feat!: implement artifact manifest type #29
Conversation
Signed-off-by: wangxiaoxuan273 <[email protected]>
Codecov Report
@@ Coverage Diff @@
## main #29 +/- ##
==========================================
- Coverage 57.18% 57.15% -0.03%
==========================================
Files 105 106 +1
Lines 10842 10914 +72
==========================================
+ Hits 6200 6238 +38
- Misses 3954 3981 +27
- Partials 688 695 +7
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. |
There was a problem hiding this comment.
How about create a new package named manifest/ociartifact since each type of manifest should have its own package? See https://github.com/oras-project/distribution/tree/main/manifest
Done. |
Signed-off-by: wangxiaoxuan273 <[email protected]>
Signed-off-by: wangxiaoxuan273 <[email protected]>
wangxiaoxuan273
changed the title
| } | ||
|
|
||
| // ArtifactManifest defines an ocischema artifact manifest. | ||
| type ArtifactManifest struct { |
There was a problem hiding this comment.
Curious why don't we use v1.Artifact?
There was a problem hiding this comment.
Because this struct needs to satisfy the manifest interface, also there are some compatibility issues as v1.Artifact uses oci.Descriptor but distribution uses distribution.Descriptor.
There was a problem hiding this comment.
ok, fair. Then there're a few disparities between the 2 structs:
- artifactType is a required field, so we don't need omitempty
- Better use
*distribution.Descriptorfor the optional subject
There was a problem hiding this comment.
Ok, I read the latest oci artifact manifest spec, artifactType is actually SHOULD.
There was a problem hiding this comment.
Changed subject to pointer after offline discussion. This gives func (m Manifest) References() a better design since it's much easier to decide whether it is nil. Thanks Wei!
Signed-off-by: wangxiaoxuan273 <[email protected]>
|
I think a builder struct (present for ocischema manifest) will be needed for artifact manifest, so that we can enforce some rules on artifact manifest creation. I'll add it in future when necessary, for now I'll just create artifact manifest by hand. |
Signed-off-by: wangxiaoxuan273 <[email protected]>
Implemented ArtifactManifest and related structs and functions. Unit tests included.
Part 2 of #21
Signed-off-by: wangxiaoxuan273 [email protected]