Implement `stop` as an alias for `disconnect`. This way LB4 applications
don't have to include custom `stop` implementation in every datasource
file.
Signed-off-by: Miroslav Bajtoš <mbajtoss@gmail.com>
Add a new hidden property `__parent` that's automatically set on all
instances of embedded models.
For backwards compatibility, this feature is not enabled by default.
You can turn it on by adding the following line to `server/server.js`
file:
app.registry.modelBuilder.settings.parentRef = true;
- You can do new Array(4) to generate an array with 4 empty items
- Doing the same with list caused an error
- This broke lodash.deepclone
- This PR restores compatibility with the array constructor
Fixes#1798
If establishing a database connection is slow
and database migration runs and there are many
models, sql operations are queued up and this
leads to the node.js max emitters exceeded
warning.
A default value for max emitters has now
been introduced, and it can also be configured
in datasources.json.
Co-authored-by: Dominique Emond <dremond@ca.ibm.com>
Before this change, when both the PK value (`id`) and the `data` object
were provided as plain-data values (e.g. as received in a JSON request),
and the connector was using a complex PK type (e.g. `ObjectID`
in MongoDB), then `replaceById` operation was printing confusing
warnings:
WARNING: id property cannot be changed from 5d39775a59f5f541513c5e05
to 5d39775a59f5f541513c5e05 for model:Post
in 'before save' operation hook
WARNING: id property cannot be changed from 5d39775a59f5f541513c5e05
to 5d39775a59f5f541513c5e05 for model:Post
in 'loaded' operation hook
This commit fixes the problem by applying the same type coercion on the
PK value (`id`) as has been applied by the model constructor on the PK
property (`data.id`).
Signed-off-by: Miroslav Bajtoš <mbajtoss@gmail.com>
When writing tests, for performance reasons we often want to reuse
the same data-source instance for many tests suites. At the same time,
we want to keep such test suites independent and allow them to reuse
the same model name for different model classes.
Juggler does support redefinition of a model with the same name.
This change is adding a new API called that allows tests to remove all
old models before creating new ones. This API would be typically
called from a `before` hook.
Signed-off-by: Miroslav Bajtoš <mbajtoss@gmail.com>
Defer automigrate/autoupdate until we are connected, so that connection
errors can be reported back to callers.
Fix postInit handler to not report connection error to console.log
and via dataSource "error" event in case there is already an operation
queued. When this happens, we want the error to be handled by the
queued operation and reported to its caller.
Signed-off-by: Miroslav Bajtoš <mbajtoss@gmail.com>
add beginTransaction method which calls begin
method from the Transaction class which in turn
calls the connector's beginTransaction method if
it supports transactions.
Co-Authored-By: Miroslav Bajtoš <mbajtoss@gmail.com>
Rework setup of the Operation Hooks test suite for memory connector
so that the "unoptimized" variant disables not only atomic CRUD
operations, but also geo queries.
This way we can test both "near" querying implementations:
- When the connectors supports "near" queries natively.
- When "near" queries are executed at LoopBack side in-memory.
Fix DataSource constructor to create a shallow copy of the settings
object provided by the caller. This prevents surprising behavior
where changes made to `ds.settings` were picked up by the provided
config object, as observed e.g. in tests of our MongoDB connector.
- On include = true or include = 1 the lib crashes with
"TypeError: includes.forEach is not a function".
- checking for boolean and number type and return empty array.
- fix: include symbol and function in normalization
- fix: review changes; adding debug stmt + early exit for tests
Before this change, when a property was configured with a default value
at LoopBack side and the database was returned a record with a missing
value for such property, then we would supply use the configured
default.
This behavior is problematic for reasons explained in #1692.
In this commit, we are fixing DAO operations like `find` and
`findOrCreate` so that they do NOT apply default property values on
data returned by the database (connector).
Please note that most of the other CRUD methods were already not
applying default values on database data as long as the connector
provided native implementation of the operation.
When converting plain-data object values into model instances,
correctly handle the case where the constructor functions is a class
constructor and must be invoked via `new`.
Miroslav Bajtoš
Dimitris Xalatsis
Agnes Lin
Diana Lau
Hage Yaapa
Jeremy Nagel
Francisco Buceta
Ayeni Olusegun
Dominique Emond
Biniam Admikew
Janny
spurreiter
Bill Matson
Agnes Lin
Raymond Feng