Interesting - I wonder why GNUTLS is a depedency (I'm building it from source on IRIX right now, no GNUTLS (currently built) on there). On the GPG website libksba is listed as optional, and npth is listed as 'you don't need it but probably want it'.

I'm all for replacing a lot of GNU software because of issues like that. Tying to build it manually ends up being a nightmare, because a lot of it depends on other GNU software they you may not want on there for various reasons.

Except if you put it in systemd, then it becomes confined to Linux. Side note, GPG gets used across quite a few platforms (I see OS X, Windows, and VMS listed on the binaries page, and seems to be good on other Unix systems too), so it makes for a great utility for others to be able to use to verify whatever.

Hey - you can even use Google's TF for your home SSH box.

I don't think the OP was talking about his laptop.

What makes you think systemd addresses issues? Or, I should say, that the BSD people seem them as issues?

You know, I've generally found Linus to be quite the opposite, and he will even submit patches to other projects (such as GNOME) when he finds issues, rather then creating trouble.

On a similar note, I've also found Theo de Raadt to not be an issue to work with. Neither take Mr. Poettering's approach of 'the users don't like the software, users must be dense' when people have issues with their projects.

I think the bigger complaint is that it's being added to systemd, not that it exists (Note that GRUB can already be used with secure boot). Lennart Poettering is pretty disliked for his abandonment of UNIX principles (the biggest one being portability), and somehow his software becomes the de facto standard in the Linux world, long before it is ready (PulseAudio anyone)? He creates issues and fractures the community, and then blames everyone else for the problems.

Ubuntu, like Red Hat, tries to keep the version of components the same for the life of a (LTS) release as not to risk breaking compatibility and application certification. Fedora's kernel updates can occasionally break things, and not breaking compatibility is very important for a long term release. (Compare Ubuntu LTS to RHEL or SuSE Enterprise)

There is a good chance that I can get Solaris to build (especially Solaris 11). I'm not sure there is much interest in me porting to IRIX.

I'd consider using it if there was a Solaris build. One of the things about Open Office when Sun had it was the cross platform support (I have OO on my Solaris workstation (SunBlade 1500) AND on my IRIX box (SGI Octane)

Actually Fedora and Red Hat are the only one that matters. If you haven't noticed, what Red Hat does, everyone else does too.

I was referencing libressl in particular on this one. MIPS Pro, when set to c99, seems to handle most code that isn't bound to GCC with some exceptions. I'm building everything 32-bit here, so I can't speak for the 64 bit quirks. I'm also using GCC 4.7 for libressl, among other things that won't build right (php-5.6 and httpd-2.4 being the big ones on that particular box, I've built a lot of things on IRIX to bring stuff up to date though).

Windows yes. I haven't seen anyone submit patches for OpenVMS, and I don't have an OpenVMS system to even try on. It does, however, support OS X, Net and Free BSD, HP-UX, and Solaris. I used to have it working for AIX and IRIX, but changes in the 2.1.x release have broken them for now, and I'm not sure that I'll ever get IRIX working again.

I forgot Solaris in the supported list! Tested on Solaris 10-11.

Actually, libressl supports OS X and HP-UX as well. Some groundwork is in place for supporting AIX and IRIX (I no longer have access to AIX to continue porting, and I'm not sure IRIX will ever work right). If you really wanted it to work with MSVC, you could write, test, and propose the patches to make it work. I'm all for eliminating GCCisms (the areas I've been poking at the code I'm not trying to eliminate GCCisms, not my priority).