[Toybox] me

[Rob Landley]

On 3/3/20 3:39 PM, lamiaworks wrote:
> i'm still around, just haven't complained for a while, which has kept Rob happy
> i'm sure.

And this wound up in moderation because you're subscribed from your lamiaworks
address, not from skymesh. (Which again, is not the fault of the mailing list
plumbing...)

> I'm awaiting at present patiently for toybox to have it's toyshell included and
> released so I can 'play' and complain "it doesn't to this and I need that
> function". Plus a few others in pending of course.

I've been blogging. https://landley.net/notes-2020.html#27-02-2020 was because
I'm trying to make the IFS separators in toybox work with utf8 chars, and it's
fiddly.

I'm currently debugging why 'for i in "$@"; do echo $i; done' is trying to exec
"in", and also why "for i; do echo $i; done" is saying bad ; right after the i.
(Which SHOULD use "$@" as the argument list, and the plumbing is there to make
it do so, but it was presumably never testes do needs debugged in the parsing
pass...)

I've also checked in a lot of tests that toysh doesn't pass yet, but I'm trying
to make it do so before release. (Although I may do the "screw it, cut a
release" thing in another week or so because it's waaaaay overdue".)

I'm now to the point where the test infrastructure is together enough I can add
real tests instead of appending notes-to-self to a file, but I haven't gone
through the file to convert the old notes-to-self into real tests yet.

> smiles wryly
> 
> scsijon

Rob

P.S. Although right NOW I'm off on a tangent trying to git bisect why
musl-cross-make stopped installing the linux headers for m68k and s390x targets,
which makes toybox defconfig not build because rfkill.c includes linux/rfkill.h
which isn't there, which means mkroot build on all targets has some targets
fail, and I taught "scripts/cross.sh all" to STOP when it gets a build failure...