[Toybox] [landley/toybox] The kernel's old kconfig had _shipped files, the new one should too, but until (a236381)

[Rob Landley]

On 06/21/2018 09:42 AM, Elliott Hughes wrote:
> isn't |ping| promoted now? or does it need more work before it can be used here?

You're right, toybox ping can be used here, and I should remove it from the list.

That commit was a quick and dirty fix because the mkroot build broke under the
new kernel without it. (And the _proper_ fix is the _shipped stuff, or at
_least_ teaching the kernel to call lex and yacc instead of flex and bison.)

A proper mkroot release is next up on my agenda after the toybox release. I
should really do them together, aboriginal linux building Linux From Scratch was
my main regression test for early toybox releases, and running real world data
through this stuff beats any canned test suite. The new one isn't quite to that
point yet because I've only got scraps of time/energy to work on this. New
$DAYJOB isn't as all-consuming as the previous one wound up being, but I'm still
pretty tired at the end of the day.

Rob

P.S. Today I had a meeting correlating my Jira issues with the same issues in
Rational and the local github instance (no idea why the github hook stopped
detecting the jira and rational issue numbers mentioned in the pull request and
updating those issues with a link, I should open a support ticket), another
meeting going over scheduling for the next "sprint", I missed a morning skype
session with the european developer who ported Mono to SH4 because I overslept
(back up to 3 energy drinks a day), taught another developer how to create core
files (ulimit -c/-Sc 999999999 and echo /path/to/blah%p >
/proc/sys/kernel/core_pattern, with links to relevant docs), helped diagnose two
different causes of latency spikes in other people's code, redid the build
system of a userspace battery monitor package that didn't match up with the
battery driver since I forward ported the product from a 3.18 kernel with a 4.14
kernel right before my week in Austin (fixing it involved breaking a circular
dependency and changing where the driver was installing its header file in the
SDK toolchain, the resulting battery app refuses to run but at least it's
building and installing now, the "exits immediately with an error" is somebody
else's problem), wound up with a networking issue back on my plate (because I
looked up iptables rules to ratelimit certain broadcast packets and handed them
off to the guy who wrote the firewall management stuff since it's ratelimiting
to a port that moves when you change configuration in the gui, _but_ those rules
depend on an iptables config option being switched on which means I need to
rebuild the kernel), and I never -did_ figure out why the guy in the next "open
floor plan" cubicle (who's visiting from India for a week and speaks almost
exclusively an indian dialect with other developers from India) kept saying "I
am Groot" (often in response to what people said to him, which was the only part
of their conversation in english) but luckily I have headphones.

And then in the evenings, I get to work on toybox.