Is your feature request related to a problem? Please describe.
createdAt and updatedAt are timestamps ;
userId is the id of the person that created the document (this is needed to make the "owners" permission group to work)
This configuration could belong to the database driver OR use a convention in Vulcan
The problem is that they rely on "onCreate"/"onUpdate" callbacks, that themselves require a server framework.
Describe the solution you'd like
createdAt, updatedAt and userId should be edge cases. They should be added automatically to all schemas (eg during model creation?) and handled accordingly in resolvers.
Except if the user explicitely override "onCreate"/"onUpdate" to be null (or a custom function) in a server schema, in which case we should not run the default function, and run the custom function if there is one.
_id schema could also be simplified to "_id: {}" or "_id: true" so that we don't need to make it explicitely optional,
Describe alternatives you've considered
Setting at the database level using drivers, but this works only for timestamps.
Using a "createSchema" function? The problem is that it means having this extra step for all Vulcan schemas, instead of a nice, raw JSON.
Providing a "baseSchema" that we can spread into each schema? Doesn't solve createdAt/updatedAt issue
Is your feature request related to a problem? Please describe.
createdAt and updatedAt are timestamps ; userId is the id of the person that created the document (this is needed to make the "owners" permission group to work) This configuration could belong to the database driver OR use a convention in Vulcan
The problem is that they rely on "onCreate"/"onUpdate" callbacks, that themselves require a server framework.
Describe the solution you'd like createdAt, updatedAt and userId should be edge cases. They should be added automatically to all schemas (eg during model creation?) and handled accordingly in resolvers.
Except if the user explicitely override "onCreate"/"onUpdate" to be null (or a custom function) in a server schema, in which case we should not run the default function, and run the custom function if there is one.
_id schema could also be simplified to "_id: {}" or "_id: true" so that we don't need to make it explicitely optional,
Describe alternatives you've considered
Setting at the database level using drivers, but this works only for timestamps.
Using a "createSchema" function? The problem is that it means having this extra step for all Vulcan schemas, instead of a nice, raw JSON.
Providing a "baseSchema" that we can spread into each schema? Doesn't solve createdAt/updatedAt issue