All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Benoît Thébaudeau" <benoit.thebaudeau@advansee.com>
To: buildroot@busybox.net
Subject: [Buildroot] Startup files numbering policy
Date: Tue, 21 May 2013 23:10:40 +0200 (CEST)	[thread overview]
Message-ID: <537887127.1024616.1369170640439.JavaMail.root@advansee.com> (raw)
In-Reply-To: <20130521221707.64d0815d@skate>

Dear Thomas Petazzoni,

On Tuesday, May 21, 2013 10:17:07 PM, Thomas Petazzoni wrote:
> Dear Beno?t Th?baudeau,
> 
> On Tue, 21 May 2013 21:31:02 +0200 (CEST), Beno?t Th?baudeau wrote:
> 
> > So far, BuildRoot has always used arbitrary Sxx numbers for the startup
> > files in
> > /etc/init.d/, e.g. S91smb in the samba package. Depending on each board
> > needs,
> > these numbers might be more or less appropriate for the startup order. Do
> > you
> > plan to add a mechanism to choose non-default custom numbering for those
> > scripts, or do you have a policy such as "It's up to each board to run
> > post-build scripts for such adjustments."?
> 
> I don't think we will want a mechanism to customize that, a post-build
> script is definitely appropriate.
> 
> However, we could clarify, adjust and/or document the current numbering
> policy, in order to make it clearer and more usable.
> 
> What specific problems do you have with the current numbering? Would a
> change in the numbering policy solve those problems?
> 
> So far, in the projects I've done, I don't remember having issues with
> the numbering policy used by Buildroot. I just added my own scripts, at
> different levels, and the scripts installed by Buildroot were living
> enough "free" numbers at various stages in the initialization to do
> whatever I needed. But it is very well possible that I didn't come
> across the cases you have in your project.

I don't have the exact use case in mind, but I have encountered this issue at
least once on a project with many startup scripts. The board needed many custom
startup scripts in its target skeleton, and in the end there was a collision
with BuildRoot's predefined startup script numbers that broke the expected
startup order. It was also unclear how to number the board-specific startup
scripts before launching the 1st build, and without a list there is always the
risk of a collision when enabling a new package.

I don't think that changing the current script numbering would be a solution,
because that could cause an issue for someone else. There could perhaps be a
document listing the numbering of the startup scripts installed by all BuildRoot
packages. That would make it clear how to assign a new startup script number,
what to expect from BuildRoot, and how to perform custom adjustments (i.e.
post-build script).

Best regards,
Beno?t

  reply	other threads:[~2013-05-21 21:10 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <1882993282.1023650.1369164262752.JavaMail.root@advansee.com>
2013-05-21 19:31 ` [Buildroot] Startup files numbering policy Benoît Thébaudeau
2013-05-21 20:17   ` Thomas Petazzoni
2013-05-21 21:10     ` Benoît Thébaudeau [this message]
2013-05-21 21:24       ` Thomas Petazzoni
2013-05-21 22:08         ` Benoît Thébaudeau
2013-05-22  7:33           ` Thomas Petazzoni

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=537887127.1024616.1369170640439.JavaMail.root@advansee.com \
    --to=benoit.thebaudeau@advansee.com \
    --cc=buildroot@busybox.net \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.