Re: [dev] A replacement for at.

2016-01-01 Thread Mattias Andrée
On Fri, 1 Jan 2016 10:19:01 +0100 Mattias Andrée wrote: > Hi! > > I'm written an alternative to at, called sat (for simple > at): https://github.com/maandree/sat sat is incompatible > with at, but I have tried to make sure that a > compatibility-layer can be written. > > sat is basically at wit

Re: [dev] A replacement for at.

2016-01-01 Thread Kamil Cholewiński
On Fri, 01 Jan 2016, Greg Reagle wrote: > On Fri, Jan 01, 2016 at 06:05:58PM +0100, Kamil Cholewiński wrote: >> Yes, please let's stop writing process management code into daemons and >> instead solve this problem in a portable and non-sucky way. > > It's called runit: http://smarden.org/runit/

Re: [dev] A replacement for at.

2016-01-01 Thread Mattias Andrée
On Fri, 01 Jan 2016 19:05:50 +0100 Kamil Cholewiński wrote: > > sat will create a PID file in $XDG_RUNTIME_DIR. > > PID files are a flawed concept, race conditions and > everything. Well. But they are standard. satd does not use the PID file to determine whether it is running, it uses flock.

Re: [dev] A replacement for at.

2016-01-01 Thread Greg Reagle
On Fri, Jan 01, 2016 at 06:05:58PM +0100, Kamil Cholewiński wrote: > Yes, please let's stop writing process management code into daemons and > instead solve this problem in a portable and non-sucky way. It's called runit: http://smarden.org/runit/

Re: [dev] A replacement for at.

2016-01-01 Thread Mattias Andrée
On Fri, 01 Jan 2016 18:05:58 +0100 Kamil Cholewiński wrote: > > Wouldn't you need a service supervisor with at's > > functionallity? > > No. Separation of concerns. > > > If you are paranoid about sat crashing > > When in doubt, assume the component will crash/fail. > > > as long as you c

Re: [dev] A replacement for at.

2016-01-01 Thread Kamil Cholewiński
> Wouldn't you need a service supervisor with at's functionallity? No. Separation of concerns. > If you are paranoid about sat crashing When in doubt, assume the component will crash/fail. > as long as you can have user-private services. Yes, please let's stop writing process management code i

Re: [dev] A replacement for at.

2016-01-01 Thread Mattias Andrée
If you are paranoid about sat crashing and not get your jobs executed, it is possible to start sat under service supervision, as long as you can have user-private services. On Fri, 01 Jan 2016 15:17:43 +0100 Kamil Cholewiński wrote: > > satd is an unprivileged daemon that is user-private, and >

Re: [dev] A replacement for at.

2016-01-01 Thread Mattias Andrée
Wouldn't you need a service supervisor with at's functionallity? Are there any? If you want to do things really simply you can use sleep-until (https://github.com/maandree/sleep-until), but then you cannot as easily list jobs and run the before scheduled. On Fri, 01 Jan 2016 15:17:43 +0100 Kamil

Re: [dev] A replacement for at.

2016-01-01 Thread Kamil Cholewiński
> satd is an unprivileged daemon that is user-private, and > starts and exits automatically. Why not use a service supervisor?

[dev] A replacement for at.

2016-01-01 Thread Mattias Andrée
Hi! I'm written an alternative to at, called sat (for simple at): https://github.com/maandree/sat sat is incompatible with at, but I have tried to make sure that a compatibility-layer can be written. sat is basically at without a lot of features that does not need to be there. sat is also written