Re: To gnupg or not to gnupg ... (was Re: Assertion while pui-list-packages)
Andy Piper <andyp(a)parallax.co.uk> writes:
>I know what John Martin thinks about this. Any other comments?
At
>this time I'm strongly in favor of moving forward to gnupg.
Sounds ok to me. If people really care about checking the signature (and I
don't) then they will make sure they have the right version installed.
Is it really this simple? I'm looking at the GnuPG homepage[1] and it
appears GnuPG currently supports a limited number of platforms. So
those of us who run one of the unsupported platforms won't be able to
verify the GnuPG signatures even if care about doing so. In light of
this I think the move to GnuPG should be delayed until the software
matures and is able to support more platforms.
Footnotes:
1. <URL:http://www.d.shuttle.de/isil/gnupg/>
Supported Systems
GnuPG works fine on GNU/Linux with x86, alpha, sparc64, m68k or
powerpc CPUs. (x86 is my primary development system, the other CPUs
are only checked from time to time)
It compiles okay on GNU/Hurd but because Mach has no random device,
it should not be used for real work. It should be easy to add the
random device driver from Linux to the Hurd - Anyone?
FreeBSD with x86 CPU works fine. OpenBSD works fine (x86 CPU?).
GnuPG compiles and runs on many more systems, but due to the lack of
a good random number source, it should not be used on these
systems. I have reports on these systems: HPUX v9.x and v10.x with
HPPA CPU, IRIX v6.3 with MIPS R10000 CPU, SunOS, Solaris on sparc
and x86, Windoze 95 and WNT with x86 CPUs.