How to convert OE hardware platform from ARM into X86

How to convert OE hardware platform from ARM into X86

[刘芳华]

Hi!

I need to convert OE hardware platform from ARM into X86,but I don't know the correct step, I tried as follows:
      Change the contents of file  ~/stuff/build/conf/local.conf, the changes are:
                           MACHINE = "i586-generic";
                           #TARGET_ARCH = "arm"
                           TARGET_OS = "linux"
then, build ,but a error occured:
  ERROR: Build of /home/build/stuff/openembedded/packages/xorg-lib/libx11_1.2.1.bb do_compile failed
  ERROR: Task 3164 (/home/build/stuff/openembedded/packages/xorg-lib/libx11_1.2.1.bb, do_compile) failed

Now,I don't know what to do, I tried to do some changes,but this error still exit,call for help! 
   

                                                                                      LiuFanghua

_________________________________________________________________
Messenger安全保护中心，免费修复系统漏洞，保护Messenger安全！
http://im.live.cn/safe/
From marcin@juszkiewicz.com.pl Mon Jun 01 13:57:32 2009
Received: from [195.149.226.213] (helo=smtp.host4.kei.pl)
	by linuxtogo.org with esmtp (Exim 4.69)
	(envelope-from <marcin@juszkiewicz.com.pl>) id 1MB68y-0004is-P2
	for openembedded-devel@openembedded.org; Mon, 01 Jun 2009 13:57:32 +0200
Received: (qmail 21679 invoked by uid 813007); 1 Jun 2009 11:48:48 -0000
X-clamdmail: clamdmail 0.18a
Received: from 195.149.226.213 (HELO home.localnet)
	(marcin@juszkiewicz.com.pl@195.149.226.213)
	by 195.149.226.213 with ESMTPA; 1 Jun 2009 11:48:48 -0000
From: Marcin Juszkiewicz <marcin@juszkiewicz.com.pl>
To: openembedded-devel@lists.openembedded.org
Date: Mon, 1 Jun 2009 13:49:16 +0200
User-Agent: KMail/1.11.2 (Linux/2.6.30-rc6-00082-g7aff192-dirty; KDE/4.2.2;
	x86_64; ; )
References: <20090529233958.GA16815@misibook.miska.isageek.net>
	<20090530194215.GA1186@misibook.miska.isageek.net>
	<200906010053.30975.bluelightning@bluelightning.org>
In-Reply-To: <200906010053.30975.bluelightning@bluelightning.org>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Message-Id: <200906011349.17080.marcin@juszkiewicz.com.pl>
Subject: Re: [oe] [PATCH] zuc: fixed build
X-BeenThere: openembedded-devel@lists.openembedded.org
X-Mailman-Version: 2.1.11
Precedence: list
Reply-To: openembedded-devel@lists.openembedded.org
List-Id: Using the OpenEmbedded metadata to build Distributions
	<openembedded-devel.lists.openembedded.org>
List-Unsubscribe: <http://lists.linuxtogo.org/cgi-bin/mailman/options/openembedded-devel>,
	<mailto:openembedded-devel-request@lists.openembedded.org?subject=unsubscribe>
List-Archive: <http://lists.linuxtogo.org/pipermail/openembedded-devel>
List-Post: <mailto:openembedded-devel@lists.openembedded.org>
List-Help: <mailto:openembedded-devel-request@lists.openembedded.org?subject=help>
List-Subscribe: <http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel>,
	<mailto:openembedded-devel-request@lists.openembedded.org?subject=subscribe>
X-List-Received-Date: Mon, 01 Jun 2009 11:57:33 -0000

Dnia poniedzia=C5=82ek, 1 czerwca 2009 o 01:53:30 Paul Eggleton napisa=C5=
=82(a):
> I did notice that we are using an OZ sources mirror on
> ewi546.ewi.utwente.nl for a few recipes but it doesn't seem to be
> responding right now for me. Should we still be fetching from this
> server or do we need to switch to a different mirror?

This machine is no longer available so those recipes needs to be=20
changed.

Regards,=20
=2D-=20
JID:      hrw@jabber.org
Website:  http://marcin.juszkiewicz.com.pl/
LinkedIn: http://www.linkedin.com/in/marcinjuszkiewicz

Re: How to convert OE hardware platform from ARM into X86

[GNUtoo]

On Mon, 2009-06-01 at 18:39 +0800, 刘芳华 wrote:
> Hi!
> 
> I need to convert OE hardware platform from ARM into X86,but I don't know the correct step, I tried as follows:
>       Change the contents of file  ~/stuff/build/conf/local.conf, the changes are:
>                            MACHINE = "i586-generic";
>                            #TARGET_ARCH = "arm"
>                            TARGET_OS = "linux"
> then, build ,but a error occured:
>   ERROR: Build of /home/build/stuff/openembedded/packages/xorg-lib/libx11_1.2.1.bb do_compile failed
>   ERROR: Task 3164 (/home/build/stuff/openembedded/packages/xorg-lib/libx11_1.2.1.bb, do_compile) failed
we need more info on the build error
such as the log of the build
Denis.

Re: How to convert OE hardware platform from ARM into X86

[刘芳华]

Hi,Denis:
   Thank you for you attention!
   Let me describe the details for you now:
   The hardware target of OE in my system is ARM.And it was configured and changed by ourselves for some purpose.Now we need to convert the hardware target into X86 on the base of current environment. I didn't  know the exactly step,just tried by myself.But while I go back to the  ARM hardware target environment,after I build ,I meet error:

NOTE: Handling BitBake files: - (6654/6654) [100 %]
NOTE: Parsing finished. 6367 cached, 1 parsed, 286 skipped, 0 masked.
NOTE: build 200906051003: started
sh: git: command not found
sh: git: command not found

OE Build Configuration:
BB_VERSION        = "1.8.13"
METADATA_BRANCH   = "<unknown>"
METADATA_REVISION = "<unknown>"
TARGET_ARCH       = "arm"
TARGET_OS         = "linux-gnueabi"
MACHINE           = "qsd8x50"
DISTRO            = "millos"
DISTRO_VERSION    = "20090605020341"
TARGET_FPU        = "hard"

NOTE: Resolving any missing task queue dependencies
NOTE: preferred version 6.5.2 of mesa-dri not available (for item libgl)
NOTE: preferred version 6.5.2 of mesa-dri not available (for item libglu)
NOTE: preferred version 1.15 of lcms not available (for item lcms)
NOTE: preferred version 0.2 of ttf-liberation not available (for item ttf-liberation)
NOTE: preferred version 2.34 of gpsd not available (for item python-pygps)
NOTE: preferred version 2.34 of gpsd not available (for item gpsd)
NOTE: preferred version 2.1.4 of libmusicbrainz not available (for item libmusicbrainz)
NOTE: preferred version 0.1.5 of vala not available (for item vala)
NOTE: preferred version 1.15 of lcms not available (for item lcms)
NOTE: Runtime target 'gnome-system-tools' is unbuildable, removing...
Missing or unbuildable dependency chain was: ['gnome-system-tools', 'liboobs-1']
ERROR: Required build target 'millos-image' has no buildable providers.
Missing or unbuildable dependency chain was: ['millos-image', 'gnome-system-tools', 'liboobs-1']


I tried my best,but still can't resovel the error,help me please,and tell me the correct step for this convert!

 FangHua




> From: GNUtoo@no-log.org
> To: openembedded-devel@lists.openembedded.org
> Date: Thu, 4 Jun 2009 21:38:40 +0200
> Subject: Re: [oe] How to convert OE hardware platform from ARM into X86
> 
> On Mon, 2009-06-01 at 18:39 +0800, 刘芳华 wrote:
> > Hi!
> > 
> > I need to convert OE hardware platform from ARM into X86,but I don't know the correct step, I tried as follows:
> >       Change the contents of file  ~/stuff/build/conf/local.conf, the changes are:
> >                            MACHINE = "i586-generic";
> >                            #TARGET_ARCH = "arm"
> >                            TARGET_OS = "linux"
> > then, build ,but a error occured:
> >   ERROR: Build of /home/build/stuff/openembedded/packages/xorg-lib/libx11_1.2.1.bb do_compile failed
> >   ERROR: Task 3164 (/home/build/stuff/openembedded/packages/xorg-lib/libx11_1.2.1.bb, do_compile) failed
> we need more info on the build error
> such as the log of the build
> Denis.
> 
> 
> 
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

_________________________________________________________________
Messenger安全保护中心，免费修复系统漏洞，保护Messenger安全！
http://im.live.cn/safe/
From elvis.dowson@mac.com Fri Jun 05 09:12:25 2009
Received: from [17.148.16.89] (helo=asmtpout014.mac.com)
	by linuxtogo.org with esmtp (Exim 4.69)
	(envelope-from <elvis.dowson@mac.com>) id 1MCTbE-0004Ln-Ju
	for openembedded-devel@lists.openembedded.org;
	Fri, 05 Jun 2009 09:12:25 +0200
MIME-version: 1.0
Received: from [87.201.169.29] by asmtp014.mac.com
	(Sun Java(tm) System Messaging Server 6.3-8.01 (built Dec 16 2008;
	32bit)) with ESMTPSA id <0KKR00GWB7L9FF50@asmtp014.mac.com> for
	openembedded-devel@lists.openembedded.org; Fri,
	05 Jun 2009 00:03:14 -0700 (PDT)
Message-id: <A98BBCED-33BD-498A-A42F-2BFB11C2D1ED@mac.com>
From: Elvis Dowson <elvis.dowson@mac.com>
To: OpenEmbedded Developers <openembedded-devel@lists.openembedded.org>
Date: Fri, 05 Jun 2009 11:03:07 +0400
X-Mailer: Apple Mail (2.935.3)
Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes
Content-Transfer-Encoding: 7BIT
X-Content-Filtered-By: Mailman/MimeDel 2.1.11
Subject: [oe] How to copy Ralink RT2571w firmware image to /lib/firmware
	folder
X-BeenThere: openembedded-devel@lists.openembedded.org
X-Mailman-Version: 2.1.11
Precedence: list
Reply-To: openembedded-devel@lists.openembedded.org
List-Id: Using the OpenEmbedded metadata to build Distributions
	<openembedded-devel.lists.openembedded.org>
List-Unsubscribe: <http://lists.linuxtogo.org/cgi-bin/mailman/options/openembedded-devel>,
	<mailto:openembedded-devel-request@lists.openembedded.org?subject=unsubscribe>
List-Archive: <http://lists.linuxtogo.org/pipermail/openembedded-devel>
List-Post: <mailto:openembedded-devel@lists.openembedded.org>
List-Help: <mailto:openembedded-devel-request@lists.openembedded.org?subject=help>
List-Subscribe: <http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel>,
	<mailto:openembedded-devel-request@lists.openembedded.org?subject=subscribe>
X-List-Received-Date: Fri, 05 Jun 2009 07:12:25 -0000

Hi,
	I have created a bitbake recipe for the Ralink RT2571w 1.6.4 WiFi  
network module.

Part of the installation process requires me to copy the a firmware  
file called rt73.bin to the /lib/firmware folder.

How can I do this?

Here is a link to the firmware file: http://www.ralinktech.com.tw/data/RT71W_Firmware_V1.8.zip

Here is a copy of my bitbake recipe:

Filename: rt2570-k2wrlz_1.6.4.bb

DESCRIPTION = "Enhanced Driver for Ralink rt2570 USB 802.11g WiFi  
sticks"
HOMEPAGE = "http://homepages.tu-darmstadt.de/~p_larbig/wlan"
SECTION = "kernel/modules"
LICENSE = "GPL"

PR = "r1"

SRC_URI = "http://homepages.tu-darmstadt.de/~p_larbig/wlan/${PN}-$ 
{PV}.tar.bz2"

inherit module

S = "${WORKDIR}/${PN}-${PV}/Module/"

do_compile_prepend_arm () {
   MAKE_TARGETS="arm"
}

do_compile() {
   export KERNDIR=${STAGING_KERNEL_DIR}
   module_do_compile
}

do_install() {
	install -d ${D}${base_libdir}/modules/${KERNEL_VERSION}/extra
	install -m 0644 rt2570${KERNEL_OBJECT_SUFFIX} ${D}${base_libdir}/ 
modules/${KERNEL_VERSION}/extra/
}


Best regards,

Elvis

Re: How to convert OE hardware platform from ARM into X86

[GNUtoo]

On Fri, 2009-06-05 at 10:12 +0800, 刘芳华 wrote:
> The hardware target of OE in my system is ARM.And it was configured
> and changed by ourselves for some purpose.Now we need to convert the
> hardware target into X86 on the base of current environment. I didn't
> know the exactly step,just tried by myself.But while I go back to the
> ARM hardware target environment,
hi, why don't you use 2 different local.conf with 2 different local
directories?(note that you can't rename or move a temp directory so
chose carefully its name at the beginning) and use a script to switch
between them with simlinks?
for instance for the openmoko I use:
$ cat openmoko.sh 
cd build/conf/
unlink local.conf 
ln -s openmoko.conf local.conf
cd ${HOME}/
unlink oetmp
ln -s oetmp_openmoko oetmp

and here's my local.conf:
$ cat build/conf/local.conf | sed 's/#.*//' | sed '/^$/d'
DL_DIR = "${HOME}/sources"
BBFILES="${HOME}/oe/local_recipes/recipes/*/*.bb
${HOME}/oe/org.openembedded.dev/recipes/*/*.bb"
BBFILE_COLLECTIONS = "local upstream"
BBFILE_PATTERN_upstream = "^${HOME}/oe/org.openembedded.dev/recipes/"
BBFILE_PATTERN_local = "^${HOME}/oe/local_recipes/recipes/"
BBFILE_PRIORITY_upstream = "1"
BBFILE_PRIORITY_local = "2"
BBMASK = ""
PREFERRED_PROVIDERS = "virtual/qte:qte virtual/libqpe:libqpe-opie"
PREFERRED_PROVIDERS += " virtual/libsdl:libsdl-x11"
PREFERRED_PROVIDERS += "
virtual/${TARGET_PREFIX}gcc-initial:gcc-cross-initial"
PREFERRED_PROVIDERS += "
virtual/${TARGET_PREFIX}gcc-intermediate:gcc-cross-intermediate"
PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}gcc:gcc-cross"
PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}g++:gcc-cross"
TMPDIR = ${HOME}/oetmp_openmoko
MACHINE = "om-gta02"
DISTRO = "openmoko"
IMAGE_FSTYPES = "tar"
BBINCLUDELOGS = "yes"
PREFERRED_VERSION_qemu-native = "0.10.3"
PREFERRED_VERSION_qemu = "0.10.3"
PREFERRED_VERSION_frozen-bubble = "2.2.0"
PREFERRED_VERSION_libsdl-mixer = "1.2.8"
PREFERRED_VERSION_boost = "1.38.0"
PREFERRED_VERSION_boost-jam-native = "3.1.17"
PREFERRED_VERSION_wesnoth = "1.6a"
PREFERRED_VERSION_gnash = "0.8.5"
PREFERRED_VERSION_libosip2 = "2.2.2"
PREFERRED_VERSION_openssh = "5.2p1"
OLDEST_KERNEL = "2.6.28"

note my tmpdir is not oetmp but oetmp_openmoko
the oetmp simlink is only for letting me find the good directory

Denis.

Re: How to convert OE hardware platform from ARM into X86

[GNUtoo]

On Fri, 2009-06-05 at 10:12 +0800, 刘芳华 wrote:
> NOTE: Runtime target 'gnome-system-tools' is unbuildable, removing...
> Missing or unbuildable dependency chain was: ['gnome-system-tools',
> 'liboobs-1']
> ERROR: Required build target 'millos-image' has no buildable
> providers.
> Missing or unbuildable dependency chain was: ['millos-image',
> 'gnome-system-tools', 'liboobs-1']
I've done a fake recipe named gnutoo with a non-existent rdepend and I
got:
$ bitbake gnutoo
__builtin__:74: DeprecationWarning: the sets module is deprecated
__builtin__:74: DeprecationWarning: the sets module is deprecated
NOTE: Handling BitBake files: - (6881/6881) [100 %]
NOTE: Parsing finished. 6602 cached, 1 parsed, 278 skipped, 0 masked.
NOTE: Resolving any missing task queue dependencies
ERROR:
'['/home/embedded/oe/local_recipes/recipes/test/gnutoo_1.0.0.bb']'
RDEPENDS/RRECOMMENDS or otherwise requires the runtime entity
'aslkjdalskdj' but it wasn't found in any PACKAGE or RPROVIDES variables
NOTE: Runtime target 'aslkjdalskdj' is unbuildable, removing...
Missing or unbuildable dependency chain was: ['aslkjdalskdj']
ERROR: Required build target 'gnutoo' has no buildable providers.
Missing or unbuildable dependency chain was: ['gnutoo', 'aslkjdalskdj']
so...it seems that you don't have gnome-system-tools and liboobs-1
already built(by the way I don't find them in oe) or at all
Denis.

OE Build :error configure failed to discover size of unix socket path

[刘芳华]

Hi!

   I met a error while I build:

/mono/io-layer/wapi-private.h:22:2: error: #error configure failed to discover size of unix socket path
| make[3]: *** [collection.lo] Error 1
| make[3]: Leaving directory `/home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/mono-2.4/mono/io-layer'
| make[2]: *** [all-recursive] Error 1
| make[2]: Leaving directory `/home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/mono-2.4/mono'
| make[1]: *** [all-recursive] Error 1
| make[1]: Leaving directory `/home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/mono-2.4'
| make: *** [all] Error 2
| FATAL: oe_runmake failed
NOTE: Task failed: /home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/temp/log.do_compile.10880
NOTE: package mono-2.4-r0: task do_compile: failed
ERROR: TaskFailed event exception, aborting
NOTE: package mono-2.4: failed
ERROR: Build of /home/build/stuff/openembedded/packages/mono/mono_2.4.bb do_compile failed
ERROR: Task 9 (/home/build/stuff/openembedded/packages/mono/mono_2.4.bb, do_compile) failed
NOTE: Tasks Summary: Attempted 7 tasks of which 0 didn't need to be rerun and 1 failed.
ERROR: '/home/build/stuff/openembedded/packages/mono/mono_2.4.bb' failed
NOTE: build 200906101437: completed

Any advice will be appreciated!


_________________________________________________________________
Messenger10年嘉年华，礼品大奖等你拿！
http://10.msn.com.cn
From esbenhaabendal@gmail.com Wed Jun 10 09:03:20 2009
Received: from [74.125.92.146] (helo=qw-out-1920.google.com)
	by linuxtogo.org with esmtp (Exim 4.69)
	(envelope-from <esbenhaabendal@gmail.com>) id 1MEHq9-00009K-42
	for openembedded-devel@lists.openembedded.org;
	Wed, 10 Jun 2009 09:03:18 +0200
Received: by qw-out-1920.google.com with SMTP id 14so359686qwa.36
	for <openembedded-devel@lists.openembedded.org>;
	Tue, 09 Jun 2009 23:53:37 -0700 (PDT)
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=gamma;
	h=domainkey-signature:mime-version:received:in-reply-to:references
	:date:message-id:subject:from:to:content-type
	:content-transfer-encoding;
	bh=aFGK1ifdLqUdTxrlAisQXzgHU/p29kiCpVPFUu3PFmg=;
	b=qT66Rt5fawbhkrzwRshCgCx+bGZ+qua04kSSJIq35pQNv7IH0DIETFB7Q9A7+G52gb
	NMCSb0SKpBCH5KEHMZdxH8nLC2FqUQIzemlRXac+CHP1BT6EhgC1w7TSg7/fR7un0wdj
	dNMC2GwEW6CwHDjl26ultzsfCZVRCTGhADzfY=
DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma;
	h=mime-version:in-reply-to:references:date:message-id:subject:from:to
	:content-type:content-transfer-encoding;
	b=cC/ORus5B9t45/YHWRi7r2tXnjZeiqQ2oy69jSPznJLrgKTbDalqMkKI7r6ryXGULo
	xQDXM371O3ZrcdULxopmfNyvt53E/4GhD2F9vKLCqrKMzyiuJkt94iZgqCCXDPsYIei7
	lJin2Yljlol8V7nwDlK0mdYH5+O8p/2xiyZuM=
MIME-Version: 1.0
Received: by 10.224.67.1 with SMTP id p1mr1368984qai.263.1244616817166; Tue, 
	09 Jun 2009 23:53:37 -0700 (PDT)
In-Reply-To: <200906091121.59329.mickey@vanille-media.de>
References: <f96d234e0904300544g4f933c15x1abaa430373868fa@mail.gmail.com>
	<f96d234e0906051443y3d0cfcf0m595469c780c9fed@mail.gmail.com>
	<200906060846.30822.mickey@vanille-media.de>
	<200906091121.59329.mickey@vanille-media.de>
Date: Wed, 10 Jun 2009 08:53:37 +0200
Message-ID: <d2b9ea600906092353o4fe9ea02vbc7df3b103b401c4@mail.gmail.com>
From: Esben Haabendal <esbenhaabendal@gmail.com>
To: openembedded-devel@lists.openembedded.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
Subject: Re: [oe] progress on OE organization issues
X-BeenThere: openembedded-devel@lists.openembedded.org
X-Mailman-Version: 2.1.11
Precedence: list
Reply-To: openembedded-devel@lists.openembedded.org
List-Id: Using the OpenEmbedded metadata to build Distributions
	<openembedded-devel.lists.openembedded.org>
List-Unsubscribe: <http://lists.linuxtogo.org/cgi-bin/mailman/options/openembedded-devel>,
	<mailto:openembedded-devel-request@lists.openembedded.org?subject=unsubscribe>
List-Archive: <http://lists.linuxtogo.org/pipermail/openembedded-devel>
List-Post: <mailto:openembedded-devel@lists.openembedded.org>
List-Help: <mailto:openembedded-devel-request@lists.openembedded.org?subject=help>
List-Subscribe: <http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel>,
	<mailto:openembedded-devel-request@lists.openembedded.org?subject=subscribe>
X-List-Received-Date: Wed, 10 Jun 2009 07:03:20 -0000

On Tue, Jun 9, 2009 at 11:21 AM, Michael 'Mickey'
Lauer<mickey@vanille-media.de> wrote:
> Alas, right now we have
>
> =A0* 20 subscribers to the member-list, of
> =A0* 70 folks with commit access, of
> =A0* several hundred subscribers to the devel-list,
> =A0let alone all the stakeholders "in the dark" ;)
>
> So, please join the e.V. and request membership to the e.V. via the membe=
rs
> list, in case you did not have already.

As for the last bullet point, what is the status of membership for
those of us without commit access? I seem to have seen mixed signals
wether we are invited to be member of the e.V.  Some messages
indicates that only people with commit access are going to be
accepted, while others (like mickey) request all interested in OE to
join the e.V.

Could someone put a small statement on Wiki on this?

/Esben

--=20
Esben Haabendal, Senior Software Consultant
Dor=E9Development ApS, Ved Stranden 1, 9560 Hadsund, DK-Denmark
Phone: +45 51 92 53 93, E-mail: eha@doredevelopment.dk
WWW: http://www.doredevelopment.dk

Re: OE Build :error configure failed to discover size of unix socket path

[Lin_Lynn]

Maybe you need give more information when you submit a bug
you can post the log file


Lynn Lin
-----Original Message-----
From: openembedded-devel-bounces@openembedded.org [mailto:openembedded-devel-bounces@openembedded.org] On Behalf Of 刘芳华
Sent: Wednesday, June 10, 2009 2:49 PM
To: openembedded-devel@openembedded.org
Subject: [oe] OE Build :error configure failed to discover size of unix socket path



Hi!

   I met a error while I build:

/mono/io-layer/wapi-private.h:22:2: error: #error configure failed to discover size of unix socket path
| make[3]: *** [collection.lo] Error 1
| make[3]: Leaving directory `/home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/mono-2.4/mono/io-layer'
| make[2]: *** [all-recursive] Error 1
| make[2]: Leaving directory `/home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/mono-2.4/mono'
| make[1]: *** [all-recursive] Error 1
| make[1]: Leaving directory `/home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/mono-2.4'
| make: *** [all] Error 2
| FATAL: oe_runmake failed
NOTE: Task failed: /home/build/stuff/workspace/work/i586-ThunderST-linux/mono-2.4-r0/temp/log.do_compile.10880
NOTE: package mono-2.4-r0: task do_compile: failed
ERROR: TaskFailed event exception, aborting
NOTE: package mono-2.4: failed
ERROR: Build of /home/build/stuff/openembedded/packages/mono/mono_2.4.bb do_compile failed
ERROR: Task 9 (/home/build/stuff/openembedded/packages/mono/mono_2.4.bb, do_compile) failed
NOTE: Tasks Summary: Attempted 7 tasks of which 0 didn't need to be rerun and 1 failed.
ERROR: '/home/build/stuff/openembedded/packages/mono/mono_2.4.bb' failed
NOTE: build 200906101437: completed

Any advice will be appreciated!


_________________________________________________________________
Messenger10年嘉年华，礼品大奖等你拿！
http://10.msn.com.cn
_______________________________________________
Openembedded-devel mailing list
Openembedded-devel@lists.openembedded.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel

Re: OE Build :error configure failed to discover size of unix socket path

[Stanislav Brabec]

刘芳华 wrote:
> 
> Hi!
> 
>    I met a error while I build:
> 
> /mono/io-layer/wapi-private.h:22:2: error: #error configure failed to discover size of unix socket path

MONO_SIZEOF_SUNPATH was not detected.

While cross-compiling you have to provide this value manually - for
example run configure once on the target platform and then read the
value in config.h.

Then you need to set mono_cv_sizeof_sunpath in
conf/{your_platform_file}. See site/arm-linux for example. If it works,
you can submit a patch.

-- 
Stanislav Brabec
http://www.penguin.cz/~utx/zaurus

Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Marc Olzheim]

[-- Attachment #1.1: Type: text/plain, Size: 255 bytes --]

I'm not sure how you tested this, but there are quotes missing, so that
the "-E" argument to "grep" is passed to "export" instead. I guess some
shells have a "-E" flag to "export", so it is silently ignored, but mine
don't.

Patch attached.

Marc

[-- Attachment #1.2: oe-libtool.patch --]
[-- Type: text/x-diff, Size: 1878 bytes --]

commit 79eb6925a932b93a0ebcbdfa96f74a9cc5f8c99d
Author: Marc Olzheim <marc@iphion.nl>
Date:   Thu Jul 30 12:48:46 2009 +0200

    2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 broke do_configure.

diff --git a/recipes/libtool/libtool_2.2.4.bb b/recipes/libtool/libtool_2.2.4.bb
index ad4b0a0..5d59ce0 100644
--- a/recipes/libtool/libtool_2.2.4.bb
+++ b/recipes/libtool/libtool_2.2.4.bb
@@ -17,10 +17,10 @@ EXTRA_AUTORECONF = "--exclude=libtoolize"
 do_configure_prepend () {
 	# Skip this for native build:
 	if test -n "$CONFIG_SITE" ; then
-		export ac_cv_path_SED=${ac_cv_path_GREP=/bin/sed}
-		export ac_cv_path_GREP=${ac_cv_path_GREP=/bin/grep}
-		export ac_cv_path_EGREP=${ac_cv_path_EGREP=/bin/grep -E}
-		export ac_cv_path_FGREP=${ac_cv_path_FGREP=/bin/grep -F}
+		export ac_cv_path_SED="${ac_cv_path_GREP=/bin/sed}"
+		export ac_cv_path_GREP="${ac_cv_path_GREP=/bin/grep}"
+		export ac_cv_path_EGREP="${ac_cv_path_EGREP=/bin/grep -E}"
+		export ac_cv_path_FGREP="${ac_cv_path_FGREP=/bin/grep -F}"
 	fi
 }
 
diff --git a/recipes/libtool/libtool_2.2.6a.bb b/recipes/libtool/libtool_2.2.6a.bb
index 6268576..6c87660 100644
--- a/recipes/libtool/libtool_2.2.6a.bb
+++ b/recipes/libtool/libtool_2.2.6a.bb
@@ -17,10 +17,10 @@ EXTRA_AUTORECONF = "--exclude=libtoolize"
 do_configure_prepend () {
 	# Skip this for native build:
 	if test -n "$CONFIG_SITE" ; then
-		export ac_cv_path_SED=${ac_cv_path_GREP=/bin/sed}
-		export ac_cv_path_GREP=${ac_cv_path_GREP=/bin/grep}
-		export ac_cv_path_EGREP=${ac_cv_path_EGREP=/bin/grep -E}
-		export ac_cv_path_FGREP=${ac_cv_path_FGREP=/bin/grep -F}
+		export ac_cv_path_SED="${ac_cv_path_GREP=/bin/sed}"
+		export ac_cv_path_GREP="${ac_cv_path_GREP=/bin/grep}"
+		export ac_cv_path_EGREP="${ac_cv_path_EGREP=/bin/grep -E}"
+		export ac_cv_path_FGREP="${ac_cv_path_FGREP=/bin/grep -F}"
 	fi
 }
 

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 204 bytes --]

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Stanislav Brabec]

Marc Olzheim wrote:
> I'm not sure how you tested this, but there are quotes missing, so that
> the "-E" argument to "grep" is passed to "export" instead. I guess some
> shells have a "-E" flag to "export", so it is silently ignored, but mine
> don't.

I tried both versions - quoted and unquoted - before the commit. Both
worked for me, that is why I committed the shorter one.

Well, adding quotes is safe. Can you check your configure output and
host shell version?

My configure output:

checking for a sed that does not truncate output... (cached) /bin/sed
checking for grep that handles long lines and -e... (cached) /bin/sed
checking for egrep... (cached) /bin/grep -E
checking for fgrep... (cached) /bin/grep -F

It indicates, that it worked correctly.

My native shell is:

/OE/openembedded/build> sh --version
GNU bash, version 4.0.24(1)-release (x86_64-suse-linux-gnu)

-- 
Stanislav Brabec
http://www.penguin.cz/~utx/zaurus

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Marc Olzheim]

[-- Attachment #1: Type: text/plain, Size: 1773 bytes --]

On Thu, Jul 30, 2009 at 01:14:24PM +0200, Stanislav Brabec wrote:
> > I'm not sure how you tested this, but there are quotes missing, so that
> > the "-E" argument to "grep" is passed to "export" instead. I guess some
> > shells have a "-E" flag to "export", so it is silently ignored, but mine
> > don't.
> 
> I tried both versions - quoted and unquoted - before the commit. Both
> worked for me, that is why I committed the shorter one.
> 
> Well, adding quotes is safe. Can you check your configure output and
> host shell version?
> 
> My configure output:
> 
> checking for a sed that does not truncate output... (cached) /bin/sed
> checking for grep that handles long lines and -e... (cached) /bin/sed
> checking for egrep... (cached) /bin/grep -E
> checking for fgrep... (cached) /bin/grep -F
> 
> It indicates, that it worked correctly.
> 
> My native shell is:
> 
> /OE/openembedded/build> sh --version
> GNU bash, version 4.0.24(1)-release (x86_64-suse-linux-gnu)

Hmm, what I got was:

(task 365) ERROR: function do_configure failed
(task 365) ERROR: log data follows
(/home/marcolz/src/iphion-new-oe3/embbuild/tmp/work/i686-oe-linux/libtool-cross-2.2.6a-r0/temp/log.do_configure.22690)
| export: 929: -E: bad variable name
(task 365) NOTE: Task failed:
/home/marcolz/src/iphion-new-oe3/embbuild/tmp/work/i686-oe-linux/libtool-cross-2.2.6a-r0/temp/log.do_configure.22690
(task 365) ERROR: TaskFailed event exception, aborting
(task 365) ERROR: Build of
/home/marcolz/src/iphion-new-oe3/embbuild/openembedded.iphion/recipes/libtool/libtool-cross_2.2.6a.bb
do_configure failed


My local bash version is:

GNU bash, version 3.2.39(1)-release (i486-pc-linux-gnu)
Copyright (C) 2007 Free Software Foundation, Inc.

Marc

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 204 bytes --]

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Stanislav Brabec]

Marc Olzheim wrote:

> (/home/marcolz/src/iphion-new-oe3/embbuild/tmp/work/i686-oe-linux/libtool-cross-2.2.6a-r0/temp/log.do_configure.22690)
> | export: 929: -E: bad variable name

> My local bash version is:
> 
> GNU bash, version 3.2.39(1)-release (i486-pc-linux-gnu)

OK. Next time I will not use unquoted ${a=b c}. Some shells don't like
it.

Committed-By: Stanislav Brabec <utx@penguin.cz>

Just curious, is your /bin/sh link to bash? I am not able to reproduce:

~> sh --version
malloc: using debugging hooks
GNU bash, version 3.2.39(1)-release (x86_64-suse-linux-gnu)
Copyright (C) 2007 Free Software Foundation, Inc.
~> export a=${a=b c}
~> echo $a
b c

-- 
Stanislav Brabec
http://www.penguin.cz/~utx/zaurus

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Marc Olzheim]

[-- Attachment #1: Type: text/plain, Size: 748 bytes --]

On Thu, Jul 30, 2009 at 02:14:13PM +0200, Stanislav Brabec wrote:
> Just curious, is your /bin/sh link to bash? I am not able to reproduce:
> 
> ~> sh --version
> malloc: using debugging hooks
> GNU bash, version 3.2.39(1)-release (x86_64-suse-linux-gnu)
> Copyright (C) 2007 Free Software Foundation, Inc.
> ~> export a=${a=b c}
> ~> echo $a
> b c

Ah, *doh*...

My /bin/sh is a dash, I should have known, my apologies.  I have a patch
in place that allows dash as /bin/sh as well, as that works just as well
for me (and /bin/bash for /bin/sh breaks other scripts for me), up until
now.

Building the micro-image, the libtool build was suddenly failing though.

Still I should have known that this would be the issue.

Marc

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 204 bytes --]

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Leon Woestenberg]

Hello,

On Thu, Jul 30, 2009 at 2:50 PM, Marc Olzheim<zlo@zlo.nu> wrote:

> My /bin/sh is a dash, I should have known, my apologies.  I have a patch
>
Testing with dash rather than bash is GOOD due to it adhering to POSIX
without much of the bash-specifics.

So please do not introduce 'bashisms'  (dependencies on bash behaviour).

Regards,
-- 
Leon

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Marc Olzheim]

[-- Attachment #1.1: Type: text/plain, Size: 625 bytes --]

On Thu, Jul 30, 2009 at 03:22:19PM +0200, Leon Woestenberg wrote:
> > My /bin/sh is a dash, I should have known, my apologies.  I have a patch
> >
> Testing with dash rather than bash is GOOD due to it adhering to POSIX
> without much of the bash-specifics.

Well, I agree, but judging by commit
e0ed375a8dfb292bdd72f5891917ef81f361f15c to classes/sanity.bbclass, I
assumed the policy was to require /bin/bash for /bin/sh instead of
fixing bashisms.

To not be obligated to use bash, I added the attached patch.

> So please do not introduce 'bashisms'  (dependencies on bash behaviour).

I agree ;-)

Marc

[-- Attachment #1.2: dash.patch --]
[-- Type: text/x-diff, Size: 1039 bytes --]

commit d37e8ae3fb24b589793ae92062d665cc6384151e
Author: Marc Olzheim <marc@iphion.nl>
Date:   Thu Jun 11 12:44:34 2009 +0200

    Make allowing dash as /bin/sh a configuration option.

diff --git a/classes/sanity.bbclass b/classes/sanity.bbclass
index 29624c3..3ee9a0a 100644
--- a/classes/sanity.bbclass
+++ b/classes/sanity.bbclass
@@ -112,11 +112,12 @@ def check_sanity(e):
 		missing = missing.rstrip(',')
 		messages = messages + "Please install following missing utilities: %s\n" % missing
 
-	try:
-	    if os.path.basename(os.readlink('/bin/sh')) == 'dash':
-		    messages = messages + "Using dash as /bin/sh causes various subtle build problems, please use bash instead.\n"
-	except:
-		pass
+	if not data.getVar( 'ALLOW_DASH_SH', e.data, True ):
+		try:
+		    if os.path.basename(os.readlink('/bin/sh')) == 'dash':
+			    messages = messages + "Using dash as /bin/sh causes various subtle build problems, please use bash instead.\n"
+		except:
+			pass
 
 	omask = os.umask(022)
 	if omask & 0755:

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 204 bytes --]

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Philip Balister]

[-- Attachment #1: Type: text/plain, Size: 872 bytes --]

Marc Olzheim wrote:
> On Thu, Jul 30, 2009 at 03:22:19PM +0200, Leon Woestenberg wrote:
>>> My /bin/sh is a dash, I should have known, my apologies.  I have a patch
>>>
>> Testing with dash rather than bash is GOOD due to it adhering to POSIX
>> without much of the bash-specifics.
> 
> Well, I agree, but judging by commit
> e0ed375a8dfb292bdd72f5891917ef81f361f15c to classes/sanity.bbclass, I
> assumed the policy was to require /bin/bash for /bin/sh instead of
> fixing bashisms.
> 
> To not be obligated to use bash, I added the attached patch.
> 
>> So please do not introduce 'bashisms'  (dependencies on bash behaviour).
> 
> I agree ;-)

No sense us creating bash only scripts, but some of the scripts used to 
build things in OE (created by others) may have bashism's. Easier for us 
to use bash than patch other people's scripts.

Philip

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/x-pkcs7-signature, Size: 3303 bytes --]

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Graeme Gregory]

Marc Olzheim wrote:
> On Thu, Jul 30, 2009 at 03:22:19PM +0200, Leon Woestenberg wrote:
>   
>>> My /bin/sh is a dash, I should have known, my apologies.  I have a patch
>>>
>>>       
>> Testing with dash rather than bash is GOOD due to it adhering to POSIX
>> without much of the bash-specifics.
>>     
>
> Well, I agree, but judging by commit
> e0ed375a8dfb292bdd72f5891917ef81f361f15c to classes/sanity.bbclass, I
> assumed the policy was to require /bin/bash for /bin/sh instead of
> fixing bashisms.
>
> To not be obligated to use bash, I added the attached patch.
>
>   
OE is always supposed to run on dash, unfortuneately a lot of the stuff 
it builds is written by people less caring. I beleive that commit was 
spurred by a libtool failure on dash.

Graeme

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Marc Olzheim]

[-- Attachment #1: Type: text/plain, Size: 765 bytes --]

On Thu, Jul 30, 2009 at 10:57:40AM -0400, Philip Balister wrote:
> No sense us creating bash only scripts, but some of the scripts used to  
> build things in OE (created by others) may have bashism's. Easier for us  
> to use bash than patch other people's scripts.

Well, there's a difference between 'to use bash' and 'to require /bin/sh
to be a bash', the /bin/sh of the host system nonetheless.  As bash is
non-POSIX compliant in different ways than other shells are non-POSIX
compliant, some (non-OE) #!/bin/sh scripts of other packages break.

IMO it would make more sense to use bash as /bin/sh for that specific
package (if possible) but even more sense to fix the bashism in the
first place (and notify upstream to fix it as well)...

Marc

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 204 bytes --]

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Koen Kooi]

On 30-07-09 17:01, Graeme Gregory wrote:
> Marc Olzheim wrote:
>> On Thu, Jul 30, 2009 at 03:22:19PM +0200, Leon Woestenberg wrote:
>>>> My /bin/sh is a dash, I should have known, my apologies. I have a patch
>>>>
>>> Testing with dash rather than bash is GOOD due to it adhering to POSIX
>>> without much of the bash-specifics.
>>
>> Well, I agree, but judging by commit
>> e0ed375a8dfb292bdd72f5891917ef81f361f15c to classes/sanity.bbclass, I
>> assumed the policy was to require /bin/bash for /bin/sh instead of
>> fixing bashisms.
>>
>> To not be obligated to use bash, I added the attached patch.
>>
> OE is always supposed to run on dash, unfortuneately a lot of the stuff
> it builds is written by people less caring. I beleive that commit was
> spurred by a libtool failure on dash.

Gtk people abusing libtool to be more precise :)

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Mike Westerhof]

Stanislav Brabec wrote:
> Marc Olzheim wrote:
>> I'm not sure how you tested this, but there are quotes missing, so that
>> the "-E" argument to "grep" is passed to "export" instead. I guess some
>> shells have a "-E" flag to "export", so it is silently ignored, but mine
>> don't.
> 
> I tried both versions - quoted and unquoted - before the commit. Both
> worked for me, that is why I committed the shorter one.
> 
> Well, adding quotes is safe. Can you check your configure output and
> host shell version?
> 
> My configure output:
> 
> checking for a sed that does not truncate output... (cached) /bin/sed
> checking for grep that handles long lines and -e... (cached) /bin/sed
> checking for egrep... (cached) /bin/grep -E
> checking for fgrep... (cached) /bin/grep -F
> 
> It indicates, that it worked correctly.
> 
> My native shell is:
> 
> /OE/openembedded/build> sh --version
> GNU bash, version 4.0.24(1)-release (x86_64-suse-linux-gnu)
> 

The original commit seems to break libltdl.so as well (which was not
fixed by the subsequent patch to add quotes).  Libtool builds, but
certain dependencies fail.  Host platforms affected include Fedora and
Ubuntu, and neither uses dash.

Failing packages fail in do_configure, with a missing symbol in libltdl:
logs from the nslu2 autobuilder are up on tinderbox:

http://tinderbox.openembedded.net/public/logs/task/2207578.txt

Here's the message in config.log from that build:

configure:16081: checking for lt_dlinit in -lltdl
configure:16116: arm-linux-gnueabi-gcc -march=armv5te -mtune=xscale
-mno-thumb-interwork -mno-thumb -mno-thumb-interwork -mno-thumb -o
conftest
-isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
-fexpensive-optimizations -fomit-frame-pointer -frename-registers -Os
-Wall -Wmissing-declarations -Wmissing-prototypes
-isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
-I$(top_srcdir) -I$(top_builddir) -I$(top_srcdir)/libgphoto2_port
-I$(top_srcdir)/libgphoto2 -I$(top_builddir)/libgphoto2
-isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
-Wall -g -D_GPHOTO2_INTERNAL_CODE
-L/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib
-Wl,-rpath-link,/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib
-Wl,-O1 -Wl,--hash-style=gnu conftest.c -lltdl   >&5
/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib/libltdl.so:
undefined reference to `lt_libltdl_LTX_preloaded_symbols'
collect2: ld returned 1 exit status
configure:16122: $? = 1


Mike (mwester)

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Mike Westerhof]

[top-posting] Any thoughts, anyone?  Should I revert the original change
(and the follow-on)?  Or is this unique to SlugOS again, and I should
add an override to the change?  Or might anyone who knows libtool and
libltdl have any suggestions on where to look?  It would even be helpful
if anyone else can tell me if libgphoto2 and pulseaudio build correctly
for their distro!

Thanks.
Mike (mwester)


Mike Westerhof wrote:
> Stanislav Brabec wrote:
>> Marc Olzheim wrote:
>>> I'm not sure how you tested this, but there are quotes missing, so that
>>> the "-E" argument to "grep" is passed to "export" instead. I guess some
>>> shells have a "-E" flag to "export", so it is silently ignored, but mine
>>> don't.
>> I tried both versions - quoted and unquoted - before the commit. Both
>> worked for me, that is why I committed the shorter one.
>>
>> Well, adding quotes is safe. Can you check your configure output and
>> host shell version?
>>
>> My configure output:
>>
>> checking for a sed that does not truncate output... (cached) /bin/sed
>> checking for grep that handles long lines and -e... (cached) /bin/sed
>> checking for egrep... (cached) /bin/grep -E
>> checking for fgrep... (cached) /bin/grep -F
>>
>> It indicates, that it worked correctly.
>>
>> My native shell is:
>>
>> /OE/openembedded/build> sh --version
>> GNU bash, version 4.0.24(1)-release (x86_64-suse-linux-gnu)
>>
> 
> The original commit seems to break libltdl.so as well (which was not
> fixed by the subsequent patch to add quotes).  Libtool builds, but
> certain dependencies fail.  Host platforms affected include Fedora and
> Ubuntu, and neither uses dash.
> 
> Failing packages fail in do_configure, with a missing symbol in libltdl:
> logs from the nslu2 autobuilder are up on tinderbox:
> 
> http://tinderbox.openembedded.net/public/logs/task/2207578.txt
> 
> Here's the message in config.log from that build:
> 
> configure:16081: checking for lt_dlinit in -lltdl
> configure:16116: arm-linux-gnueabi-gcc -march=armv5te -mtune=xscale
> -mno-thumb-interwork -mno-thumb -mno-thumb-interwork -mno-thumb -o
> conftest
> -isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
> -fexpensive-optimizations -fomit-frame-pointer -frename-registers -Os
> -Wall -Wmissing-declarations -Wmissing-prototypes
> -isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
> -I$(top_srcdir) -I$(top_builddir) -I$(top_srcdir)/libgphoto2_port
> -I$(top_srcdir)/libgphoto2 -I$(top_builddir)/libgphoto2
> -isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
> -Wall -g -D_GPHOTO2_INTERNAL_CODE
> -L/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib
> -Wl,-rpath-link,/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib
> -Wl,-O1 -Wl,--hash-style=gnu conftest.c -lltdl   >&5
> /home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib/libltdl.so:
> undefined reference to `lt_libltdl_LTX_preloaded_symbols'
> collect2: ld returned 1 exit status
> configure:16122: $? = 1
> 
> 
> Mike (mwester)
> 
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
> 
> 

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Mike Westerhof]

Mike Westerhof wrote:
> [top-posting] Any thoughts, anyone?  Should I revert the original change
> (and the follow-on)?  Or is this unique to SlugOS again, and I should
> add an override to the change?  Or might anyone who knows libtool and
> libltdl have any suggestions on where to look?  It would even be helpful
> if anyone else can tell me if libgphoto2 and pulseaudio build correctly
> for their distro!

And I think the answer is no, it doesn't build correctly for anybody.
 The following line seems to be incorrect in the original patch:

+               export ac_cv_path_SED=${ac_cv_path_GREP=/bin/sed}
                                                   ^^^^

Mike (mwester)


> Thanks.
> Mike (mwester)
> 
> 
> Mike Westerhof wrote:
>> Stanislav Brabec wrote:
>>> Marc Olzheim wrote:
>>>> I'm not sure how you tested this, but there are quotes missing, so that
>>>> the "-E" argument to "grep" is passed to "export" instead. I guess some
>>>> shells have a "-E" flag to "export", so it is silently ignored, but mine
>>>> don't.
>>> I tried both versions - quoted and unquoted - before the commit. Both
>>> worked for me, that is why I committed the shorter one.
>>>
>>> Well, adding quotes is safe. Can you check your configure output and
>>> host shell version?
>>>
>>> My configure output:
>>>
>>> checking for a sed that does not truncate output... (cached) /bin/sed
>>> checking for grep that handles long lines and -e... (cached) /bin/sed
>>> checking for egrep... (cached) /bin/grep -E
>>> checking for fgrep... (cached) /bin/grep -F
>>>
>>> It indicates, that it worked correctly.
>>>
>>> My native shell is:
>>>
>>> /OE/openembedded/build> sh --version
>>> GNU bash, version 4.0.24(1)-release (x86_64-suse-linux-gnu)
>>>
>> The original commit seems to break libltdl.so as well (which was not
>> fixed by the subsequent patch to add quotes).  Libtool builds, but
>> certain dependencies fail.  Host platforms affected include Fedora and
>> Ubuntu, and neither uses dash.
>>
>> Failing packages fail in do_configure, with a missing symbol in libltdl:
>> logs from the nslu2 autobuilder are up on tinderbox:
>>
>> http://tinderbox.openembedded.net/public/logs/task/2207578.txt
>>
>> Here's the message in config.log from that build:
>>
>> configure:16081: checking for lt_dlinit in -lltdl
>> configure:16116: arm-linux-gnueabi-gcc -march=armv5te -mtune=xscale
>> -mno-thumb-interwork -mno-thumb -mno-thumb-interwork -mno-thumb -o
>> conftest
>> -isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
>> -fexpensive-optimizations -fomit-frame-pointer -frename-registers -Os
>> -Wall -Wmissing-declarations -Wmissing-prototypes
>> -isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
>> -I$(top_srcdir) -I$(top_builddir) -I$(top_srcdir)/libgphoto2_port
>> -I$(top_srcdir)/libgphoto2 -I$(top_builddir)/libgphoto2
>> -isystem/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/include
>> -Wall -g -D_GPHOTO2_INTERNAL_CODE
>> -L/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib
>> -Wl,-rpath-link,/home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib
>> -Wl,-O1 -Wl,--hash-style=gnu conftest.c -lltdl   >&5
>> /home/slug/slugos/tmp/staging/armv5te-linux-gnueabi/usr/lib/libltdl.so:
>> undefined reference to `lt_libltdl_LTX_preloaded_symbols'
>> collect2: ld returned 1 exit status
>> configure:16122: $? = 1
>>
>>
>> Mike (mwester)
>>
>> _______________________________________________
>> Openembedded-devel mailing list
>> Openembedded-devel@lists.openembedded.org
>> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>>
>>
> 
> 
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
> 
> 

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Koen Kooi]

On 02-08-09 03:32, Mike Westerhof wrote:
> [top-posting] Any thoughts, anyone?  Should I revert the original change
> (and the follow-on)?  Or is this unique to SlugOS again, and I should
> add an override to the change?  Or might anyone who knows libtool and
> libltdl have any suggestions on where to look?  It would even be helpful
> if anyone else can tell me if libgphoto2 and pulseaudio build correctly
> for their distro!

Sorry for taking so long, but no, the breakage is not unique to slugos, 
I see pulseaudio breaking with libtool as well.
I don't have time to look into it today, so if anyone wants to revert in 
the meantime, you have my ack.

regards,

Koen

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Koen Kooi]

On 02-08-09 05:56, Mike Westerhof wrote:
> Mike Westerhof wrote:
>> [top-posting] Any thoughts, anyone?  Should I revert the original change
>> (and the follow-on)?  Or is this unique to SlugOS again, and I should
>> add an override to the change?  Or might anyone who knows libtool and
>> libltdl have any suggestions on where to look?  It would even be helpful
>> if anyone else can tell me if libgphoto2 and pulseaudio build correctly
>> for their distro!
>
> And I think the answer is no, it doesn't build correctly for anybody.
>   The following line seems to be incorrect in the original patch:
>
> +               export ac_cv_path_SED=${ac_cv_path_GREP=/bin/sed}


Changing that fixes, will push a fix later today. Turns one to be one of 
the easier libtool problems to fix :)

regards,

Koen

Re: Commit 2e0fa4918d643ce6ec0299cd31cfe1c346e4bce3 breaks libtool's do_configure [PATCH]

[Marc Olzheim]

[-- Attachment #1.1: Type: text/plain, Size: 404 bytes --]

On Sat, Aug 01, 2009 at 10:56:48PM -0500, Mike Westerhof wrote:
> And I think the answer is no, it doesn't build correctly for anybody.
>  The following line seems to be incorrect in the original patch:
> 
> +               export ac_cv_path_SED=${ac_cv_path_GREP=/bin/sed}

Ah, doh, lol... How could we've overlooked this. :-/

Anyway, I'd say fix this for libtool_2.2.4.bb as well ;-)

Marc

[-- Attachment #1.2: lt224.patch --]
[-- Type: text/x-diff, Size: 864 bytes --]

commit 89c75a691b4c92aab6d29a85599824813813b205
Author: Marc Olzheim <marc@iphion.nl>
Date:   Mon Aug 3 10:23:10 2009 +0200

    libtool: Apply 0131d2ec1d1e49a1facbc02234d5a70f8b67425e on libtool 2.2.4 as well.

diff --git a/recipes/libtool/libtool_2.2.4.bb b/recipes/libtool/libtool_2.2.4.bb
index 5d59ce0..a36dd6e 100644
--- a/recipes/libtool/libtool_2.2.4.bb
+++ b/recipes/libtool/libtool_2.2.4.bb
@@ -17,7 +17,7 @@ EXTRA_AUTORECONF = "--exclude=libtoolize"
 do_configure_prepend () {
 	# Skip this for native build:
 	if test -n "$CONFIG_SITE" ; then
-		export ac_cv_path_SED="${ac_cv_path_GREP=/bin/sed}"
+		export ac_cv_path_SED="${ac_cv_path_SED=/bin/sed}"
 		export ac_cv_path_GREP="${ac_cv_path_GREP=/bin/grep}"
 		export ac_cv_path_EGREP="${ac_cv_path_EGREP=/bin/grep -E}"
 		export ac_cv_path_FGREP="${ac_cv_path_FGREP=/bin/grep -F}"

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 204 bytes --]