Misleading docs on SyncManager pull() implementation

[ddenev]

The current documentation on SyncManager's pull() method implementation says the following:

The code example is misleading and contradicting the text above because:

1. it uses lastFinishedSyncStart to limit the returned result set (i.e. the returned items will be different than the full data set for the collection) and
2. it returns { items: data } which is expected to return the full data set for the collection (the SyncManager internally does a snapshot comparison to find the delta b/w the current collection and the returned collection)

Implementing the above example leads to SyncManager emptying the local collection after a 2nd sync when nothing has changed on the backend.

IMHO, the example should be updated to show both cases:

1. when using the lastFinishedSyncStart to get only the changes (added, modified, removed) and then return { changes }
2. when not using lastFinishedSyncStart to get all items in the collection and then return { items }

[maxnowack]

You’re absolutely right. Would you be willing to open a pull request to update the documentation? 🙂

[ddenev]

@maxnowack, sure, I can do that. And I can also suggest an update to the https://signaldb.js.org/orm/#adding-static-methods-to-collections example - the method cannot be static as it uses this and has to be an instance method 🙂. I would also rename the getPublishedPosts to findPublishedPosts as it returns a 'Cursor` and not the actual posts, if you don't mind.

[maxnowack]

Another good catch! Yes, that would be great as well!