dewclaw.git - the second-worst way to configure openwrt, by some metric

	Commit message (Collapse)	Author	Age	Files	Lines
*	add licenseHEAD main	pennae	2023-10-08	1	-0/+661
\|
*	add some documentation	pennae	2023-10-01	8	-22/+200
\|
*	add reload-only deployment option	pennae	2023-09-28	2	-15/+93
\| \| \| \| \|	this is mostly useful for things like static dhcp lease tables. reboots may be a bit harsh for those.
*	diagnose failed-but-unrecovered deployment attemps	pennae	2023-09-27	1	-0/+5
\|
*	improve on-device logging slightly	pennae	2023-09-27	2	-5/+12
\| \| \| \| \| \|	tell the user that a deployment has failed and log reboot messages from the device, just in case the network broke before the reboot (in which case logging the reboot event from the deploying host would be misleading)
*	don't provide log formats as environments	pennae	2023-09-27	2	-8/+8
\| \| \| \|	environments work, but this seems a little less fragile.
*	combine all scripts into a single environment	pennae	2023-09-26	1	-3/+10
\| \| \| \| \|	this makes it easier to generate a bunch of scripts from a single configuration without having to deal with multiple result links.
*	check correctness of UCI identifiers during build	pennae	2023-09-25	1	-2/+17
\| \| \| \| \| \| \|	uci import does not abort properly on config errors, so the deployment process continues and leaves the device with a partially invalid config. checking identifiers during build is also a lot faster than having the targets reboot due to import failures and subsequent rollbacks.
*	initial commit	pennae	2023-09-22	13	-0/+820
	without warranty of any kind, express or impliend