Re: Has the update procedure changed?

From: Mark Millard <marklmi_at_yahoo.com>
Date: Tue, 08 Aug 2023 03:53:37 UTC
On Aug 7, 2023, at 19:50, Robert Huff <roberthuff@rcn.com> wrote:

> Mark Millard <marklmi@yahoo.com> writes:
> 
>>  The material in 26.6.5 does not repeat all that, it is
>>  more of a summary that is presented.
>> 
>>  There are also instructions in UPDATING (near the end)
>>  and yet other instructions in Makefile (leading
>>  comments).
>> 
>>  I've not done detailed comparisons of such in some
>>  time, but they well not be exact matches. As I remember,
>>  the UPDATING material was more explicit about various
>>  cases/contexts and what was appropriate for them.
>> 
>>  It can be appropriate to review them all.
> 
> 1) It would be really nice is someone would take a look and make
> sure these 100% non-contradictory.

To be effective, this may require the additional criteria of
being very explicit each time there is a partial listing
of commands that summarize only some of the overall steps, the
ones involved for the subject(s) in the local subsection but
not other steps. The handbook 26.6.5 command lists would be
an example.

> 2) If I can only use one of these, which one should I choose?

The only one that covers a variety of types of contexts
is the one in UPDATING, if I remember right. It is hard
to avoid referencing that, even if only to figure out
which type of context one happens to be interested in
for the specific update at hand.

I'm not aware of installkernel depending on etcupdate
activity, but "man etcupdate"  reports, in part:

     -p             Enable “pre-world” mode.  Only merge changes to files that
                    are necessary to successfully run ‘make installworld’ or
                    ‘make installkernel’. . . .

To my knowledge not even UPDATING shows etcupdate -p
being used before installkernel. "man etcupdate" may be
trying to cover potential future updates that could
change the status? Still, it looks contradictory.

> I'm procrastinating on doing a najor update, and want to keep
> the long record of nothing going Horribly Wrong(tm).




===
Mark Millard
marklmi at yahoo.com