maritz

Currently there are different handlings for different paths of load/find where no result means either an error or an emtpy array. In some cases this can be a bit confusing...

Add option to mark behavior property as "use scored sorting"

1

Currently when you define a property with type behavior, it will never use scored indexing for sorts. This means that for example creating a createdAt behavior automatically means that sorting...

Currently you have to create and handle the very common createdAt and updatedAt fields yourself. It would be useful to have this done automatically on every model with the option...

Searching for a unique while other searches would prevent finding that unique still results in unique

2

When searching for a property that is indexed as a unique and that unique exists, it will always be returned - even if another search property would prevent it.

They have been removed during the rewrite for v2 and have not made it into the 2.0.0 release.

instead of the old continue_on_link_error behaviour, we should add a way to deepValidate before saving. Meaning all relationChanges (only link) get validated and if one of them is not valid,...

There are still quite a few things that aren't perfectly typed or even completely untyped. Any ideas or discussion about that should go in this issue.

Currently all unique string properties are case insensitive. Making this optionally case sensitive might be needed for some use cases (for example certain ids can be case sensitive base64).

Currently the defaultValue is always set during the execution of the constructor for each property. When using the defaultValue function to create cryptographically strong tokens (as an example) this produces...

Currently the last version on npm is still 0.7.21. (also git tags are missing for both 0.7.21 and 0.7.22)