[OpenWrt-Devel] [PATCH] netifd: fix 14_migrate-dhcp-release script

Hans Dedecker dedeckeh at gmail.com
Sat Apr 4 15:29:43 EDT 2020

On Sat, Apr 4, 2020 at 9:04 PM <mail at adrianschmutzler.de> wrote:

> >>> prepend 'uci' to 'commit network'
> >> Can you explain why the prepend of 'uci' is required ?
> >> In other words what is not working as I fail to spot the issue
> >
> > I think the author is right. "commit" is no command but an argument to
> the uci command.
> > I've tested the script and it works fine for me ..
> > So I want to understand what is failing ...
> uci-defaults "scripts" are implemented here:
> https://github.com/openwrt/openwrt/blob/master/package/base-files/files/lib/functions.sh#L256-L263
> As you see, after all of them have been sourced a global "uci commit" is
> called anyway.
> So, after all, the "uci commit something" statements in uci-defaults files
> are actually not strictly necessary. However, many authors prefer to have
> committed what they touch directly in the uci-defaults script.
> This is handled relatively inconsistently across OpenWrt main/packages
> repos, I once had a look into it and found that it was about 50:50 back
> then (having a commit in the file vs. not having it).
I observed the same inconsistency across the code base which makes it hard
to refer to a default way of working

> However, I'd still expect the "commit network" to produce some error
> anywhere.
I did not see an issue when doing a quick test as the global 'uci commit'
covered up the issue for me

> But instead of adding the "uci" in front, another working solution would
> be to just drop the line entirely.
Agree and I would prefer this solution


> Best
> Adrian
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.infradead.org/pipermail/openwrt-devel/attachments/20200404/99bd059d/attachment.htm>
-------------- next part --------------
openwrt-devel mailing list
openwrt-devel at lists.openwrt.org

More information about the openwrt-devel mailing list