Re: OLPC build creation failed

2012-10-04 Thread Jose Prous
Thanks Martin, I could build bios-crypto from git. It signs and it looks
like everything works :)

Thanks everybody for the help.

On Tue, Oct 2, 2012 at 12:57 PM, Jerry Vonau jvo...@shaw.ca wrote:

 On Thu, 2012-09-13 at 17:51 -0700, John Gilmore wrote:
   On Thu, Sep 13, 2012 at 3:57 PM, Jose Prous josepr...@gmail.com
 wrote:
Yes it's a x86 machine, I guess that is the problem. Thanks.
  
   Glad that we found the reason. We should add an explicit check in OOB
   that gives you a more useful error msg.
 
  Instead, you should fix OOB so it works to cross-compile.

 Sorry John, I don't think OOB is fixable without first fixing yum as
 imgcreater uses yum to resolve dependencies and install the rpms. The
 problem is that yum deduces what arch yum is running on by calling uname
 and offers no way to override that variable[1]. The PTB didn't view this
 behaviour as a bug, which later cropped up in other places in Fedora
 like anaconda[2]. Even Martin felt the effects of this decision[3].

 Jerry

 1.https://bugzilla.redhat.com/show_bug.cgi?id=172080
 2.https://bugzilla.redhat.com/show_bug.cgi?id=367731
 3.
 http://www.redhat.com/archives/fedora-buildsys-list/2009-February/msg00028.html



  We at
  Cygnus spent years making the GNU tools able to cross-compile, and it
  all works.  We built a big test infrastructure and made sure that the
  resulting executables were bit-for-bit identical no matter what the
  build host was.  It's just your builder script that's unable to
  do it -- and that is under your control.
 



 ___
 Devel mailing list
 Devel@lists.laptop.org
 http://lists.laptop.org/listinfo/devel

___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] [DESIGN] Proposal: Contol-Panel packaging

2012-10-04 Thread Peter Robinson
On Fri, Jul 27, 2012 at 2:51 AM, Jerry Vonau jvo...@shaw.ca wrote:
 Hi All:

 I would like to propose a feature for discussion and inclusion in the
 0.98 cycle is packaging all control-panel applets as rpms. As this
 discussion does not impact the UI and more of a packaging issue I'm an
 not creating a Features page. The discussion can take place here on the
 mailing-list.

 The reasoning is that it should make it easier for downstream users of
 sugar to exclude applets that don't apply to their use case without
 having to patch them out. For example OLPC removes via their .spec file,
 the keyboard and updater applets from their builds and uses their own
 version of an updater supplied as an rpm.

 Dextrose is using this idea to now to partly split up the what is
 available to install at rpm generation time[1]. I have found this useful
 from a deployment perspective by being able to exclude applets that are
 unwanted or need further development from the final image.

 The current code base and workflow would not be changed, except a
 revised sugar.spec file would generate more that just the sugar rpm when
 run, like how it is done now for sugar-emulator. Deployment level users
 of sugar would then need to state which applets to include in their
 image at image creation time. This will allow development of applets to
 evolve without having to reinstall all of sugar in the field for a
 change to an applet.

 Any XO specific user tool like About my Computer and Power should
 not really be part of sugar but should be available to install on demand
 like OLPC's sugar-update-control and olpc-switch-desktop that are added
 to OLPC's sugar installation. SoaS might benefit from not shipping all
 the applets, omitting the ones that apply to XO hardware.

 This change might help development of new features in the control-panel
 area that later be incorporated into sugar once proven to work.

 Feedback and comments welcome,

I've pushed this and it will appear in the next build, feedback welcome.

I've left the AboutComputer and AboutMe built in and all the rest are
sub packages. There's a meta-package called sugar-cp-all which pulls
all the Control Panels packages in as well.

Sorry about the delay.

Peter
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] [DESIGN] Proposal: Contol-Panel packaging

2012-10-04 Thread Jerry Vonau
On Thu, 2012-10-04 at 21:29 +0100, Peter Robinson wrote:
 On Fri, Jul 27, 2012 at 2:51 AM, Jerry Vonau jvo...@shaw.ca wrote:
  Hi All:
 
  I would like to propose a feature for discussion and inclusion in the
  0.98 cycle is packaging all control-panel applets as rpms. As this
  discussion does not impact the UI and more of a packaging issue I'm an
  not creating a Features page. The discussion can take place here on the
  mailing-list.
snip
  Feedback and comments welcome,
 
 I've pushed this and it will appear in the next build, feedback welcome.
 
 I've left the AboutComputer and AboutMe built in and all the rest are
 sub packages. There's a meta-package called sugar-cp-all which pulls
 all the Control Panels packages in as well.
 
 Sorry about the delay.

Thank you very much, reviewing now.

Jerry


___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Linux 3.7 Kernel To Support Multiple ARM Platforms - Slashdot

2012-10-04 Thread Sameer Verma
Up until now there has been a separate Linux kernel build for each of
the ARM platforms or SoCs, which is one of the several problems when
it comes to ARM based Linux. The merging of ARM multi-platform support
into Linux 3.7 will put an end to this problem, enabling the new
kernel to not only target multiple platforms but also be more in line
with its x86 counterpart.

http://linux.slashdot.org/story/12/10/04/1942201/linux-37-kernel-to-support-multiple-arm-platforms
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


qemu wine experiments on ARM

2012-10-04 Thread Martin Langhoff
Just a quick outline of one experiment to see if simple Win32
content-ware apps can be run. Commands may be missing switches,
options and steps. Caveat reader. BYO smarts.

On x87 host F17
 - install livecd-tools spin-kickstarts
 - copy fedora-live-mini.ks, add wine -- remove some stuff
 - sudo setarch i686 livecd-creator fedora-live-mini.ks
 - wait
 - copy resulting iso to ext disk
 - yumdownloader sgabios-bin-0-0.20110622SVN.fc17.noarch
seabios-bin-1.7.0-1.fc17.noarch -- they are noarch but missing from
arm repos. copy to ext disk

On ARM XO F17
 - install seabios-bin sgabios-bin rpms
 - yum -ty install qemu-system-x86 qemu-user
 - qemu-386 -m 512 -cdrom /some/media/fedora-mini.iso
   inside of the qemu window, in the live cd boot text menu press tab
to remove rhbg and quiet. boot slowly. Probably. Eventually.

Clearly, qemu can emulate x86 on ARM. Slowly. This is of course
impractical, mainly because you need to get through a bootprocess
before you can run an app and the app is not in a transparently shared
environment (it's in qemu's odd fb window).

Qemu's user-mode is a lot more practical. We would need a minimal
fedora x86 chroot that has wine, and use qemu over that. After a bit
of googling, I found PRoot, which seems to provide the right glue to
make qemu's user mode actually usable.

http://cedric-vincent.github.com/PRoot/
http://adt.cs.upb.de/quf/quf11/quf2011_13.pdf

For a truly minimal rootfs, mock is likely to be more useful than livecd-tools.



m
-- 
 mar...@laptop.org -- Software Architect - OLPC
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: qemu wine experiments on ARM

2012-10-04 Thread Martin Langhoff
On Fri, Oct 5, 2012 at 12:22 AM, Martin Langhoff mar...@laptop.org wrote:
 Qemu's user-mode is a lot more practical. We would need a minimal
 fedora x86 chroot that has wine, and use qemu over that. After a bit
 of googling, I found PRoot, which seems to provide the right glue to
 make qemu's user mode actually usable.

 http://cedric-vincent.github.com/PRoot/
 http://adt.cs.upb.de/quf/quf11/quf2011_13.pdf

proot only offers x86 rpms, so a rebuild for armv7hl/f17 at
http://dev.laptop.org/~martin/proot/ - untested

 For a truly minimal rootfs, mock is likely to be more useful than 
 livecd-tools.

On x86 fedora...

mock -r fedora-17-i386 --init
mock -r fedora-17-i386 --install wine

then tar up /var/lib/mock/fedora-17-i386/root, untar on your XO.


m
-- 
 mar...@laptop.org -- Software Architect - OLPC
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: Linux 3.7 Kernel To Support Multiple ARM Platforms - Slashdot

2012-10-04 Thread Peter Robinson
On Fri, Oct 5, 2012 at 1:17 AM, Sameer Verma sve...@sfsu.edu wrote:
 Up until now there has been a separate Linux kernel build for each of
 the ARM platforms or SoCs, which is one of the several problems when
 it comes to ARM based Linux. The merging of ARM multi-platform support
 into Linux 3.7 will put an end to this problem, enabling the new
 kernel to not only target multiple platforms but also be more in line
 with its x86 counterpart.

 http://linux.slashdot.org/story/12/10/04/1942201/linux-37-kernel-to-support-multiple-arm-platforms

It supports basically 3 usable platforms at the moment, and likely not
all components of them. It's a marker point in what began just prior
to 3.0 with the ARM rant from Linus. There's still along way to go.
You should see the first Fedora kernel supporting this in the coming
weeks on 3 server platforms.

Peter
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel