v1: support user uploaded content (HMS-5212) #1443
Conversation
|
/jira-epic HMS-4930 |
schutzbot
changed the title
croissanne
changed the title
croissanne
changed the title
Codecov ReportAttention: Patch coverage is
|
croissanne
force-pushed
the
repos-by-ids
branch
from
bde9dcf to
13284a1
Compare
croissanne
force-pushed
the
repos-by-ids
branch
2 times, most recently
from
5f6659d to
55862c7
Compare
| } | ||
| } | ||
|
|
||
| return result, nil | ||
| } |
There was a problem hiding this comment.
Not huge fan of putting so much logic into a HTTP client. Ideally, the whole client.go could be auto-generated from OpenAPI, I remember we sligthly discussed it on Slack and I forgot what was the reason not to generate it. But there is a lot of boiler plate that could be generated.
There was a problem hiding this comment.
Currently it's a balance between putting stuff in v1 and clients. We could probably autogenerate some stuff and then have some logic around it in clients, as v1 is pretty big.
There was a problem hiding this comment.
Yeah not the right time and place, but I am all for generating as much as possible for OpenAPI clients. The generator we use supports selective generation, only endpoints we are interested in can be generated.
croissanne
force-pushed
the
repos-by-ids
branch
3 times, most recently
from
ed706a6 to
3c50a0e
Compare
|
Not sure if you meant to fix those nitpicks, I see no changes. It needs a rebase once again FYI. |
croissanne
force-pushed
the
repos-by-ids
branch
from
3c50a0e to
46a42a7
Compare
still working on the integration test |
croissanne
force-pushed
the
repos-by-ids
branch
10 times, most recently
from
4e53f5f to
a7c5d9b
Compare
The upload origin refers to repositories created from content the user uploaded directly.
This offloads most of the logic to the content sources client. This paves the way for supporting repositories with an upload origin, which only have a repository ID available. Furthermore this will allow us to treat the external origin repositories in the same way, based on ID, whilst only using the URLs for the red hat repositories.
This allows users to specify repository UUIDs from the content sources service instead of baseurls when using snapshots.
Uses content sources to fill in any data that's not already specified.
Uses content sources to fill in any data that's not already specified.
A small mock is needed, it can just return empty data, but image builder now needs content sources to do a simple compose to check if any of the repositories are defined in content sources.
croissanne
force-pushed
the
repos-by-ids
branch
from
a7c5d9b to
72bf558
Compare
This required accepting (snapshotted) payload and custom repositories by uuid.