Notable side-effects:
- loopback no longer exports "caller" and "arguments" properties
- kv-memory connector is now properly added to the connector registry
- the file "test/support.js" was finally removed
- Fix for current implimentation that returned all models that had any
assigned roles. Context was not carried into listByPrincipalType,
setting roleId as null.
In strong-remoting 3.x, we have stricken the coercion of inputs
methods that are expecting an Object will nolonger accept an array
as input, to preserve backwards compatibility we have added flag
allowArray in remote arguments, which would accept an array of objects
- Add a new User setting 'allowEternalTokens'
- Enhance 'AccessToken.validate' to support eternal tokens with ttl
value -1 when the user model allows it.
Fix the definition of "data" argument to
{ type: 'object', model: modelName, ... }
That way strong-remoting passed the request body directly to the model
method (does not create a new model instance), but the swagger will
still provide correct schema for these arguments.
This fixes a bug where upsert in relation methods was adding default
property values to request payload.
Current implementation of `app.model(modelName, settings)`
works as a sugar for model creation. In 3.0, this is
not supported anymore. This implementation reports an
error when sugar is used for model creation.
Includes:
- Updated app.model() method
- Fixed test cases reflecting the change
- Expose "keys()" at "GET /keys"
- Add a dummy implementation for "iterateKeys" to serve a useful error
message when the model is not attached correctly.
Fix token middleware to check if `req.loopbackContext` is active.
The context is not active for example when express-session calls
setImmediate which breaks CLS.
Rework the test suite to always report errors and correctly signal
when async tests are done.
This should prevent spurious test failures on CI servers that are
difficult to troubleshoot, because the error is reported for different
test case.
The remote-connector test has misconfigured the client (remote) model,
where the client model was trying to keep track of changes. That's
redundant because it's up to the server model (attached directly to the
database) to track changes.
This commit fixes that configuration and also cleans up the test code
a little bit to make it easier to distinguish between the remote
(client) model and the server model.
Use local registry in test fixtures to prevent collision in globally
shared models.
Fix issues discoverd in auth implementation where the global registry
was used instead of the correct local one.