On dv., ag. 09 2019, Vincent Bernat wrote:
❦ 9 août 2019 09:22 +02, Martin Steigerwald <martin@lichtvoll.de>:Reality seems different. Almost nothing was using inetd (tftpd is theI note that you wrote "seems". But still:As if there would just be *one* reality. Actually there is. But I never saw any human being being able to express it in words. So I'd rather not. I believe it can be experienced at any time. But for me it isbeyond words and so much else.With arguing about what reality might be and claiming it is this or such… the trouble starts. Cause then people who somehow dare to manage to experience a different reality can easily be made wrong. I think this has been one of the core issues around the conflicts regarding Systemd. How dare you see things different than me? But you just need to talk to ten people to recall a situation they experienced together and you willreceive ten different story. Now: which one would be right?The one which provides real-world examples. One says "socket activation is not used for decades, let's just not use it", I say "it is used right now, see the following examples". You come and you say "I don't use it with dovecot". Sorry, but upstream did implement socket activation for areason, not out of the blue of nothing.
Not that it's too relevant, but most of this sub-thread already isn't: Socket activation was used in low end embedded devices running Debian, precisely as Simon described, already 10-15 years ago. Those devices just didn't have the RAM (32M!) to have the processes running all the time, but they could afford swapping and starting the services on-demand for shorts period of time. That one hasn't first-hand experienced (or noticed) certain things doesn't mean they were/are not there.
In any case, and this bit *may* be relevant, but that's for each person to judge and my understanding / perception may be wrong: it'd look as if lately on this ML many technical topics derive in some bits of the Debian community pushing for "let's drop sysvinit" or (wrongly) claiming that "sysvinit is bit rotting and nobody is using it" and that in turn results in long discussions like this.
My theory on this is: those that were very vocal against systemd in a non-constructive way moved away from Debian, those who were vocal *for alternatives* in a constructive way are trying to do what they can where they can (which is not always directly Debian); I perceive systemd-bashing to be mostly not a thing *here* (and that's good!); but it looks like this kind of threads, discussing details of wildly different use-cases for Debian or how "systemd can do X and Y can't > but you also can do it without it in a perceived simpler way" are more of a magnet for quick postings than those actually tackling issues, which usually have very good points and different perspectives.
Basically: the Debian community has been able to make the change a big part of it wanted to make (adopting systemd by default), but that means the other bits of that process should also be respected. That is: sysvinit is supposed to be supported and that's not going to be always in the same way; sometimes that will require not blindly adopting systemd's upstream's way of doing things for everything, but talking things through and seeing what would be best for Debian as a whole, sometimes supporting other init's will be a bit of an after-thought, sometimes just having different init's be possible in the same OS it will help find hard-to-spot bugs. If that's accepted and respected, it's not a thing to argue about or waste brain-cycles in, in the meantime things are getting both calmer and better.
-- Evilham