| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
| |
I blithely replaced runcmd(['mv', src_dir, dest_dir]) with
os.rename(src_dir, dest_dir), without realising that mv does
extra work when the target is a directory that already exists.
|
| |
|
|\ |
|
| | |
|
|/
|
|
|
| |
This reverts commit cc04afc00f35dc00e03c9ab207bbc2e5513812d0, reversing
changes made to 656295846bffb4eda9fa3869d04fd7c042b2e37a.
|
| |
|
| |
|
|\ |
|
| |
| |
| |
| | |
systems.
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Accordingly the rsync manual:
"A trailing slash on the source changes this behavior to avoid
creating an additional directory level at the destination. You can
think of a trailing / on a source as meaning "copy the contents of
this directory" as opposed to "copy the directory by name".
|
| |
| |
| |
| | |
Unmounting the root disk as the side effect of turn it to be read only
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Or else this
ln -s -f update1 /mp/systems/default
will do this
'/pp/systems/default/update1' -> 'update1
When we want
'/pp/systems/default' -> 'update1
|
| | |
|
| |
| |
| |
| |
| |
| | |
We now only support one system-kind, so there's no need for the
factory stuff, and at this point it only serves to obfuscate
and complexify.
|
| | |
|
| |
| |
| |
| |
| | |
Also delete the infrastructure for disk-based system-kinds
from builder2.py.
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Because the whole cachedir was considered, and not just the
artifacts tree, it would recursively consider every file an
artifact, incuding cached gits and ccache.
This made it slower and remove things before they were due, since
only the artifact cache is guaranteed to set mtimes on access.
This wouldn't cause incorrect builds, since any cached objects get
re-generated, but it was definitely not desired behaviour.
|
| |
| |
| |
| |
| | |
This was left over from experimentation, it works with sh, and is
faster.
|
|\ \
| | |
| | |
| | |
| | |
| | |
| | | |
'origin/baserock/richardmaw/S8043-timestamp-extension-output'
I verified that a deployment still works, and that the output is
time-stamped.
|
| | |
| | |
| | |
| | |
| | | |
This is required to get real-time output and the timestamps meaning
anything useful.
|
| | |
| | |
| | |
| | |
| | | |
This requires a patch to cliapp to actually print the output, as there
is a bug that causes it to always output to a pipe.
|
| | |
| | |
| | |
| | |
| | | |
Also clarify how the directory containing it gets created
to avoid a bit of code duplication.
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | | |
We now only support one system-kind, so there's no need for the
factory stuff, and at this point it only serves to obfuscate
and complexify.
|
| | | |
|
|/ /
| |
| |
| |
| | |
Also delete the infrastructure for disk-based system-kinds
from builder2.py.
|
|\ \
| |/
|/|
| |
| |
| | |
'origin/baserock/richarddale/check-repo-alias-option'
Reviewed-By: Daniel Silverstone <daniel.silverstone@codethink.co.uk>
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
-s, or --protect-args prevents the file path components of destination
or source paths being interpreted by the remote shell.
This is for wildcards or other shell features, but it breaks when paths
have whitespace.
We tend to always use absolute paths, so all uses of rsync now pass -s.
kvm.write needs it, since the disk can be written to a path with spaces.
Nfsboot and ssh-rsync need it because version labels are used, which may
have spaces, and temporary directories are used, which could have spaces
in weird TMPDIR configurations.
|
| | |
|
| | |
|
| |
| |
| |
| | |
I don't know how that one slipped through.
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Verify if are using and older extlinux configuration and upgrade it
if the case, by checking if the "default" symbolic link exists on the target.
Note that with the symbolic link we don't need to update extlinux configuration after an upgrade
|
| | |
| | |
| | |
| | | |
systems.
|
| | | |
|
|\ \ \
| |/ /
|/| |
| | |
| | |
| | |
| | |
| | | |
git://git.baserock.org/baserock/baserock/morph
There were no major objections to the design and it has been sitting in
the merge queue long enough that it was decided to merge it.
It has been reviewed and lightly tested by Jonathan Maw.
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This removes staging areas and extracted chunks from --tempdir.
Then asks the local artifact cache what artifacts it
has and how old they are, removing all sources older than
--cachedir-artifact-delete-older-than, and may delete other sources that
are younger than --cachedir-artifact-keep-younger-than if it still needs
to make space.
|
| | | |
|
| | | |
|
| | | |
|
| |/ |
|
|\ \
| | |
| | |
| | |
| | |
| | | |
git://git.baserock.org/baserock/baserock/morph
Reviewed by Richard Maw, Lars Wirzenis, Daniel Silverstone and Jonathan Maw.
|
| |/
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The VirtualBox deployment write extension will configure networking in
the following manner:
- One host-only network interface will be used to group the virtual
machines together (and to give access between the host and the virtual
machines). It will be bound to eth0 of the virtual machine.
- One NAT (or Bridge) network interface will be used to allow the
virtual machines access to the wider network. This will be bound to eth1
of the virtual machine.
Now deployment to virtual box will require that both HOST_IPADDR and NETMASK
environment variables be set, and also configuration for eth0 and eth1 is
expected in the NETWORK_CONFIG environment variable.
This commit also replaces some run_cmd with ssh_runcmd.
|
|\ \
| | |
| | |
| | |
| | |
| | | |
git://git.baserock.org/baserock/baserock/morph
Reviewed by Richard Maw and Lars Wirzenius
|