Bug#960940: RFS: runit/2.1.2-37 [RC] -- system-wide service supervision

2020-09-16 Thread Lorenzo Puliti
Package: sponsorship-requests
Followup-For: Bug #960940

Dear mentors,

I'm still looking for a sponsor and I've uploaded again
with the following changes:

 * autopkgtest on #953875 (unwanted init switch) marked as superficial
 * Change the name of the special rundirectory to run selected service 
   in container: renamend from 'isolation' to 'solo' (to avoid confusion
   with docker isolation feature)
 * New runit-run package: do not restart the systemd service on upgrade.
   Restarting the whole supervision tree can be dangerous if certain services
   that do not support runtime restart are supervised (think dbus or any 
display manager)
   (this issue is unreported on BTS for now and affects runit-systemd package)
 * New runit-run package: use 'telinit q' to start the supervision tree after
   install and to stop it after removal, when running under sysvinit. 
   The current (somehow buggy) behaviour with runit-sysv package is that changes
   in /etc/inittab are effective only after a system reboot. 

Full datail below


 * Package name: runit
   Version : 2.1.2-37
   Upstream Author : Gerrit Pape 
 * URL : http://smarden.org/runit/
 * License : BSD-3-clause
 * Vcs : https://salsa.debian.org/debian/runit
   Section : admin

It builds those binary packages:

  runit - system-wide service supervision
  runit-run - service supervision (systemd and sysv integration)
  getty-run - runscripts to supervise getty processes
  runit-init - system-wide service supervision (as init system)

To access further information about this package, please visit the following 
URL:

  https://mentors.debian.net/package/runit/

Alternatively, one can download the package with dget using this command:

  dget -x https://mentors.debian.net/debian/pool/main/r/runit/runit_2.1.2-37.dsc

updated git repo:

  https://salsa.debian.org/Lorenzo.ru.g-guest/runit/-/tree/release-2.1.2-37

Changes since the last upload:

 runit (2.1.2-37) experimental; urgency=medium
 .
   * Bring back the test about forced-rescan
   * Fix forced rescan test failure on sbuild (Closes: #941322)
   * Clean the 'current' symlink on purge
   * Invoke-run: add a DEBUG option to finish-default
   * Finish-default: print invoke-run instead of runsv
   * Get non default rundirectory from kernel command line
   * Add a 'solo' rundirectory for containers (Closes: #950851)
   * Merge runit-sysv and runit-systemd into runit-run (Closes: #953875)
   * Add lintian overrides for manpage without executable
   * Update copyright for debian files
   * Add an autopkgtest to catch regression on #953875
   * runit-run: don't stop the service on upgrade (systemd)
   * runit-run: rescan inittab on install/remove (sysvinit)

Regards,
--
  Lorenzo Puliti



Bug#960940: RFS: runit/2.1.2-37 [RC] -- system-wide service supervision

2020-05-19 Thread Lorenzo Puliti
Package: sponsorship-requests
Followup-For: Bug #960940

On 5/19/20 12:12 AM, David Kalnischkies wrote:
> Hi,
>
> Disclaimer: Not sponsoring as I am oblivious to init systems of all
> kinds, so I don't feel confident reviewing changes in that area.
>
> (I am not stalking you – I promise! – I just couldn't let that
>  opportunity pass by after pushing another set of apt resolver changes
>  triggered by your packages :P)

Thanks you for your review and also for the quick response to bug reports that
I have made on apt :)

After your review, I've made the following changes;

* versioned Breaks+Replaces
* upload to experimental instead of unstable
* runit-run.postinst: add a newline before inserting the text
  avoid code duplication
* runit-run.postrm: remove newline at the end of the inittab file,
  but only if more than one is detected 

> Also, but that is bike shed material: "-run" ?
> I don't have a good suggestion, but I think for the uninitiated
> -run and -init will be indistinguishable (also: run-it-run, init!).

I haven't addressed this: the "-run" suffix is common for runit related
packages (and also runit users) and it's used to mark a tiny config package
that ships a runit service, for example
- getty-run --> runit service for getty
- git-daemon-run --> runit service for git daemon
- mini-httpd-run --> runit service for mini-httpd

So runit-run is for "run runit as a service under another init system" as
opposite to runit-init which is for "use runit as full init replacement".
I think the uninitiated user can get an idea by reading the package description
of runit, runit-init and runit-run.
Anyway I'm open to suggestions if someone thinks that the package description is
not helping or there is an idea for a better name then "runit-run".

I've re-uploaded the package:


Info:
https://mentors.debian.net/package/runit

Download:
dget -x 
https://mentors.debian.net/debian/pool/main/r/runit/runit_2.1.2-37.dsc

updated git repo:

https://salsa.debian.org/Lorenzo.ru.g-guest/runit/-/tree/release-2.1.2-37


Best Regards,
Lorenzo



-- System Information:
Debian Release: bullseye/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 5.5.9-van (SMP w/4 CPU cores; PREEMPT)
Kernel taint flags: TAINT_FIRMWARE_WORKAROUND, TAINT_OOT_MODULE
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), 
LANGUAGE=en_US:en (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: runit (via /run/runit.stopit)


Bug#960940: RFS: runit/2.1.2-37 [RC] -- system-wide service supervision

2020-05-18 Thread David Kalnischkies
Hi,

Disclaimer: Not sponsoring as I am oblivious to init systems of all
kinds, so I don't feel confident reviewing changes in that area.

(I am not stalking you – I promise! – I just couldn't let that
 opportunity pass by after pushing another set of apt resolver changes
 triggered by your packages :P)

On Mon, May 18, 2020 at 05:43:55PM +0200, Lorenzo Puliti wrote:
>* Merge runit-sysv and runit-systemd into runit-run (Closes: #953875)

You are using unversioned Breaks+Replaces – please make them versioned.
It is perhaps unlikely that packages with that name will reappear in the
archive, but never say never…

Related, but as a matter of style only, I would release package
reorganisations always to experimental first: a) you need to clear NEW
and b) there is usually at least one bug in it you want to fix quickly,
but a) makes it unpredictable then that emergency will be.
(also c) you need another upload to move to testing as NEW needs
binaries to be uploaded, but they wont go into testing).

Also, but that is bike shed material: "-run" ?
I don't have a good suggestion, but I think for the uninitiated
-run and -init will be indistinguishable (also: run-it-run, init!).


| debian/runit-run.postinst

I see that the same code was used before, but:
a) make that a function instead of two code copies
b) consider that a file might not have an ending newline:
$ echo -n 'foo' > foo
$ cat foo - < bar
heredoc> EOF
foobar
$


Best regards

David Kalnischkies


signature.asc
Description: PGP signature


Bug#960940: RFS: runit/2.1.2-37 [RC] -- system-wide service supervision

2020-05-18 Thread Lorenzo Puliti
Package: sponsorship-requests
Severity: normal

Dear mentors,

I am looking for a sponsor for my package "runit"

 * Package name: runit
   Version : 2.1.2-37
   Upstream Author : Gerrit Pape 
 * URL : http://smarden.org/runit/
 * License : BSD-3-clause
 * Vcs : https://salsa.debian.org/debian/runit
   Section : admin

It builds those binary packages:

  runit - system-wide service supervision
  runit-run - service supervision (systemd and sysv integration)
  getty-run - runscripts to supervise getty processes
  runit-init - system-wide service supervision (as init system)

To access further information about this package, please visit the following 
URL:

  https://mentors.debian.net/package/runit

Alternatively, one can download the package with dget using this command:

  dget -x https://mentors.debian.net/debian/pool/main/r/runit/runit_2.1.2-37.dsc

Updated git repo:

  https://salsa.debian.org/Lorenzo.ru.g-guest/runit/-/tree/release-2.1.2-37

Changes since the last upload:

   * Bring back the test about forced-rescan
   * Fix forced rescan test failure on sbuild (Closes: #941322)
   * Clean the current symlink on purge
   * Invoke-run: add a DEBUG option to finish-default
   * Finish-default: print invoke-run instead of runsv
   * Get non default rundirectory from kernel command line
   * Add an isolation rundirectory for containers (Closes: #950851)
   * Merge runit-sysv and runit-systemd into runit-run (Closes: #953875)
   * Add lintian overrides for manpage without executable
   * Update copyright for debian files
   * Add an autopkgtest to catch regression on 953875

Regards,

--
  Lorenzo Puliti



-- System Information:
Debian Release: bullseye/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 5.5.9-van (SMP w/4 CPU cores; PREEMPT)
Kernel taint flags: TAINT_FIRMWARE_WORKAROUND, TAINT_OOT_MODULE
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), 
LANGUAGE=en_US:en (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: runit (via /run/runit.stopit)