Development Process

Introduction

There are tools in place which provide feedback to the developer to verify that the created work is correct. This is done via certain automated checks. The following workflow needs to be followed to make use of the automated checks.

Branches

In the repository for the connector, the following branches always exist:

• The main branch.

  • The main branch always contains releasable connectors and is not used for development.
  • It is only possible to merge to main via a pull request.
  • It is only possible to merge to main with approval from the UiPath team.
  • The pipelines running in the repository need to succeed before the pull request can be merged.

• A validation branch.

  • The validation branch is used to perform the validations.
  • It is only possible to merge to validation via a pull request.
  • The pipelines running in the repository need to succeed before the pull request can be merged.

Development Flow

Next to the existing branches main and validation, you can use your own preferred workflow for connector development and create the branches necessary. The only limitation is that only rebase commits are allowed when merging between branches.

Step	Description
1	When the work is finished, the work needs to be merged to the validation branch via a pull request (PR).
2	Merge to the main branch. This can only be done via a validation branch after a connector passes validation.

See the illustration below.

Pull Requests (PR)

Whenever work is pushed to a remote branch, it is possible to create a pull request containing the work. This can be achieved by going to GitHub. A pop-up will be shown where you can easily create a pull request. See the illustration below.

Step	Description
1	When you open up a pull request, make sure you set your base branch. If you open up a pull request to merge work to validation, select validation. Otherwise, choose your own branch to which you want to merge the work.
2	Add a meaningful title and a short message to describe the work.
3	Create the pull request.
4	Whenever all checks succeed, you can merge the pull request. You can do this by clicking the button Rebase and merge at the bottom of the pull request. See the illustration below.

Feedback

During development, it is possible to see the status of the automated check. To get this feedback, make sure you create a pull request to merge your work. Each pull request has a status section at the bottom of the conversation page. In this section, you will see the state of each pipeline separately. See the illustration below.

When one of these checks is not successful, a log file will be created containing the errors which are encountered. If such a pipeline errors, then a Slack message will be sent to the channel created for the development of the connector. See the illustration below for an example.

Note: Some of the automatic checks make use of the sample data provided in the connector. It is recommended to make this sample data available as early as possible during development, to make the best use of these automated checks.

Automated Checks

The following automated checks are available.

Transformations

To create robust and correct transformations, an automated check is executed for each pull request (PR), and when code is pushed to a remote branch, to validate that the transformations are correct. This automated check does the following tasks:

• The sample data provided by the connector is uploaded to SQL Server and Snowflake and performs all the transformations steps (executing the models and tests). The automated check will provide feedback when transformations or tests fail.
• All transformations are checked to see if they conform to the coding guidelines. The automated check will provide feedback when the coding guidelines are not met.

Deliverables Check

When a connector is ready to be validated, the content should be merged in the validation branch by means of a pull request (PR). This PR can only be merged when the deliverables check is successful, which checks whether the required content is present.

Automatic Spelling Check

To support writing connector documentation without errors, an automatic spell check is performed for every pull request (PR). The check validates all .md files and provides suggestions for incorrect words. Install the recommended extension code-spell-checker to perform the spell check while writing the documentation.