Magnus Eriksson <magetoo@fastmail.fm> writes:

>   Do I understand the partition handling right?  And if so, what caused
> the change in interpretation?

I don't sure but I found one bug under 1.6-RELEASE. It was my first
NetBSD installation. As usually, I made first partition under
NetBSD slice (/dev/Xc part of disk) for swap (wd0b and sd0b on
different machines). Reason was simple: first cylinders are the fastest,
so if swap will be activated it would be faster. It's work under
Linux, OpenBSD and FreeBSD and it was not work under NetBSD. These
harddrives were unbootble. 2 months later on my month holiday I had
discovered that if I make first partition Xa then NetBSD is bootble.
I noticed David Leight about this issue (he actively developed the
i386 sysinst and bootstrap parts at that time). One month later I
read in source-changes about new scheme. When I made release and
installed it on my spare comp I discovered that sysinst doesn't
allow to choose order of partitions any more. First is always Xa now.
I guess that disklabel code and other related programs were binded
to this new scheme. IIRC it was ~2 years ago.

FYI: I didn't parse source code so my conclusions are logical only. :)

> Magnus

HTH,
agkhram
-- 
= Alexey G. Khramkov (agkhram) @ Infotech <agkhram{at}mercdev{dot}com> =
= GPG fingerprint : 944D 0C8B 343B 6C8D 50A1  061A E2DA 3E11 7765 6B47 =