Age | Commit message (Collapse) | Author |
|
If the user hasn't defined any bundles a mega-image won't be
created and there's no need to recopy the mega image rootfs.
The original rootfs can be used in this case.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Rather than trying to determine the manifest link name by removing the value
of the DATETIME variable from the image name duplicate the logic from
rootfs-postcommands.bbclass in OE-Core to derive the manifest name in the
same way.
This prevents issues when the IMAGE_NAME (and thus manifest name) differ
significantly from the IMAGE_NAME's used in OE-Core.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
This reverts commit ba2aeec203b09d96c6e81c85cdd57054bc670c93.
Manifest files are not currently writting with IMAGE_NAME_SUFFIX
|
|
Use manifest files written to DEPLOY_DIR_IMAGE, rather than
SWUPDMANIFESTDIR, during create_rootfs as the latter are (currently)
only written for package based bundles.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
The image name suffix is modifiable via the IMAGE_NAME_SUFFIX
variable, therefore when constructing image filenames we should
use the same variable.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Consistently use docstrings for documenting methods in lib/swupd/*
and ensure every method is documented.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
The manifest files are used for various things so we must be sure
they are available, even if the swupd inputs were staged from a
shared state.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
When creating a composed rootfs for swupdimage images we also write
an updated manifest of the composed image's content. To better match
behaviour of other manifests in DEPLOY_DIR_IMAGE also create a
symlink to the latest manifest.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
When swupdimage rootfs are constructed they rely on reading the
contents of the rootfs.manifest files for each bundle they contain
in order to determine the image's contents. However it's possible,
especially with the addition of a sstate accelerated
do_stage_swupd_inputs, that the manifests will have been generated
in a different bitbake invocation and therefore that the DATETIME
in the IMAGE_NAME will be different.
This change accounts for that eventuality by falling back to the
manifest's symlink when the composed manifest filename (including
DATETIME) doesn't exist.
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|
|
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
|