Re: [PATCH] tests/vm/haiku.x86_64: Update the Haiku VM to Beta 4

[Philippe Mathieu-Daudé]

On 16/1/23 09:30, Thomas Huth wrote:
> The old Haiku VM based on Beta 3 does not work anymore since it
> fails to install the additional packages now that Beta 4 has been
> released. Thanks to Alexander von Gluck IV for providing a new
> image based on Beta 4, we can now upgrade the test image in our
> QEMU CI, too, to get this working again.
>
> Note that Haiku Beta 4 apparently finally fixed the issue with
> the enumeration of the virtio-block devices (see the ticket at
> https://dev.haiku-os.org/ticket/16512 ) - the tarball disk can
> now be found at index 1 instead of index 0.
>
> Signed-off-by: Thomas Huth <thuth@redhat.com>
> ---
>   tests/vm/haiku.x86_64 | 6 +++---
>   1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/tests/vm/haiku.x86_64 b/tests/vm/haiku.x86_64
> index 29668bc272..8b0d1b8954 100755
> --- a/tests/vm/haiku.x86_64
> +++ b/tests/vm/haiku.x86_64
> @@ -48,8 +48,8 @@ class HaikuVM(basevm.BaseVM):
>       name = "haiku"
>       arch = "x86_64"
>   
> -    link = "https://app.vagrantup.com/haiku-os/boxes/r1beta3-x86_64/versions/20220216/providers/libvirt.box";
> -    csum = "e67d4aacbcc687013d5cc91990ddd86cc5d70a5d28432ae2691944f8ce5d5041"
> +    link = 
> "https://app.vagrantup.com/haiku-os/boxes/r1beta4-x86_64/versions/20230114/providers/libvirt.box";
> +    csum = "6e72a2a470e03dbc3c5e808664e057bb4022b390dca88e4c7da6188f26f6a3c9"

Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>

To keep builds reproducible I'd rather stick the repo
with the base image used, something like:

 release_tag = 'r1beta4'
 self.ssh_root_check("pkgman drop-repo -y Haiku")
 self.ssh_root_check("pkgman drop-repo -y HaikuPorts")
 self.ssh_root_check(f"pkgman add-repo -y 
https://eu.hpkg.haiku-os.org/haiku/{release_tag}/$(getarch)/current")
 self.ssh_root_check(f"pkgman add-repo -y 
https://eu.hpkg.haiku-os.org/haikuports/{release_tag}/$(getarch)/current")

Note I used ssh_root_check() because the current script failed on a
heavily loaded machine, since ssh_root() doesn't check for failure.

Comments related to pre-existing issue, so:
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>

I can send my suggestions as followup patches.

Regards,

Phil.