[oe] [PATCH] bitbake.conf: Don't remove nonexisting paths from FILESPATH

2010-04-22 Thread Chris Larson
From: clarson Doing so coupled with amend.bbclass results in caching an __depends which doesn't include all possible amend.inc files, only all the possible amend.inc files *within already existing directories*. Leaving all the possible dirs in FILESPATH resolves this. Signed-off-by: Chris Larso

Re: [oe] [PATCH] change INHIBIT_PACKAGE_STRIP to PACKAGE_STRIP with more options

2010-04-22 Thread Chris Larson
On Mon, Apr 19, 2010 at 1:41 AM, Roman I Khimov wrote: > Originally approached with "package.bbclass: introduce KEEP_GNU_DEBUGLINK": > http://patchwork.openembedded.org/patch/886/ > > Redone with a change from INHIBIT_PACKAGE_STRIP to PACKAGE_STRIP which > allows more options. By default it does

Re: [oe] [PATCH] python-2.6.4: make Modules/_ctypes/libffi configure

2010-04-22 Thread Martin Jansa
On Thu, Apr 22, 2010 at 5:00 PM, Antonio Ospite wrote: > This fixes configure issues with recent autoconf, e.g: >  autoreconf: Entering directory `Modules/_ctypes/libffi' >  autoreconf: configure.ac: not using Gettext >  autoreconf: running: aclocal --force >  configure.ac:26: error: m4_copy: won'

Re: [oe] [PATCH] python-2.6.4: make Modules/_ctypes/libffi configure

2010-04-22 Thread Koen Kooi
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 You forgot to bump PR, so this change won't get pickup by the package manager On 22-04-10 17:00, Antonio Ospite wrote: > This fixes configure issues with recent autoconf, e.g: > autoreconf: Entering directory `Modules/_ctypes/libffi' > autoreconf:

[oe] [PATCH] python-2.6.4: make Modules/_ctypes/libffi configure

2010-04-22 Thread Antonio Ospite
This fixes configure issues with recent autoconf, e.g: autoreconf: Entering directory `Modules/_ctypes/libffi' autoreconf: configure.ac: not using Gettext autoreconf: running: aclocal --force configure.ac:26: error: m4_copy: won't overwrite defined macro: \ _AC_ARG_VAR_PRECIOUS conf

Re: [oe] [oe-commits] Sergey Lapin : busybox: prevent postinst script from failing

2010-04-22 Thread Sergey Lapin
On Thu, Apr 22, 2010 at 6:05 PM, Roman I Khimov wrote: >> recent addition in rootfs_ipk.bbclass which runs postinst scripts with >> -e option, while very useful, prevents busybox from configuring at image >> build time. It will be configured at first image boot if your file >> system is writable,

Re: [oe] [oe-commits] Sergey Lapin : busybox: prevent postinst script from failing

2010-04-22 Thread Roman I Khimov
> recent addition in rootfs_ipk.bbclass which runs postinst scripts with > -e option, while very useful, prevents busybox from configuring at image > build time. It will be configured at first image boot if your file > system is writable, but in case of e.g. squashfs, that will be fatal > error. Th

Re: [oe] About having a static /dev

2010-04-22 Thread Koen Kooi
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 22-04-10 14:46, Antonio Ospite wrote: > On Thu, 22 Apr 2010 12:55:43 +0100 > Phil Blundell wrote: > >> On Thu, 2010-04-22 at 13:43 +0200, Antonio Ospite wrote: >>> 2. If I install any package which depends on udev then udev is brought >>> in

Re: [oe] About having a static /dev

2010-04-22 Thread Phil Blundell
On Thu, 2010-04-22 at 14:46 +0200, Antonio Ospite wrote: > A per-MACHINE way would be better IMHO, I am thinking about > adding a check on something like /dev/.staticdev (suggested in [1]) to > both udev init script and to the 'device' script from initscripts, and > make the image class add the /de

Re: [oe] About having a static /dev

2010-04-22 Thread Dr. Michael Lauer
Am 22.04.2010 um 14:46 schrieb Antonio Ospite: > On Thu, 22 Apr 2010 12:55:43 +0100 > Phil Blundell wrote: > >> On Thu, 2010-04-22 at 13:43 +0200, Antonio Ospite wrote: >>> 2. If I install any package which depends on udev then udev is brought >>>in and the static layout is gone at the next

Re: [oe] About having a static /dev

2010-04-22 Thread Antonio Ospite
On Thu, 22 Apr 2010 12:55:43 +0100 Phil Blundell wrote: > On Thu, 2010-04-22 at 13:43 +0200, Antonio Ospite wrote: > > 2. If I install any package which depends on udev then udev is brought > > in and the static layout is gone at the next boot. > > I think your best option is probably to ma

Re: [oe] About having a static /dev

2010-04-22 Thread Koen Kooi
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 22-04-10 13:43, Antonio Ospite wrote: > For my machine (a780) I'd like to have a static /dev whatever image I > build, I tried setting: > IMAGE_DEV_MANAGER = "" > IMAGE_DEVICE_TABLES = "files/device_table-minimal.txt \ >files

Re: [oe] About having a static /dev

2010-04-22 Thread Phil Blundell
On Thu, 2010-04-22 at 13:43 +0200, Antonio Ospite wrote: > 2. If I install any package which depends on udev then udev is brought > in and the static layout is gone at the next boot. I think your best option is probably to make udev be a DISTRO_FEATURE and not set it for your own distribution

[oe] About having a static /dev

2010-04-22 Thread Antonio Ospite
Hi, For my machine (a780) I'd like to have a static /dev whatever image I build, I tried setting: IMAGE_DEV_MANAGER = "" IMAGE_DEVICE_TABLES = "files/device_table-minimal.txt \ files/device_table-ezx.txt" But this is still not enough: 1. the 'devices' script from initscrip