To allow generic manifest walking, we define an interface method of
`References` that returns the referenced items in the manifest. The
current implementation does not return the config target from schema2,
making this useless for most applications.
The garbage collector has been modified to show the utility of this
correctly formed `References` method. We may be able to make more
generic traversal methods with this, as well.
Signed-off-by: Stephen J Day <stephen.day@docker.com>
When we push a schema1 manifest, we encode history information from the
image JSON into v1Compatibility strings for the respective layers. The
"author" field was not being set in these v1Compatibility strings, so if
a parent layer had an author set, it would not be preserved after
pushing through a schema1 manifest and repulling, so the image ID would
change after the pull. This change preserves the authorship information
for parent layers so that the image ID does not change.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
Most places in the registry were using string types to refer to
repository names. This changes them to use reference.Named, so the type
system can enforce validation of the naming rules.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
Add schema2 manifest implementation.
Add a schema2 builder that creates a schema2 manifest from descriptors
and a configuration. It will add the configuration to the blob store if
necessary.
Rename the original schema1 manifest builder to ReferenceBuilder, and
create a ConfigBuilder variant that can build a schema1 manifest from an
image configuration and set of descriptors. This will be used to
translate schema2 manifests to the schema1 format for backward
compatibliity, by adding the descriptors from the existing schema2
manifest to the schema1 builder. It will also be used by engine-side
push code to create schema1 manifests from the new-style image
configration, when necessary to push a schema1 manifest.
Signed-off-by: Aaron Lehmann <aaron.lehmann@docker.com>
