testing branch 2010-10-14

testing branch 2010-10-14

[Cliff Brake]

master has been merged to the testing-next branch and is now ready for
clean builds.  I'll finish tagging and updating the testing branch
later today.  It looks like enough things passed to justify moving the
testing branch to testing_2010-10-08.

Thanks,
Cliff

-- 
=================
http://bec-systems.com

Re: testing branch 2010-10-14

[Cliff Brake]

On Thu, Oct 14, 2010 at 5:56 PM, Cliff Brake <cliff.brake@gmail.com> wrote:
> master has been merged to the testing-next branch and is now ready for
> clean builds.  I'll finish tagging and updating the testing branch
> later today.  It looks like enough things passed to justify moving the
> testing branch to testing_2010-10-08.

tag created:
http://cgit.openembedded.org/cgit.cgi/openembedded/tag/?id=testing_2010-10-08

For more information on the testing effort, please see:
http://wiki.openembedded.org/index.php/Testing

The testing effort has turned up a number of issues, so I think it is
proving beneficial.  At some point we may want to consider making the
testing branch the default branch on checkout, or at least suggest
that new users start with it, and a known working build combination.

Thanks,
Cliff

-- 
=================
http://bec-systems.com

Re: testing branch 2010-10-14

[Petr Štetiar]

Cliff Brake <cliff.brake@gmail.com> [2010-10-14 17:56:24]:

> master has been merged to the testing-next branch and is now ready for
> clean builds.  I'll finish tagging and updating the testing branch
> later today.  It looks like enough things passed to justify moving the
> testing branch to testing_2010-10-08.

Hi,

ts72xx/angstrom-2010.x/minimal-image fails again on udev[1]. Adding that
CFLAGS_append = "-Os" fixed it, but I wonder if it's intention to use broken
recipe/compiler?  Wouldn't it be better to add this workaround into the recipe
and remove it once it's fixed in gcc? It's interesting, that the same
machine/distro combination builded fine with testing_2010-10-08.

1. http://tinderbox.openembedded.net/public/logs/task/8761618.txt

-- ynezz

Re: testing branch 2010-10-14

[Frans Meulenbroeks]

For me things in this branch work out quite well. 5 out of 6 targets
build; still peeking at #6.

But actually reason for mailing was to drop the suggestion from Crofton on irc.
He suggested testing_2010-xx-yy was confusing to some and that we
perhaps better could use tested_2010-xx-yy
I think he has a good point.

Frans

Re: testing branch 2010-10-14

[Cliff Brake]

On Tue, Oct 19, 2010 at 2:21 AM, Frans Meulenbroeks
<fransmeulenbroeks@gmail.com> wrote:
> For me things in this branch work out quite well. 5 out of 6 targets
> build; still peeking at #6.

I got a build failure with beagleboard/beagleboard-linuxtag2010-demo-image:

http://tinderbox.openembedded.net/builds/96943/

Koen, if you happen to read this, do you know if the following change
fixes this?

commit c9ac99f13f603d15e9a852916cc4daa4a4d210bc
Author: Koen Kooi <koen@openembedded.org>
Date:   Fri Oct 15 10:46:24 2010 +0200


                                                   python-pygobject:
forcefully disable  introspection

It will pick up the .pc from the native sysroot and then fail to build
since we don't have target introspection libs to link against

> But actually reason for mailing was to drop the suggestion from Crofton on irc.
> He suggested testing_2010-xx-yy was confusing to some and that we
> perhaps better could use tested_2010-xx-yy
> I think he has a good point.

Yes, I agree that makes sense.  We could even rename the "testing"
branch to "tested", which is the semi-stable branch that comes out of
this effort.

Thanks,
Cliff

-- 
=================
http://bec-systems.com

Re: testing branch 2010-10-14

[Koen Kooi]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 19-10-10 13:06, Cliff Brake wrote:
> On Tue, Oct 19, 2010 at 2:21 AM, Frans Meulenbroeks
> <fransmeulenbroeks@gmail.com> wrote:
>> For me things in this branch work out quite well. 5 out of 6 targets
>> build; still peeking at #6.
> 
> I got a build failure with beagleboard/beagleboard-linuxtag2010-demo-image:
> 
> http://tinderbox.openembedded.net/builds/96943/
> 
> Koen, if you happen to read this, do you know if the following change
> fixes this?

It does for me :)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Darwin)

iD8DBQFMvX1DMkyGM64RGpERAruPAJ9gfoW/81Qv9YxykT6bUq5txq+CRQCeMRSR
DS03gt7oo2ZYunYN1y3mtYI=
=TCHC
-----END PGP SIGNATURE-----

Re: testing branch 2010-10-14 - configure epiphany still fails

[Sledz, Steffen]

do_configure still fails with the ca-certificates problem [1]. :(

If nobody will find a solution for this problem (for my first diagnosis see [2]) until the next testing branch creation, i'll push my original patch which disables the ca-file use [3].

Steffen

[1] <http://tinderbox.openembedded.org/packages/859111/>
[2] <http://thread.gmane.org/gmane.comp.handhelds.openembedded/37925/focus=38034>
[3] <http://thread.gmane.org/gmane.comp.handhelds.openembedded/37453/focus=37585>

Re: testing branch 2010-10-14 - configure epiphany still fails

[Steffen Sledz]

On 10/17/2010 02:27 PM, Sledz, Steffen wrote:
> do_configure still fails with the ca-certificates problem [1]. :(
> 
> If nobody will find a solution for this problem (for my first diagnosis see [2]) until the next testing branch creation, i'll push my original patch which disables the ca-file use [3].
> 
> Steffen
> 
> [1] <http://tinderbox.openembedded.org/packages/859111/>
> [2] <http://thread.gmane.org/gmane.comp.handhelds.openembedded/37925/focus=38034>
> [3] <http://thread.gmane.org/gmane.comp.handhelds.openembedded/37453/focus=37585>

OK, i tried to fix for myself. Patches follow.

Steffen

[PATCH 1/2] ca-certificates: unbreak do_configure of packages depending on ca-bundle.pem

[Steffen Sledz]

just create an empty ca-bundle file to satisfy build time
dependencies in do_configure of other packages (e.g. epiphany),
content is updated later in postinst

Signed-off-by: Steffen Sledz <sledz@dresearch.de>
---
 .../ca-certificates_20090814+nmu2.bb               |    9 +++++++++
 1 files changed, 9 insertions(+), 0 deletions(-)

diff --git a/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb b/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb
index c8e4dbd..a37b01a 100644
--- a/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb
+++ b/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb
@@ -3,6 +3,7 @@ HOMEPAGE = "http://packages.debian.org/sid/ca-certificates"
 SECTION = "misc"
 PRIORITY = "optional"
 LICENSE = "GPL"
+PR = "r1"
 
 SRC_URI = "${DEBIAN_MIRROR}/main/c/ca-certificates/ca-certificates_${PV}.tar.gz \
            file://remove-c-rehash.patch"
@@ -27,13 +28,21 @@ do_install_append() {
                 crt=$(echo $crt | sed -e 's/\.\///')
                 echo $crt >> ${D}/etc/ca-certificates.conf
         done
+
+	# update-ca-certificates is not able to deal with pathes,
+	# so just create an empty ca-bundle file to satisfy build time
+	# dependencies in do_configure of other packages (e.g. epiphany),
+	# content is updated later (see pkg_postinst)
+	touch ${D}/etc/ssl/ca-bundle.pem
 }
 
 PACKAGE_ARCH = "all"
 PACKAGES = "${PN}"
 
 pkg_postinst_${PN} () {
+if test -z "$D"; then
         /usr/sbin/update-ca-certificates
+fi
 }
 
 CONFFILES_${PN} = "/etc/ca-certificates.conf"
-- 
1.7.1

Re: [PATCH 1/2] ca-certificates: unbreak do_configure of packages depending on ca-bundle.pem

[Thomas Zimmermann]

Am Mittwoch 20 Oktober 2010, 13:10:06 schrieb Steffen Sledz:
> just create an empty ca-bundle file to satisfy build time
> dependencies in do_configure of other packages (e.g. epiphany),
> content is updated later in postinst
> 
> Signed-off-by: Steffen Sledz <sledz@dresearch.de>

Acked-by: Thomas Zimmermann <ml@vdm-design.de>

> ---
>  .../ca-certificates_20090814+nmu2.bb               |    9 +++++++++
>  1 files changed, 9 insertions(+), 0 deletions(-)
> 
> diff --git a/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb
> b/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb index
> c8e4dbd..a37b01a 100644
> --- a/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb
> +++ b/recipes/ca-certificates/ca-certificates_20090814+nmu2.bb
> @@ -3,6 +3,7 @@ HOMEPAGE = "http://packages.debian.org/sid/ca-certificates"
>  SECTION = "misc"
>  PRIORITY = "optional"
>  LICENSE = "GPL"
> +PR = "r1"
> 
>  SRC_URI =
> "${DEBIAN_MIRROR}/main/c/ca-certificates/ca-certificates_${PV}.tar.gz \
> file://remove-c-rehash.patch"
> @@ -27,13 +28,21 @@ do_install_append() {
>                  crt=$(echo $crt | sed -e 's/\.\///')
>                  echo $crt >> ${D}/etc/ca-certificates.conf
>          done
> +
> +	# update-ca-certificates is not able to deal with pathes,
> +	# so just create an empty ca-bundle file to satisfy build time
> +	# dependencies in do_configure of other packages (e.g. epiphany),
> +	# content is updated later (see pkg_postinst)
> +	touch ${D}/etc/ssl/ca-bundle.pem
>  }
> 
>  PACKAGE_ARCH = "all"
>  PACKAGES = "${PN}"
> 
>  pkg_postinst_${PN} () {
> +if test -z "$D"; then
>          /usr/sbin/update-ca-certificates
> +fi
>  }
> 
>  CONFFILES_${PN} = "/etc/ca-certificates.conf"

[PATCH 2/2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Steffen Sledz]

Signed-off-by: Steffen Sledz <sledz@dresearch.de>
---
 recipes/gnome/epiphany_2.30.2.bb |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/recipes/gnome/epiphany_2.30.2.bb b/recipes/gnome/epiphany_2.30.2.bb
index 57a5d9a..bf796d8 100644
--- a/recipes/gnome/epiphany_2.30.2.bb
+++ b/recipes/gnome/epiphany_2.30.2.bb
@@ -3,7 +3,7 @@ LICENSE = "GPLv2+"
 DEPENDS = "libsoup-2.4 gnome-desktop gnome-vfs libgnomeui webkit-gtk iso-codes startup-notification ca-certificates"
 RDEPENDS_${PN} = "gnome-vfs-plugin-http iso-codes"
 RRECOMMENDS_${PN} = "ca-certificates"
-PR = "r4"
+PR = "r5"
 
 inherit gnome
 
@@ -11,7 +11,7 @@ SRC_URI[archive.md5sum] = "29b084acfa016540d91d3601ec3dff5c"
 SRC_URI[archive.sha256sum] = "cd0124e71e72142593cfeb442d58d97e99ba94ace6e31d94717fe977c0bfb98a"
 
 
-EXTRA_OECONF += "--disable-nss --with-engine=webkit --with-distributor-name=${DISTRO}"
+EXTRA_OECONF += "--disable-nss --with-engine=webkit --with-distributor-name=${DISTRO} --with-ca-file=${STAGING_DIR_HOST}/etc/ssl/ca-bundle.pem"
 
 do_configure_prepend() {
         touch ${S}/gnome-doc-utils.make
-- 
1.7.1

Re: [PATCH 2/2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Steffen Sledz]

Some ACKs here before testing-next is updated?

Steffen

Re: [PATCH 2/2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Holger Freyther]

On 10/21/2010 11:14 AM, Steffen Sledz wrote:
> Some ACKs here before testing-next is updated?

It doesn't look correct. Judging from the code epiphany will now look into
${STAGING_DIR_HOST}/etc/ssl/ca-bundle.pem at runtime. What about stripping
away this and use the etcdir variable?

Re: [PATCH 2/2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Steffen Sledz]

Am 21.10.2010 11:54, schrieb Holger Freyther:
> On 10/21/2010 11:14 AM, Steffen Sledz wrote:
>> Some ACKs here before testing-next is updated?
> 
> It doesn't look correct. Judging from the code epiphany
> will now look into ${STAGING_DIR_HOST}/etc/ssl/ca-bundle.pem
> at runtime.

OK, you're right.

> What about stripping away this and use the etcdir variable?

Sorry, no idea how to do this. Do you have an example?

Steffen

-- 
Steffen Sledz
DResearch Digital Media Systems GmbH
Otto-Schmirgal-Str.3, D-10319 Berlin, Germany
Tel: +49 (30) 515932237 mailto:sledz@DResearch.DE
Fax: +49 (30) 515932299 http://www.DResearch.DE
Geschäftsführer: Dr. Michael Weber, Werner Mögle;
Amtsgericht Berlin Charlottenburg; HRB 54412;
Ust.-IDNr. DE169013825; WEEE Reg.-Nr. DE 85995642

Re: [PATCH 2/2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Sledz, Steffen]

Am 21.10.2010 14:11, schrieb Steffen Sledz:
> Am 21.10.2010 11:54, schrieb Holger Freyther:
>> It doesn't look correct. Judging from the code epiphany
>> will now look into ${STAGING_DIR_HOST}/etc/ssl/ca-bundle.pem
>> at runtime.
> 
> OK, you're right.
> 
>> What about stripping away this and use the etcdir variable?
> 
> Sorry, no idea how to do this. Do you have an example?

I've done some further investigations here.

The problem is that the configure script checks for the ca-file (tries some default locations, or uses the given path) *at build time at the build host* und uses the resulting path *at runtime at the target*.
I believe this is an error in the epiphany package itself.

I see two options here (i prefer the 1st):

1. Use the --with-ca-file option but disable the existance check in configure.ac

2. Use --without-ca-file.

Both would obsolete the related patch i suggested for ca-certificates recipe.

Any other options?

BTW: Why does this error comes up in my builds and not in others?
I think the cause are the default locations configure tries. These are /etc/pki/tls/certs/ca-bundle.crt and /etc/ssl/certs/ca-certificates.crt.
I use openSUSE-11.3 as build host which puts its own ca bundle at /etc/ssl/ca-bundle.pem. May be Fedore or Ubuntu use one of the locations mentioned above.

Steffen

[PATCH v2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Steffen Sledz]

* Disable existence check for ca-bundle file in configure.ac because
  it does not work in OE cross compile environment (location at build
  host may differ from location at target).
* Explicitely set ca-bundle file path according to the one used by
  ca-certificates package.

Signed-off-by: Steffen Sledz <sledz@dresearch.de>
---
 recipes/gnome/epiphany/ca-file-configure.patch |   17 +++++++++++++++++
 recipes/gnome/epiphany_2.30.2.bb               |    6 ++++--
 2 files changed, 21 insertions(+), 2 deletions(-)
 create mode 100644 recipes/gnome/epiphany/ca-file-configure.patch

diff --git a/recipes/gnome/epiphany/ca-file-configure.patch b/recipes/gnome/epiphany/ca-file-configure.patch
new file mode 100644
index 0000000..b77d441
--- /dev/null
+++ b/recipes/gnome/epiphany/ca-file-configure.patch
@@ -0,0 +1,17 @@
+--- epiphany-2.30.2.orig/configure.ac	2010-10-25 11:10:31.316864572 +0200
++++ epiphany-2.30.2/configure.ac	2010-10-25 11:12:29.562614607 +0200
+@@ -272,9 +272,11 @@
+     fi
+ 
+     AC_MSG_RESULT($with_ca_file)
+-    if ! test -f "$with_ca_file"; then
+-        AC_MSG_ERROR([No such file '$with_ca_file'. Use --with-ca-file=path to set, or --without-ca-file to disable])
+-    fi
++# disable this check because it does not work in all environments
++# (location at build host may differ from target)
++#    if ! test -f "$with_ca_file"; then
++#        AC_MSG_ERROR([No such file '$with_ca_file'. Use --with-ca-file=path to set, or --without-ca-file to disable])
++#    fi
+     GTLS_SYSTEM_CA_FILE="$with_ca_file"
+ 
+     AC_DEFINE_UNQUOTED([GTLS_SYSTEM_CA_FILE], ["$GTLS_SYSTEM_CA_FILE"], [path to system Certificate Authority list])
diff --git a/recipes/gnome/epiphany_2.30.2.bb b/recipes/gnome/epiphany_2.30.2.bb
index 57a5d9a..5a0a5eb 100644
--- a/recipes/gnome/epiphany_2.30.2.bb
+++ b/recipes/gnome/epiphany_2.30.2.bb
@@ -3,15 +3,17 @@ LICENSE = "GPLv2+"
 DEPENDS = "libsoup-2.4 gnome-desktop gnome-vfs libgnomeui webkit-gtk iso-codes startup-notification ca-certificates"
 RDEPENDS_${PN} = "gnome-vfs-plugin-http iso-codes"
 RRECOMMENDS_${PN} = "ca-certificates"
-PR = "r4"
+PR = "r5"
 
 inherit gnome
 
+SRC_URI += " file://ca-file-configure.patch "
+
 SRC_URI[archive.md5sum] = "29b084acfa016540d91d3601ec3dff5c"
 SRC_URI[archive.sha256sum] = "cd0124e71e72142593cfeb442d58d97e99ba94ace6e31d94717fe977c0bfb98a"
 
 
-EXTRA_OECONF += "--disable-nss --with-engine=webkit --with-distributor-name=${DISTRO}"
+EXTRA_OECONF += "--disable-nss --with-engine=webkit --with-distributor-name=${DISTRO} --with-ca-file=/etc/ssl/certs/ca-certificates.crt"
 
 do_configure_prepend() {
         touch ${S}/gnome-doc-utils.make
-- 
1.7.1

Re: [PATCH v2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Steffen Sledz]

Am 25.10.2010 11:55, schrieb Steffen Sledz:
> * Disable existence check for ca-bundle file in configure.ac because
>   it does not work in OE cross compile environment (location at build
>   host may differ from location at target).
> * Explicitely set ca-bundle file path according to the one used by
>   ca-certificates package.
> 
> Signed-off-by: Steffen Sledz <sledz@dresearch.de>

This is a patch implementing the first option i mentioned (Use the --with-ca-file option but disable the existance check in configure.ac).

If there are no better solutions in the next time i'll push this one.

Steffen

Re: [PATCH v2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Paul Menzel]

[-- Attachment #1: Type: text/plain, Size: 2081 bytes --]

Am Montag, den 25.10.2010, 11:55 +0200 schrieb Steffen Sledz:
> * Disable existence check for ca-bundle file in configure.ac because
>   it does not work in OE cross compile environment (location at build
>   host may differ from location at target).
> * Explicitely set ca-bundle file path according to the one used by
>   ca-certificates package.
> 
> Signed-off-by: Steffen Sledz <sledz@dresearch.de>
> ---
>  recipes/gnome/epiphany/ca-file-configure.patch |   17 +++++++++++++++++
>  recipes/gnome/epiphany_2.30.2.bb               |    6 ++++--
>  2 files changed, 21 insertions(+), 2 deletions(-)
>  create mode 100644 recipes/gnome/epiphany/ca-file-configure.patch
> 
> diff --git a/recipes/gnome/epiphany/ca-file-configure.patch b/recipes/gnome/epiphany/ca-file-configure.patch
> new file mode 100644
> index 0000000..b77d441
> --- /dev/null
> +++ b/recipes/gnome/epiphany/ca-file-configure.patch
> @@ -0,0 +1,17 @@
> +--- epiphany-2.30.2.orig/configure.ac	2010-10-25 11:10:31.316864572 +0200
> ++++ epiphany-2.30.2/configure.ac	2010-10-25 11:12:29.562614607 +0200
> +@@ -272,9 +272,11 @@
> +     fi
> + 
> +     AC_MSG_RESULT($with_ca_file)
> +-    if ! test -f "$with_ca_file"; then
> +-        AC_MSG_ERROR([No such file '$with_ca_file'. Use --with-ca-file=path to set, or --without-ca-file to disable])
> +-    fi
> ++# disable this check because it does not work in all environments
> ++# (location at build host may differ from target)
> ++#    if ! test -f "$with_ca_file"; then
> ++#        AC_MSG_ERROR([No such file '$with_ca_file'. Use --with-ca-file=path to set, or --without-ca-file to disable])
> ++#    fi
> +     GTLS_SYSTEM_CA_FILE="$with_ca_file"
> + 
> +     AC_DEFINE_UNQUOTED([GTLS_SYSTEM_CA_FILE], ["$GTLS_SYSTEM_CA_FILE"], [path to system Certificate Authority list])

Would adding the following check also work?

        if test "$cross_compiling" != "yes"; then
         …
        fi

I think this would be more portable. Could you report that issue
upstream too please.

[…]


Thanks,

Paul

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 205 bytes --]

Re: [PATCH v2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Steffen Sledz]

Am 25.10.2010 13:37, schrieb Paul Menzel:
> Am Montag, den 25.10.2010, 11:55 +0200 schrieb Steffen Sledz:
>> * Disable existence check for ca-bundle file in configure.ac because
>>   it does not work in OE cross compile environment (location at build
>>   host may differ from location at target).
>> * Explicitely set ca-bundle file path according to the one used by
>>   ca-certificates package.
>> ...
>> ++# disable this check because it does not work in all environments
>> ++# (location at build host may differ from target)
>> ++#    if ! test -f "$with_ca_file"; then
>> ++#        AC_MSG_ERROR([No such file '$with_ca_file'. Use --with-ca-file=path to set, or --without-ca-file to disable])
>> ++#    fi
> 
> Would adding the following check also work?
> 
>         if test "$cross_compiling" != "yes"; then
>          …
>         fi

I don't think this is a good idea. It is not a problem of cross compiling. Cross compiling is just one case where it occurs.

Also if i build for the same architecture it is not a good idea to check for the existence of a file *at buildtime* and compile this path into the binary.

The existence of such a file should be checked and handled at runtime only.

> I think this would be more portable. Could you report that issue
> upstream too please.

I'll try.

Steffen

Re: [PATCH v2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Steffen Sledz]

Am 25.10.2010 14:07, schrieb Steffen Sledz:
> Am 25.10.2010 13:37, schrieb Paul Menzel:
>> ... Could you report that issue
>> upstream too please.
> 
> I'll try.

<https://bugzilla.gnome.org/show_bug.cgi?id=633109>

Re: [PATCH v2] epiphany-2.30.2: unbreak do_configure by setting ca-file path

[Khem Raj]

On 10/25/10, Steffen Sledz <sledz@dresearch.de> wrote:
> Am 25.10.2010 13:37, schrieb Paul Menzel:
>> Am Montag, den 25.10.2010, 11:55 +0200 schrieb Steffen Sledz:
>>> * Disable existence check for ca-bundle file in configure.ac because
>>>   it does not work in OE cross compile environment (location at build
>>>   host may differ from location at target).
>>> * Explicitely set ca-bundle file path according to the one used by
>>>   ca-certificates package.
>>> ...
>>> ++# disable this check because it does not work in all environments
>>> ++# (location at build host may differ from target)
>>> ++#    if ! test -f "$with_ca_file"; then
>>> ++#        AC_MSG_ERROR([No such file '$with_ca_file'. Use
>>> --with-ca-file=path to set, or --without-ca-file to disable])
>>> ++#    fi
>>
>> Would adding the following check also work?
>>
>>         if test "$cross_compiling" != "yes"; then
>>          …
>>         fi
>
> I don't think this is a good idea. It is not a problem of cross compiling.
> Cross compiling is just one case where it occurs.
>
> Also if i build for the same architecture it is not a good idea to check for
> the existence of a file *at buildtime* and compile this path into the
> binary.
>
> The existence of such a file should be checked and handled at runtime only.

For native build runtime and buildtimes are same even if you
are building in a chroot so its an ok assumption
adding a cross compile check is a good thing.

It works on debian derivatives because the paths match coincidently
so for target package it uses the ca files installed on build system
but for SuSE location is different so it fails.

You have my ack for the patch

Acked-by: Khem Raj <raj.khem@gmail.com>

>
>> I think this would be more portable. Could you report that issue
>> upstream too please.
>
> I'll try.
>
> Steffen
>
>
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>


-- 
-Khem