[Toybox] status.html claims sh is complete

[Rob Landley]

On 4/21/23 17:15, enh via Toybox wrote:
> subject says it all --- https://landley.net/toybox/status.html
> <https://landley.net/toybox/status.html> claims sh is complete. caused confusion
> earlier today when i was telling someone it's a work in progress and they
> pointed to the docs to prove me wrong :-)

Yeah, somebody else poked me about that a few weeks ago, I thought it was on
github but it was probably private email.

The binaries I ship are generated by mkroot, and building mkroot switches on two
commands out of pending because it needs them to function: route and sh. I need
to rewrite route based on the netlink API so it can manage multiple routing
tables, and sh is just plain big.

(Mkroot needing to switch on two commands is actually one of the sharp edges in
the design, because if there's an existing toybox .config it will use it rather
than creating its own, so if you run defconfig then run mkroot without doing a
distclean first, the result won't boot because it hasn't got a shell. The
_proper_ fix is to finish and promote route and sh.)

Last release I ran scripts/mkstatus.py against the toybox binary I shipped, and
that runs ./toybox out of the build directory to see what commands are enabled.
So if it's a mkroot binary, it sees route and sh are there.

So that's _how_ it happened, and the status of those two commands is a weird
schrodinger thing, but now that it's confused more than one person I'll build a
vanilla 0.8.9 checkout and regenerate the page, and try to err on the OTHER side
next release.

Rob