Xible Versions Save

Added

• Introduced /api/flows/<flow-name>/instances/<flow-instance-id> endpoint to get statuses of a specific flow instance.

Fixed

• Better handling on broken worker IPC's when trying to stop the worker from the main process.

Added

• Support for different base href's. This allows XIBLE to be hosted from subpaths behind a reverse proxy for example. Set the BASE_HREF environment variable accordingly to ensure XIBLE correctly handles the base path.

Added

• Support for different stores. This makes it possible to store flows, flow-states, and the vault outside the filesystem. The first supported store besides FileStore is PgStore for PostgreSQL. You can use the store by providing --flow-store-type=PgStore and --connection-string=<connection-string> parameters when starting XIBLE. (#94)

Changed

• Upgrade Cypress for E2E testing.

• Updated minumum Node.JS version to 20.

Fixed

• The pattern validation for the name of a new flow in the browser was not escaped properly.

Added

• It is now possible to xiblepm flow install <user>/<flow> instead of xiblepm flow install <flow> --publish-user-name=<user>. xiblepm flow search returns a list in the same format.

• Whenever a nodepack is under development and the package.json name attribute does not start with 'xible-np-' or 'xible-nodepack-', a warning is thrown, indicating that the nodepack cannot be published to the registry.

• The CLI command xiblepm flow delete <flowname> has been added to delete a flow which is published to the registry. The documentation has been updated.

• The UI has an option to publish flows from the 'Flows' page. This acts similarly as the command line xiblepm flow publish <flowname>.

• It is now possible to navigate the node selector within the editor using the up and down keys to select different nodes, and the space key to insert a node.

Changed

• The .xiblerc.json file location is now configurable in the config.json and as a result through the xible config command.

Fixed

• Calling node.getData() on a node that has no dataStructure, would throw.

Changed

• Flows are now stored per user. This means that the registry can hold multiple flows by the same name, but published by different users. All functionality has been updated to reflect this change. This includes the UI and the cli commands such as xiblepm flow install.

Added

• Within the structure.json of a node, developers can now configure the input data fields using the dataStructure object, instead of creating a separate editor/index.htm. This improves development time for new nodes by a fair bit. If an editor/index.htm is present, it will overrule any configured dataStructure within the structure.json. (#95)

• A new convenience method node.getData(dataName, state) to fetch all data for a data field, including related inputs if so configured. This behaves similarly as a combination of fetching data through NODE.data.x and NODE.getInputByName(x).getValues(). node.getData() is the new preferred method of fetching data. (#95)

Changed

• nodeIo.isConnected() actually returns a boolean now, instead of the amount of connectors.

Fixed

• Describe/help for a node, in the flow editor, was no longer visible. When a node was in focus, and the 'h' key was pressed, the help page would not show up. Similarly when clicking on the help button for that node.

Fixed

• Some mouse actions (scroll for zoom, doubleclick to add a node) were not registered in the flow editor in certain scenarios.

Added

• A feature has been added to edit the node data/input values in a separate dock, which hosts more space. The amount of space available here is especially useful for nodes which potentially contain large amount of data, such as the function node.

• When hovering the mouse cursor over a node, a box with options appears that could previously only be triggered by hitting the right key combination. These include;

  • Open the help/details page of a node.
  • Delete a node from a flow.
  • Edit the contents of a node in a seperate side panel.

Fixed

• Within the dock of the flow editor, the 'clear'-logs button was not completely clickable because of overlap by the tab list.

Added

• Input fields within nodes that have type=password, now include a toggle to make the password visible from the editor.

• Data fields that are stored in the vault now have a shield icon available in the editor to signify this. (#90)

• The 'object' nodepack now contains a object.path node to parse jsonpath expressions. (#62)

Changed

• A focused input field now focuses the surrounding node, preventing the loss of visibility on fields that are not required. These fields hide by default of the focus is not on the node, which could make it cumbersome to edit these fields.

• The xible.flow.onstart and waitfor nodes have been renamed to follow naming conventions. They are now called xible.flow.on-start and wait-for. Existing flows which are using these nodes will have to replace it with the newly named nodes.

• Within the editor, pasting a selection of nodes will center the copy in the viewport. (#19)

• For new installations, the default amount of statuses visible on any node in the editor is set to 5. This can be changed through the settings. The value is stored in the configuration file as editor.nodes.statuses.max.

Fixed

• The 'http' and 'function' nodepacks are now loaded by default. They were included in the package.json, but not in the default nodepack loader. In previous version it is still possible to simply install these nodepacks directly from the registry.