Re: CVS commit: src/distrib/amd64

[David Brownlee <abs%netbsd.org@localhost>]

On Tue, 17 Jan 2023 at 08:57, Robert Elz <kre%munnari.oz.au@localhost> wrote:
>
>     Date:        Tue, 17 Jan 2023 08:13:12 +0100
>     From:        Martin Husemann <martin%duskware.de@localhost>
>     Message-ID:  <20230117071312.GA18963%mail.duskware.de@localhost>
>
>   | what is missing in this
>   | environment so it does not work by default?
>
> The shell started this way is not a login shell, so does
> not read any of the standard login scripts, and I assume
> also does not have ENV set to a suitable startup script
> either.

Adding an explicit -E seemed like the simplest and smallest (not
adding an additional file to the image) way to trigger this behaviour
(and also clearest to anyone reading through). (Happy to be adjusted
by someone with better sh fu :)

So essentially I think there are two questions (referencing martin's reply)

1) Should this be amd64 only:

No. I wanted to start with amd64 to establish a baseline, and then
work towards ports with very limited install space & memory

2) Is this the best mechanism:

There is an argument that any interactive shell should default to -E
in the absence of some other trigger, whether login, single user,
install sh (not -DSMALL), or just typing sh (or some subset, such as
with an explicit '-i' or interactive with '-l')

Some other systems (and shells) do this, but I feel there would need
to be a clear and coherent way to not trigger the behaviour.

David