Re: Bug#922353: ITP: socket-activate -- Run a socket-activated daemon with minimal dependencies

[Peter Pentchev]

On Sat, Feb 16, 2019 at 03:32:46PM +0100, Guillem Jover wrote:
> Hi!
> 
> On Fri, 2019-02-15 at 10:46:43 -0500, Daniel Kahn Gillmor wrote:
> > Control: clone 922353 -2
> > Control: reassign -2 dpkg
> > Control: retitle -2 start-stop-daemon should support socket-activation via 
> > the sd_listen_fds(3) convention
> > Control: severity -2 wishlist
> 
> Thanks.
> 
> > On Fri 2019-02-15 04:34:47 +0100, Guillem Jover wrote:
> > > Another option would be to implement this in start-stop-daemon, like
> > > the similar support for the systemd readiness protocol was recently
> > > implemented there too.
> > 
> > Thanks for the suggestion!  How widely-distributed is start-stop-daemon
> > outside of debian?  I see it's been ported to OpenBSD; are they
> > syncing from upstream?
> 
> . For the BSDs to use
> this more seriously the code would probably need to be split into its
> own project, so that it does not pollute their licensing. Its current
> "license" might also need to be clarified (PD), and "relicensed" into
> MIT or similar.
> 
> This is something I've actually pondered doing anyway, so this might
> be a good excuse, I guess.
> 
> > The code i have is just python3 right now (simple argument parsing made
> > development much quicker), but it's not too terrible to do it in C.
> 
> Yes, and it should be pretty generic and portable.
> 
> > I'm opening this as a wishlist issue for dpkg just so we don't lose
> > track of it, since it might take me some cycles to get the C
> > implementation in shape.  If anyone else wants to beat me to it, i
> > certainly wouldn't complain :)
> 
> I'll probably look into it once I've gone over some of the immediate
> stuff I have on my plate, if there's been no patch submitted by then.
> :)

Here you go :)

https://gitlab.com/dkg/socket-activate/merge_requests/1#note_142084524

G'luck,
Peter

-- 
Peter Pentchev  roam@{ringlet.net,debian.org,FreeBSD.org} p...@storpool.com
PGP key:http://people.FreeBSD.org/~roam/roam.key.asc
Key fingerprint 2EE7 A7A5 17FC 124C F115  C354 651E EFB0 2527 DF13


signature.asc
Description: PGP signature

Re: Bug#922353: ITP: socket-activate -- Run a socket-activated daemon with minimal dependencies

[Guillem Jover]

Hi!

On Fri, 2019-02-15 at 17:24:24 +0100, Andrej Shadura wrote:
> Speaking of which, maybe it would also make sense to merge my changes
> in?
> 
> https://bitbucket.org/andrew_shadura/start-stop-daemon-isolate/commits/8646be59e3fde0b22f84a842ef5729a5de08fd3b

This being ,
which I had the recollection was pending on your side? Maybe there was
a communication breakdown somewhere there. :)

Thanks,
Guillem

Re: Bug#922353: ITP: socket-activate -- Run a socket-activated daemon with minimal dependencies

[Guillem Jover]

Hi!

On Fri, 2019-02-15 at 10:46:43 -0500, Daniel Kahn Gillmor wrote:
> Control: clone 922353 -2
> Control: reassign -2 dpkg
> Control: retitle -2 start-stop-daemon should support socket-activation via 
> the sd_listen_fds(3) convention
> Control: severity -2 wishlist

Thanks.

> On Fri 2019-02-15 04:34:47 +0100, Guillem Jover wrote:
> > Another option would be to implement this in start-stop-daemon, like
> > the similar support for the systemd readiness protocol was recently
> > implemented there too.
> 
> Thanks for the suggestion!  How widely-distributed is start-stop-daemon
> outside of debian?  I see it's been ported to OpenBSD; are they
> syncing from upstream?

. For the BSDs to use
this more seriously the code would probably need to be split into its
own project, so that it does not pollute their licensing. Its current
"license" might also need to be clarified (PD), and "relicensed" into
MIT or similar.

This is something I've actually pondered doing anyway, so this might
be a good excuse, I guess.

> The code i have is just python3 right now (simple argument parsing made
> development much quicker), but it's not too terrible to do it in C.

Yes, and it should be pretty generic and portable.

> I'm opening this as a wishlist issue for dpkg just so we don't lose
> track of it, since it might take me some cycles to get the C
> implementation in shape.  If anyone else wants to beat me to it, i
> certainly wouldn't complain :)

I'll probably look into it once I've gone over some of the immediate
stuff I have on my plate, if there's been no patch submitted by then.
:)

Thanks,
Guillem

Re: Bug#922353: ITP: socket-activate -- Run a socket-activated daemon with minimal dependencies

[Daniel Kahn Gillmor]

Control: clone 922353 -2
Control: reassign -2 dpkg
Control: retitle -2 start-stop-daemon should support socket-activation via the 
sd_listen_fds(3) convention
Control: severity -2 wishlist

On Fri 2019-02-15 04:34:47 +0100, Guillem Jover wrote:
> Another option would be to implement this in start-stop-daemon, like
> the similar support for the systemd readiness protocol was recently
> implemented there too.

Thanks for the suggestion!  How widely-distributed is start-stop-daemon
outside of debian?  I see it's been ported to OpenBSD; are they
syncing from upstream?

The code i have is just python3 right now (simple argument parsing made
development much quicker), but it's not too terrible to do it in C.

It's great to hear that the dpkg codebase would welcome such a feature
addition if it's done right!

I'm opening this as a wishlist issue for dpkg just so we don't lose
track of it, since it might take me some cycles to get the C
implementation in shape.  If anyone else wants to beat me to it, i
certainly wouldn't complain :)

  --dkg


signature.asc
Description: PGP signature

Re: Bug#922353: ITP: socket-activate -- Run a socket-activated daemon with minimal dependencies

[Guillem Jover]

Hi!

On Thu, 2019-02-14 at 17:36:31 -0500, Daniel Kahn Gillmor wrote:
> Package: wnpp
> Severity: wishlist
> Owner: Daniel Kahn Gillmor 
> 
> * Package name: socket-activate
>   Version : 0.1
>   Upstream Author : Daniel Kahn Gillmor
> * URL : https://gitlab.com/dkg/socket-activate
> * License : GPL
>   Programming Lang: Python
>   Description : Run a socket-activated daemon with minimal dependencies

> socket-activate makes it possible to use socket-activated services on
> Unix-based systems that don't have systemd installed at all.
> 
> It implements the environment variable and file descriptor convention
> described in sd_listen_fds(3) without using any external code or
> dependencies from systemd.  The only thing it depends on is python3
> itself.
> 
> See https://bugs.debian.org/922082 for more discussion of the
> motivation for this project.  Future versions might be in C, so that
> even python isn't required, but the initial python implementation is
> intended to validate the interface.

Another option would be to implement this in start-stop-daemon, like
the similar support for the systemd readiness protocol was recently
implemented there too.

Thanks,
Guillem

Bug#922353: ITP: socket-activate -- Run a socket-activated daemon with minimal dependencies

[Daniel Kahn Gillmor]

Package: wnpp
Severity: wishlist
Owner: Daniel Kahn Gillmor 

* Package name: socket-activate
  Version : 0.1
  Upstream Author : Daniel Kahn Gillmor
* URL : https://gitlab.com/dkg/socket-activate
* License : GPL
  Programming Lang: Python
  Description : Run a socket-activated daemon with minimal dependencies

socket-activate makes it possible to use socket-activated services on
Unix-based systems that don't have systemd installed at all.

It implements the environment variable and file descriptor convention
described in sd_listen_fds(3) without using any external code or
dependencies from systemd.  The only thing it depends on is python3
itself.

See https://bugs.debian.org/922082 for more discussion of the
motivation for this project.  Future versions might be in C, so that
even python isn't required, but the initial python implementation is
intended to validate the interface.

Note that while i'm the responsible upstream on socket-activate, i'd
welcome any other contributions to it.