBSD... do we continue to support it?

Targetting these standards will enable the removal of some fallback code for some POSIX-mandated functions, and the use of more modern code contructs. But perhaps more importantly, it should remove the need to support specific OSs and allow MOC to be used on any standards-compliant UNIX-like system (although there remain some GNU-specific constructs which will be addressed later).

According to the OpenBSD sys/unistd.h header, the version of the POSIX.1 standard it targets for compliance is the 1990 one. Whatever their reasons, that standard's nearly a quarter of a century old now, so maybe in the interests of moving MOC forward we have to leave OpenBSD behind on MOC 2.5.

For those who are using MOC on OpenBSD (and I honestly don't know if there is anyone), I see a number of alternatives:

  • Stop using MOC,
  • Remain behind on MOC 2.5,
  • Move away from OpenBSD (probably to FreeBSD or NetBSD, both of which target the 2001 standard),
  • Convince the OpenBSD developers to move to the more recent POSIX standard,
  • Convince OpenBSD's MOC package maintainer (if there still is one) to provide and maintain within the OpenBSD distribution the patches necessary to downgrade MOC to the 1990 standard,
  • Patch MOC yourself to downgrade it to that standard, or
  • Have someone fork MOC for OpenBSD.

None of those options is good news for any OpenBSD user still roaming free in MOCland, so I'd be interested in your feedback. If I hear nothing within a week, I'll consider it acceptable to require POSIX.1-2001 compliance and continue moving forward.

John Fitzgerald,
MOC Maintainer.


"and I honestly don't know if there is anyone"... Me neither. I'll be happy as long as MOC works on Debian stable. ;)

After 'tedu'[0]: what POSIX features do you want/need/planning on using?




Out of the big three BSDs its probable the one I have the most interest in, sad to see that it is losing support by
targetting a older posix standard.

Its your time and your project, I'd say that OpenBSD should provide support for their differences within their
own project members through their MOC port maintener, also any chance of finding out if they do have any plans on upgrading their POSIX standard?
they're pretty conservative on upgrading code and since their core code gets reviewed and audited they might have just
been slower to get on the bandwagin with more pressing things, if it goes that way they can just use their current working copy until their
POSIX 2001 comes in, which would just make you free to do what you want.

Looks like this post has already attracted attention by OpenBSD users.

Conservative? hah! I'm a noob, they are regularly redesigning their code to remove
exploits all the time.

After a similar appeal on one of the OpenBSD mail lists, it seems that the fourth option is the one they went for. That surprised me, but they seem happy that OpenBSD is now POSIX.1-2008 compliant and so MOC can now move forwards hand-in-hand with its OpenBSD users.

I shall be working with one of the OpenBSD maintainers over the next few days to ensure that some of the problems I was experiencing in testing are now resolved and then committing the patchset which will bring MOC up to C99 and POSIX.1-2001 compliance and also make them a minimum requirement.

After investigating a potential problem in one of the previous commits, MOC 2.6-alpha2 will be next on the agenda.

it is working fine on openbsd. We'll see if the upcoming changes work with obsd.
I'll make sure they do.
the port is in openbsd-wip (work in progress)
I can't post the url because of the spamfilter, it's in 'jasperla' at github.

Thanks for that.

I have a set of patches I was working through with one of the OpenBSD folk, but it has stalled due to lack of time on my part.