<philipp64>
Oh, right. Shit. Because I'm using `$(...)` notation. I figured since it was a function it would run inside the same shell as an optimization...
<philipp64>
What if I want to kill the parent shell as well?
<philipp64>
I could capture `$$` into a variable when it's sourced and then kill -9 $THATVAR
<philipp64>
ugly. must be a better way
<quinq>
You can just kill $$
<quinq>
Out of curiosity, what's the incentive behind replacing a ~100L awk script with a ~300L shell script?
<philipp64>
I'm writing the isc-dhcp init.d script to spit out Kea configs, and I need callable functions from shell that I can use for manipulating addresses, subnets, etc. as a library.
<philipp64>
I want to have a single shell library for manipulating IPv4 addresses in every possible way.
<quinq>
Maybe aim for portable shell then, better for libraries and future-proof
<philipp64>
This should be portable.
<dwfreed>
note that a standard openwrt install is only going to have busybox ash
<quinq>
It doesn't support all portable shell features?
<philipp64>
I've tested this with busybox.
<philipp64>
busybox shell does have a couple of features that aren't posix shell which we lean on, however... like ${var//pattern/replacement}
<philipp64>
but that genie is already out of the bottle.
<quinq>
and local at a quick glance
<philipp64>
yeah... well, I didn't break the seal on anything new...
<quinq>
:)
<quinq>
Time for bed, have a good night and good luck!
<russell-->
whatever i was seeing might have been the result of trying to work around the obnoxious target/linux/qoriq/base-files/etc/board.d/02_network
<dhewg>
Ansuel: for ipq4019, do you know what the content of board-2.bin is? Is it a another binary? Or just params for the firmware?
<Ansuel>
Just params
<dhewg>
any way to get those in a human readable format?
<dhewg>
but there's google wifi, also a 4019 device, which is known to work with >2 mesh nodes
<dhewg>
so I looked at that firmware, and noticed that they run an ancient fw
<dhewg>
but before blindly trying that I wanted to look at other differences, and obviously the google wifi device ships its own board-2.bin (which also doesn't match what openwrt ships for that device)
<dhewg>
that fw has '10.4-3.2.1-00051' in it's header
<dhewg>
so I wonder if I can just try their firmware-5.bin with the current board-2.bin? Or should I expect fire, explosions and such from such an attempt? ;)
<dhewg>
fucking ath, I wish there was a mt based vdsl board...
hitech95 has joined #openwrt-devel
<russell-->
i have a uci-defaults script that sets network.wan.device to eth0.201, after a sysupgrade -n, it's still sitting at eth0 (as it was before the sysupgrade). if i do the jffs2reset and reboot, it becomes eth0.201 as intended.
<russell-->
i guess i'm also unclear on why any configuration modification is running after /etc/uci-defaults scripts
nicklowe has joined #openwrt-devel
nicklowe has quit []
nicklowe has joined #openwrt-devel
<nicklowe>
@Ansuel Curious if there's an intentional reason the hardware RNG in the ipq806x is not wired up, I'm looking at the R7800, it's missing from cat /sys/devices/virtual/misc/hw_random/rng_available
<Ansuel>
it is wired up
<nicklowe>
@Ansuel Sure? cat of rng_current will give none
nicklowe has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]