[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#835520: [PATCH 00/11] Trim obsolete/harmful info about init integration



Hello Andreas,

Thanks for working on this, appreciated!

Andreas Henriksson [2016-12-06 15:07 +0100]:
>   Drop outdated "/run needs initscripts dependency"

Seconded.

>   Update init-system title to be more agnostic

Grammar isse, see below.

>   Drop obsolete paragraph about static runlevels and update-rc.d
>   Drop outdated paragraph about sequence numbers

Both seconded.

>   Add note about update-rc.d normally used via dh

Grammar/wording, see below.

>   Add "or equivalent" to must use invoke-rc.d paragraph

Seconded.

>   Drop legacy code from invoke-rc.d example
>   Add note about invoke-rc.d normally used via dh
>   Drop obsolete paragraph about rc.boot

All three seconded.

>   Replace init example by refering to dh-make

I don't like this in the curent form, see below.

>   Drop entire section 9.4 Console messages from init.d scripts

Seconded.


Comments about individual patches (patch title in [ ]):

Andreas Henriksson [2016-12-06 15:07 +0100]:
> [Update init-system title to be more agnostic]
> -	  <heading>Interfacing with the initscript system</heading>
> +	  <heading>Interfacing with init system</heading>

system"s", otherwise it sounds wrong (Josh already mentioned this).
Otherwise, seconded.

Andreas Henriksson [2016-12-06 15:07 +0100]:
> [Add note about update-rc.d normally used via dh]
> It might not be the policys place to define how the maintainer should

policy's

> automate the packaging work, but atleast mention debhelper to not

at least

> fool people into thinking manually writing maintainer scripts is
> the preferred method of using update-rc.d.
> ---
>  policy.sgml | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/policy.sgml b/policy.sgml
> index af2d1d2..be11109 100644
> --- a/policy.sgml
> +++ b/policy.sgml
> @@ -7712,6 +7712,14 @@ test -f <var>program-executed-later-in-script</var> || exit 0
>  	      please consult its man page <manref name="update-rc.d"
>  		section="8">.
>  	    </p>
> +
> +	    <p>
> +	      Note that usage of update-rc.d is normally not done directly
> +	      by the maintainer, but rather indirectly via the debhelper
> +	      programs that adds the required maintainer-script code

"usage is not done" sounds weird, and "adds" → "add -- How about

  Note that the packaging should normally not call update-rc.d
  directly but through debhelper programs that add the required
  update-rc.d calls automatically.

> +	      automatically. (See <prgn>dh_installinit</prgn>,
> +	      <prgn>dh_systemd_enable</prgn>, etc.)


Andreas Henriksson [2016-12-06 15:07 +0100]:
> [Drop legacy code from invoke-rc.d example]
> Since several releases the *-rc.d policy tools are always available

s/Since/For/   s/are always/have always been/

> on their system. At points we've had bugs that created corner-cases

s/their/the/ ?

> which make the *-rc.d tools go missing during an upgrade phase,
> but that should not be worked around in each and every package.
> Thus remove the check if invoke-rc.d is available (as it's a bug
> in some other package if it isn't these days).
> 
> Slightly reword the paragraph to not sound like should be converting

like "the developer" (or "packager"), or use passive voice ("not sound
like anything should be converted").

> something (which isn't the normal case these days), rather than just
> use invoke-rc.d.
> 
> Closes: #833177
> ---
>  policy.sgml | 10 ++--------
>  1 file changed, 2 insertions(+), 8 deletions(-)
> 
> diff --git a/policy.sgml b/policy.sgml
> index 444c33e..431bc47 100644
> --- a/policy.sgml
> +++ b/policy.sgml
> @@ -7749,17 +7749,11 @@ test -f <var>program-executed-later-in-script</var> || exit 0
>  	    </p>
>  
>  	    <p>
> -	      Most packages will simply need to change:
> -	      <example compact="compact">/etc/init.d/&lt;package&gt;
> -	      &lt;action&gt;</example> in their <prgn>postinst</prgn>
> -	      and <prgn>prerm</prgn> scripts to:
> +	      Most packages will simply use:
>  	      <example compact="compact">
> -	if which invoke-rc.d >/dev/null 2>&1; then
>  		invoke-rc.d <var>package</var> &lt;action&gt;
> -	else
> -		/etc/init.d/<var>package</var> &lt;action&gt;
> -	fi
>  	      </example>
> +	      in their <prgn>postinst</prgn> and <prgn>prerm</prgn> scripts.
>  	    </p>
>  
>  	    <p>
> -- 
> 2.10.2


Andreas Henriksson [2016-12-06 15:07 +0100]:
> [Replace init example by refering to dh-make]
>  	  <p>
> -	    An example on which you can base your
> -	    <file>/etc/init.d</file> scripts is found in
> -	    <file>/etc/init.d/skeleton</file>.
> +	    An example on which you can base your init integration on
> +	    is available in the output generated by <prgn>dh_make</prgn>.
>  	  </p>

We should absolutely drop the reference to /etc/init.d/skeleton as it
is not even present on a default install any more (it's Priority:
optional) and has several problems. I'm not familiar with dh-make, so
I can't second this as-is in good faith, but I agree with Andrey that
introducing this into the policy feels undesirable.

So as long as we don't have anyone wanting to fix sysvinit for this,
I'd rather just drop this paragraph entirely.

Thanks,

Martin
-- 
Martin Pitt                        | http://www.piware.de
Ubuntu Developer (www.ubuntu.com)  | Debian Developer  (www.debian.org)

Attachment: signature.asc
Description: PGP signature


Reply to: