This allows to combine `org.osbuild.mkdir` with the `osbuild.deployment`
mount and with the upcoming `org.osbuild.bind` mount. The use case is
that we need to create the dir `/var/home` so that `useradd` from inside
a ostree root works (there /home is a symlink and useradd will not
follow the symlink and create a dir in the target by itself).
This allows to write:
```json
{
"type": "org.osbuild.mkdir",
"options": {
"paths": [
{
"path": "/var/home"
}
]
},
"devices": {
"disk": {
"type": "org.osbuild.loopback",
"options": {
"filename": "disk.raw",
"partscan": true
}
}
},
"mounts": [
{
"name": "part4",
"type": "org.osbuild.ext4",
"source": "disk",
"target": "/",
"partition": 4
},
{
"name": "part3",
"type": "org.osbuild.ext4",
"source": "disk",
"target": "/boot",
"partition": 3
},
{
"name": "part2",
"type": "org.osbuild.fat",
"source": "disk",
"target": "/boot/efi",
"partition": 2
},
{
"name": "ostree.deployment",
"type": "org.osbuild.ostree.deployment",
"options": {
"source": "mount",
"deployment": {
"default": true
}
}
},
{
"name": "bind",
"type": "org.osbuild.bind",
"target": "tree://",
"options": {
"source": "mount://"
}
}
]
},
```
to fix this.
To test the curl sources it is very useful to have a small httpd
server that can serve an arbitrary directory. This helper will
ensure that via:
```python
with with osbuild.testutil.net.http_serve_directory(fake_httpd_root) as httpd:
port = httpd.server_port
# download from http://localhost:{port}/<any-path-under-httpd-root>
```
When moving to `bootc install to-filesystem` we will need support
for mounting the deployed disk and writing to the deployment root
this requires that we teach the users and selinux stages to
have them available. This is a first step towards this.
It also adds tests to ensure the options can be passed.
This is a preparation to allow adding mounts/devices to the users
stage so that we can eventually support bootc install to-filesystem.
It also adds some smoke tests for the schema to ensure it's still
valid.
In CoreOS Assembler, some hyperv artifact we `zip` for compression. This
new stage is modeled after the `org.osbuild.tar` stage with necessary
modifications.
Add an org.osbuild.systemd.unit stage using the new format for the
Environment option with two instances to the test manifest.
The contents of the new dropin file at
tree/usr/lib/systemd/system/boltd.service.d/30-boltd-debug.conf are:
[Service]
Environment="G_MESSAGES_DEBUG=all"
Environment="G_MESSAGES_TRACE=none"
Update the org.osbuild.systemd.unit stage to also support multiple
Environment options where each is an object with {key: value}. Enable
the allow_no_value option in configparser so we can add the multiple
entries.
Add the new options to the b.json test and update the diff.
The new file has the following contents:
[Unit]
Description=Create directory
DefaultDependencies=False
ConditionPathExists=|!/etc/myfile
ConditionPathIsDirectory=|!/etc/mydir
[Service]
Type=oneshot
RemainAfterExit=True
ExecStart=mkdir -p /etc/mydir
ExecStart=touch /etc/myfile
Environment="DEBUG=1"
EnvironmentFile=/etc/example.env
[Install]
WantedBy=local-fs.target
RequiredBy=multi-user.target
Add an empty file to the location where the service file will be
created in the b.json version of the test. This way, we will get a
content hash of the created file which is a slightly better test than
just knowing that it was created.
Note that, in the diff, the "before" checksum is the empty file hash:
echo -n '' | sha256sum
e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 -
In 2d2cdd8097 the file was replaced by
the generated json and it went unnoticed in the PR. Reverted and
updated the options to match the generated json file.
Support the Environment and EnvironmentFile options in the Service
section of the unit file.
The Environment option is set as an object with keys "key" and "value"
and the key is validated with a pattern.
Updated the stage to special-case the Environment option.
Tiny tweak to remove some boilerplate related to tmpfile handling.
The pytest `tmp_path` fixture gives us the tmpdir without having
to worry about cleanup etc (and in a slightly more concise way).
The `amend_secrets()` does not work with real files so there is
no need to mock cachedirs or create fake input files. This commit
just removes those.
It also changes the checksum to `"1"*64` to make it very clear
that the checksum has no significance in this test.
There was a regression with the secrets adding of rhsm for the
curl source. This was my mistake (sorry!). Here is a regression
test that would have prevented this (if we have had it earlier).
Similar to the previous commit to include a `inputs_service` fixture
this does the same for `source.SourcesService` imports.
Note that we cannot easily share the helpers so we have to life with
a bit of very similar but duplicated code. To fix this we would have
to have a shared confftest.py that pytest can find. Which would mean
that we need to put the tests under a common dir that is reachable
via __init__.py files (which we currently not have because stages,
inputs etc do not have a __init__.py so python does not considers
them modules).
By default "bootc" will refuse to work on a non-selinux system if
the bootc container requires selinux. This is a sensible approach
in general but for us it's tricky because we want to be able to
generate images when running on developer machines or CI machines
that may not necessarily have selinux. So make bootc more relaxed.
