Re: [OE-core][PATCH v3] initscripts/init-system-helpers: fix mountnfs.sh dependency

["Jens Rehsack" <rehsack@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 2236 bytes --]



> Am 05.05.2020 um 16:56 schrieb Otavio Salvador <otavio.salvador@ossystems.com.br>:
> 
> On Tue, May 5, 2020 at 10:43 AM Jens Rehsack <rehsack@gmail.com> wrote:
>>> Am 04.05.2020 um 16:21 schrieb Otavio Salvador <otavio.salvador@ossystems.com.br>:
>>> 
>>> On Mon, May 4, 2020 at 11:10 AM Jens Rehsack <rehsack@gmail.com> wrote:
>>>> 
>>>> With commit c9fc9110be33fe0f24bc3a7c242b584a4ca33e04
>>>>   Author: Yue Tao <Yue.Tao@windriver.com>
>>>>   Date:   Fri May 25 10:48:08 2018 +0800
>>>> 
>>>>       initscripts: Avoid starting rpcbind daemon twice
>>>> 
>>>>       Check the status before start it to avoid duplicates.
>>>> 
>>>> the use of a script {/usr/sbin/}service is introduced - maybe earlier
>>>> provided by systemd, nowadays mostly by init-system-helpers from
>>>> debian project.
>>>> 
>>>> For the very first shot, maybe discussions and improvements based
>>>> on that script collection, use just the init-system-helpers-service
>>>> in initscripts/mountnfs.sh to avoid problems mounting NFS in later
>>>> boot stage.
>>>> 
>>>> Signed-off-by: Jens Rehsack <sno@netbsd.org>
>>> ...
>>>> +RDEPENDS_${PN}-update-rc.d = "perl"
>>> 
>>> Can't it be replaced? using perl here is not nice.
>> 
>> Be careful :P
> 
> Everyone has the right to do bad choices in life ;-P

/me nods
But comparing to up-river alternatives, there isn't much better :)

>> Seriously - it uses perl only for it's update-rc.d (which I didn't ripped off
>> to give community a chance to discuss it). Generally - especially with the
>> runit, s6, openrc ... in mind - we have to grow from there anyway and I like
>> our current update-rc.d implementation anyway.
>> 
>> But I would like then to have a 2nd commit (for documentation purposes) to
>> kill it or we just provide it and let people choose to use this one over
>> meta/recipes-core/update-rc.d
> 
> Generally, I'd prefer to not give the option as it causes more
> troubles for supporting it in future. I am totally fine with it being
> dropped on a second commit for sure.

What do others (Richard, Anuj, Khem) say?
And - making a patch series from that or let someone else kicking it?

Cheers
--
Jens Rehsack - rehsack@gmail.com


[-- Attachment #2: Message signed with OpenPGP --]
[-- Type: application/pgp-signature, Size: 849 bytes --]