On Sun, May 23, 2021 at 09:01:01PM +0200, Cyberfusion wrote: > Op 23 mei 2021 om 20:17 heeft Ercolino de Spiacico het volgende geschreven: > > > > I didn't try but I suppose the --test parameter works. However this > > is not my point. > > I would like dnsmasq to stop (prevent execution) only for very serious > > matters like incomplete IP addresses in config, binding port higher > > than 65K, etc. you name it. > > > > I believe there is a rather long list of reasons (incorrect dhcp-host > > syntax being one as per original message) that could/should be > > excluded from this process execution control. > > > > It is impossible for software to decide what is a ‘serious matter’ and what > isn’t. > > > This takes nothing away from the fact that having a clear/clean config > > is ultimately a must, I just don't want to have to call a person > > in a different continent again because of a typo on a "secondary" > > configuration parameter. > > > > You don’t have to. You can validate the config. > > > In one sentence: let's make dnsmasq as resilient as possible. > > > > The workflow should be as resilient as possible. Ergo, validate your > config. You could add the validation command to ExecStartPre to prevent > dnsmasq from restarting with a faulty config.
I do like that idea. While preparing a patch for it, I learnt it is already implemented: * https://thekelleys.org.uk/gitweb/?p=dnsmasq.git;a=blob;f=debian/systemd.service;h=b769c4ea04f973b2069f529923c1b1821a6ff59c;hb=HEAD#l13 * https://thekelleys.org.uk/gitweb/?p=dnsmasq.git;a=blob;f=debian/systemd@.service;h=16ca86e9826888c671e05fc05b2731255bc05445;hb=HEAD#l13 * https://thekelleys.org.uk/gitweb/?p=dnsmasq.git;a=blob;f=contrib/systemd/dnsmasq.service;h=ff0ed089b2062258fbeccc48d5b7becc05670557;hb=HEAD#l10 Groeten Geert Stappers -- Silence is hard to parse _______________________________________________ Dnsmasq-discuss mailing list Dnsmasq-discuss@lists.thekelleys.org.uk https://lists.thekelleys.org.uk/cgi-bin/mailman/listinfo/dnsmasq-discuss