You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Whether this is an issue or not depends on whether we think people can't support query parameters but still want to support uploading children? For example a non-code server - ala S3. However, in those cases I wonder if those are read-only servers, or the "write" path (if any) is via the native S3 interfaces instead.
So, the question is... do we want to support "import" w/o requiring ?nested ? Or do we think the "are you sure?" check is unnecessary even if they can support query parameters?
Options:
1 - keep it as is
2 - drop ?nested and let the presence of the children be good enough - remove "are you sure?" check
3 - add a /import API that implies ?nested, similar to how /export implies ?inline=*
The text was updated successfully, but these errors were encountered:
duglin
changed the title
if ?nested isn't support then import or create() with children can never be supported
if ?nested isn't supported then import or create() with children can never be supported
Jan 10, 2025
- typo and wording clarifications in main spec
- remove duplicate "compatibility" in schema, it's part of core attributes now
- replaced `?export` with `?compact` and just removes the dup info. Any
inlining would need to be requested by the client
- defined `/export` to be `/?compact&inline=*,model,capabilities`
- remove ?nested
- tweaks valid chars for ID: `[a-zA-Z0-9_][a-zA-Z0-9_.\-~@]{0,127}`
- IDs must be no longer than 128 chars
- s/$structure/$details/g
Fixes: xregistry#234Fixes: xregistry#233Fixes: xregistry#232Fixes: xregistry#220Fixes: xregistry#229Fixes: xregistry#224Fixes: xregistry#237
Signed-off-by: Doug Davis <[email protected]>
- typo and wording clarifications in main spec
- remove duplicate "compatibility" in schema, it's part of core attributes now
- replaced `?export` with `?compact` and just removes the dup info. Any
inlining would need to be requested by the client
- defined `/export` to be `/?compact&inline=*,model,capabilities`
- remove ?nested
- tweaks valid chars for ID: `[a-zA-Z0-9_][a-zA-Z0-9_.\-~@]{0,127}`
- IDs must be no longer than 128 chars
- s/$structure/$details/g
Fixes: xregistry#234Fixes: xregistry#233Fixes: xregistry#232Fixes: xregistry#220Fixes: xregistry#229Fixes: xregistry#224Fixes: xregistry#237
Signed-off-by: Doug Davis <[email protected]>
Signed-off-by: Simon Heimler <[email protected]>
Whether this is an issue or not depends on whether we think people can't support query parameters but still want to support uploading children? For example a non-code server - ala S3. However, in those cases I wonder if those are read-only servers, or the "write" path (if any) is via the native S3 interfaces instead.
So, the question is... do we want to support "import" w/o requiring
?nested
? Or do we think the "are you sure?" check is unnecessary even if they can support query parameters?Options:
1 - keep it as is
2 - drop
?nested
and let the presence of the children be good enough - remove "are you sure?" check3 - add a
/import
API that implies?nested
, similar to how/export
implies?inline=*
The text was updated successfully, but these errors were encountered: