Re: [arch-general] [arch-dev-public] [PATCH 1/1] move initramfs generation from install script to pacman hook

2016-05-19 Thread Leonid Isaev
On Thu, May 19, 2016 at 2:13 PM, Mauro Santos  
wrote:
> On 19-05-2016 00:20, Sébastien Luttringer wrote:
>> Could we use a prefix convention to order our hooks? It's usefull to build
>> modules before building initramfs and eventually run grub update at the end.
>
> Not sure triggering grub update automagically is a good idea, I maintain
> grub.cfg myself and I'm sure more people do the same.
>
> I'd guess everyone that maintains their own grub.cfg would really
> appreciate not having a bad surprise after a kernel update.

I second that. Touching bootloader config is a bad idea. For example, in my
case, I use arch's syslinux to boot multiple distros (with custom
syslinux.cfg), which don't even have a bootloader package.

Thx,
L.

-- 
Leonid Isaev
GPG fingerprints: DA92 034D B4A8 EC51 7EA6  20DF 9291 EE8A 043C B8C4
  C0DF 20D0 C075 C3F1 E1BE  775A A7AE F6CB 164B 5A6D


Re: [arch-general] [arch-dev-public] [PATCH 1/1] move initramfs generation from install script to pacman hook

2016-05-19 Thread Stefan Tatschner
On 19.05.2016 22:24, Carsten Mattner wrote:
>> I'd guess everyone that maintains their own grub.cfg would really
>> appreciate not having a bad surprise after a kernel update.

But as a sysadmin I can configure my own hooks in /etc/hooks, right? So,
I could optionally add a hook which upates grub.


Re: [arch-general] [arch-dev-public] [PATCH 1/1] move initramfs generation from install script to pacman hook

2016-05-19 Thread Carsten Mattner
On Thu, May 19, 2016 at 2:13 PM, Mauro Santos
 wrote:
> On 19-05-2016 00:20, Sébastien Luttringer wrote:
>> Could we use a prefix convention to order our hooks? It's usefull to build
>> modules before building initramfs and eventually run grub update at the end.
>
> Not sure triggering grub update automagically is a good idea, I maintain
> grub.cfg myself and I'm sure more people do the same.
>
> I'd guess everyone that maintains their own grub.cfg would really
> appreciate not having a bad surprise after a kernel update.

I didn't know Arch didn't overwrite/regenerate grub.cfg,
but because distros usually do, I use syslinux instead.
Good to know grub is an options.


Re: [arch-general] Problem with pacman hooks, alphabetic order.

2016-05-19 Thread Christian Hesse
LoneVVolf  on Thu, 2016/05/19 17:24:
> On 14-05-16 01:15, Carsten Feuls wrote:
> > Hello Everybody,
> >
> > I have some trouble with pacman hooks.
> > Arch is going to use pacman hooks in every package.
> > etckeeper was one of the first package that use pacman hooks, without any
> > trouble.
> > But now it becomes more tricky to run.
> > My Problem is that the pacman hooks run in alphabetic order.
> > And not in a Prirority order.
> >
> > How this problem could be solved?
> > Yes I know I can number every hook but I prefer a more upstream 
> > solution..
> >
> >
> > Sincerly Yours
> > Carsten Feuls
> >  
> 
> I do think there may be better way to solve this then adding a priority 
> system for hooks.
> 
> this is current trigger used by etckeeper hooks :
> [Trigger]
> Operation = Install
> Operation = Upgrade
> Operation = Remove
> Type = Package
> Target = *
> 
> I think the purpose of etckeeper is to keep track of changes in the /etc 
> folder, right?
> 
> How about using this as trigger :
> 
> [Trigger]
> Operation = Install
> Operation = Upgrade
> Operation = Remove
> Type = File
> Target = /etc/*

That is what current version does. Well, except that the install root is
omitted, so correct target is just 'etc/*'.

But that does not solve the problem: Think of etckeeper hook runs
first, followed by ca-certificates hook. The changes are not committed.
That's why we want prefix for order, so etckeeper hook can run last.
-- 
main(a){char*c=/*Schoene Gruesse */"B?IJj;MEH"
"CX:;",b;for(a/*Best regards my address:*/=0;b=c[a++];)
putchar(b-1/(/*Chriscc -ox -xc - && ./x*/b/42*2-3)*42);}


pgpbtiMdeLjid.pgp
Description: OpenPGP digital signature


Re: [arch-general] Problem with pacman hooks, alphabetic order.

2016-05-19 Thread LoneVVolf

On 14-05-16 01:15, Carsten Feuls wrote:

Hello Everybody,

I have some trouble with pacman hooks.
Arch is going to use pacman hooks in every package.
etckeeper was one of the first package that use pacman hooks, without any
trouble.
But now it becomes more tricky to run.
My Problem is that the pacman hooks run in alphabetic order.
And not in a Prirority order.

How this problem could be solved?
Yes I know I can number every hook but I prefer a more upstream 
solution..



Sincerly Yours
Carsten Feuls



I do think there may be better way to solve this then adding a priority 
system for hooks.


this is current trigger used by etckeeper hooks :
[Trigger]
Operation = Install
Operation = Upgrade
Operation = Remove
Type = Package
Target = *

I think the purpose of etckeeper is to keep track of changes in the /etc 
folder, right?


How about using this as trigger :

[Trigger]
Operation = Install
Operation = Upgrade
Operation = Remove
Type = File
Target = /etc/*



Re: [arch-general] [arch-dev-public] [PATCH 1/1] move initramfs generation from install script to pacman hook

2016-05-19 Thread Mauro Santos
On 19-05-2016 13:33, Doug Newgard wrote:
> On Thu, 19 May 2016 13:13:51 +0100
> Mauro Santos  wrote:
> 
>> On 19-05-2016 00:20, Sébastien Luttringer wrote:
>>> Could we use a prefix convention to order our hooks? It's usefull to build
>>> modules before building initramfs and eventually run grub update at the 
>>> end.  
>>
>> Not sure triggering grub update automagically is a good idea, I maintain
>> grub.cfg myself and I'm sure more people do the same.
>>
>> I'd guess everyone that maintains their own grub.cfg would really
>> appreciate not having a bad surprise after a kernel update.
>>
> 
> I don't think he means as part of a package, but as something that the 
> sysadmin
> sets up. It's just an example of how ordering hooks can be helpful.
> 

The way the sentence was worded I understood that in the future it could
possibly be implemented as part of a package and raised the concern.
When given as an example, it's a perfectly reasonable example of what a
sysadmin might want to order properly :)

Sorry for the noise.

-- 
Mauro Santos


Re: [arch-general] [arch-dev-public] [PATCH 1/1] move initramfs generation from install script to pacman hook

2016-05-19 Thread Doug Newgard
On Thu, 19 May 2016 13:13:51 +0100
Mauro Santos  wrote:

> On 19-05-2016 00:20, Sébastien Luttringer wrote:
> > Could we use a prefix convention to order our hooks? It's usefull to build
> > modules before building initramfs and eventually run grub update at the 
> > end.  
> 
> Not sure triggering grub update automagically is a good idea, I maintain
> grub.cfg myself and I'm sure more people do the same.
> 
> I'd guess everyone that maintains their own grub.cfg would really
> appreciate not having a bad surprise after a kernel update.
> 

I don't think he means as part of a package, but as something that the sysadmin
sets up. It's just an example of how ordering hooks can be helpful.


Re: [arch-general] [arch-dev-public] [PATCH 1/1] move initramfs generation from install script to pacman hook

2016-05-19 Thread Mauro Santos
On 19-05-2016 00:20, Sébastien Luttringer wrote:
> Could we use a prefix convention to order our hooks? It's usefull to build
> modules before building initramfs and eventually run grub update at the end.

Not sure triggering grub update automagically is a good idea, I maintain
grub.cfg myself and I'm sure more people do the same.

I'd guess everyone that maintains their own grub.cfg would really
appreciate not having a bad surprise after a kernel update.

-- 
Mauro Santos