Libguestfs - Feb 2010 - Debian libguestfs packages updated

Hi,
I finally updated the Debian libguestfs packages:

http://honk.sigxcpu.org/con/Libguestfs__Virtual_Machine_Image_Swiss_Army_Knife.html

Sine we can't run the fuse tests easily in a chroot (since we might be
unable to load the module) I'm using the attached patch to skip the
tests if /dev/fuse doesn't exist. This probably isn't clean enough to go
upstream though.

The other patch adjusts the test output of rhbz557655. xstrtol seems to
be behaving differently here. I didn't investigate why yet.

As a last point: is there a simple way to disable the appliance build
altogether? I'd like to ship the Debian package without it so we stay
out of the security update trouble and provide them via http.

Cheers and thanks for that great tool,
 -- Guido

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0003-Adjust-error-output-for-regressions-rhbz557655.patch
Type: text/x-diff
Size: 1198 bytes
Desc: not available
URL:
<http://listman.redhat.com/archives/libguestfs/attachments/20100222/4a30c2a2/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0004-Disable-fuser-tests-if-dev-fuse-doesn-t-exist.patch
Type: text/x-diff
Size: 678 bytes
Desc: not available
URL:
<http://listman.redhat.com/archives/libguestfs/attachments/20100222/4a30c2a2/attachment-0001.bin>

On Mon, Feb 22, 2010 at 02:41:22PM +0100, Guido G?nther
wrote:
> Hi,
> I finally updated the Debian libguestfs packages:
> 
>
http://honk.sigxcpu.org/con/Libguestfs__Virtual_Machine_Image_Swiss_Army_Knife.html
> 
> Sine we can't run the fuse tests easily in a chroot (since we might be
> unable to load the module) I'm using the attached patch to skip the
> tests if /dev/fuse doesn't exist. This probably isn't clean enough
to go
> upstream though.
Funnily enough, we [Fedora] cannot run the FUSE tests either, because
the people who run our build machines refuse to install the fuse.ko
module.  For this reason we carry the following non-upstream patch:

http://cvs.fedoraproject.org/viewvc/devel/libguestfs/libguestfs-1.0.79-no-fuse-test.patch?view=markup

I think your patch is neater, but I don't want to apply it because I'm
not happy about skipping tests magically (admittedly we do skip tests
magically elsewhere).

Probably the answer is ./configure --disable-fuse-tests.
> The other patch adjusts the test output of rhbz557655. xstrtol seems to
> be behaving differently here. I didn't investigate why yet.
[...]
> -set-memsize: memsize: integer out of range
> -set-memsize: memsize: integer out of range
> -set-memsize: memsize: integer out of range
> -set-memsize: memsize: integer out of range
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
That's unexpected, and needs further investigation.

I opened a bug about it to investigate this further:

https://bugzilla.redhat.com/show_bug.cgi?id=567567
> As a last point: is there a simple way to disable the appliance build
> altogether? I'd like to ship the Debian package without it so we stay
> out of the security update trouble and provide them via http.
There's not a way to disable it at the moment, but I would
accept a patch which added ./configure --disable-appliance.

Have you thought about distributing the supermin appliance?  That
should have no security update implications, and it's what we do in
Fedora, precisely for this reason, also because it's much smaller.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
virt-p2v converts physical machines to virtual machines.  Boot with a
live CD or over the network (PXE) and turn machines into Xen guests.
http://et.redhat.com/~rjones/virt-p2v

On Mon, Feb 22, 2010 at 02:41:22PM +0100, Guido G?nther
wrote:
> The other patch adjusts the test output of rhbz557655. xstrtol seems to
> be behaving differently here. I didn't investigate why yet.
There's a fix for this upstream now:

https://bugzilla.redhat.com/show_bug.cgi?id=567567
> to make the testsuite pass.
> ---
>  regressions/rhbz557655-expected.stderr |    8 ++++----
>  1 files changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/regressions/rhbz557655-expected.stderr
b/regressions/rhbz557655-expected.stderr
> index ea560e8..bc62baa 100644
> --- a/regressions/rhbz557655-expected.stderr
> +++ b/regressions/rhbz557655-expected.stderr
> @@ -1,7 +1,7 @@
> -set-memsize: memsize: integer out of range
> -set-memsize: memsize: integer out of range
> -set-memsize: memsize: integer out of range
> -set-memsize: memsize: integer out of range
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
> +set-memsize: memsize: invalid integer parameter (xstrtol returned 1)
>  set-memsize: memsize: invalid integer parameter (xstrtol returned 4)
>  set-memsize: memsize: invalid integer parameter (xstrtol returned 2)
>  set-memsize: memsize: invalid integer parameter (xstrtol returned 2)
which hopefully means you should be able to drop this patch
from Debian.

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
virt-top is 'top' for virtual machines.  Tiny program with many
powerful monitoring features, net stats, disk stats, logging, etc.
http://et.redhat.com/~rjones/virt-top