[Buildroot] Move openldap package and enable LDAP server?

From: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
To: buildroot@busybox.net
Subject: [Buildroot] Move openldap package and enable LDAP server?
Date: Sat, 12 Dec 2015 14:01:45 +0100	[thread overview]
Message-ID: <20151212140145.5a7583dc@free-electrons.com> (raw)
In-Reply-To: <5669CC92.8080504@gmx.de>

Hello Andreas,

On Thu, 10 Dec 2015 20:03:46 +0100, Andreas Ehmanns wrote:

> please find attached the patch for adding the openldap server. It 
> contains the server option, moves the configuration to "Networking 
> Applications" and adds an init script for the ldap server.
> 
> Please let me know what you think about the patch and if there are 
> things that I should change.

Thanks. But could you please send the patch with 'git send-email'
rather than as an attachment? It allows the patch to be reviewed
properly, and also easily applied.

A few comments though:

 * The commit title is too long. Something like:

	openldap: add support to build the server

   would be preferable.

 * I think you should keep the "default y" on
   BR2_PACKAGE_OPENLDAP_CLIENTS in order to preserve the existing
   behavior, and not breaking things for current users.

 * In the S75sldapd script, I believe checking for the presence of the
   daemon and config file is not really needed.

   If the -4 in ARGS is to force IPv4 only, I believe it's not really
   great. Why not also support IPv6 ?

 * There is a fix of tabs and spaces for the indentation in the shell
   script, please try to be consistent.

 * Please use "printf" rather than "echo -n" in the shell script, since
   printf is POSIX, while "echo -n" is not.

 * We normally use a PID file, see S50dropbear for example. Is slapd
   already managing its own PID file ?

 * In the .mk file, the OPENLDAP_USERS definition is inside a condition
   that is commented out, it doesn't look good.

 * Don't do ifeq ($(BR2_PACKAGE_OPENLDAP_SERVER),n), but just ifeq
   ($(BR2_PACKAGE_OPENLDAP_SERVER),). An option will never have the
   value 'n': when an option is disabled, its value is the empty string.

Could you take those comments into account and send an updated version?

Thanks a lot!

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com