* Start rerouting VSCode cell execution APIs.
* Add more conversion code.
* Convert VSCode notebook registrations into ADS equivalents.
* Update vscode notebook provider kernels when notebook controller's supportedLanguages are set.
* Update an error message.
* Add another session argument.
* Add base classes for converting notebook serializers.
* Disable some vscode notebook methods.
* Disable more vscode APIs.
* Disable more stuff.
* Start implementing serializer notebook data conversions.
* Use direct references to extension host notebook methods, rather than azdata ones.
* Add a comment.
* Remove a space.
* Use import type to fix module loading errors.
* Use internal cancellation token class.
* Start adding cell output conversion.
* Convert data from byte array to a string.
* More output work.
* Use a Set for proxy filtering.
* Start adding tests.
* Include metadata in cell conversion. Fix other test failures.
* Fix serialize tests.
* Add more tests.
* Remove wildcard characters from vscode filenames.
* Start implementing session details.
* Add more kernel info.
* Add kernel spec.
* Add Future callback wrapper class.
* Start implementing execute conversion.
* Pass notebook URI to requestExecute.
* Start working on CellExecution methods.
* Move some code around to fix layering issues.
* Use proxy to access browser code, rather than direct imports.
* Move files around to fix layering issues.
* Remove unused imports.
* Start implementing some notebook cell execution behaviors.
* Revert some unnecessary extHost API changes.
* Check for nbformat.
* Also handle nbformat in serialize case.
* Active notebook extensions when resolving NotebookInput.
* Fix nbformat handling.
* Disable VSCode notebooks code.
* Filter out notebook services from registration assertion.
* Wait for providers to load before calling canResolve.
* Use controller's viewType for notebook provider ID, instead of controller ID.
* Start adding extHostNotebook tests for new APIs.
* Re-order proxy calls.
* Remove commented code.
* Move vscode provider files to browser folder. Fix RPC serialization issues by using readonly field instead of getter for providerId.
* Add a comment.
* Remove unnecessary dispose call.
* Handle disposable from registerExecuteProvider.
* Remove a comment.
* Remove unnecessary provider fields.
* Remove reference to notebook service to fix circular reference issue in stringify.
* Add object types for methods in ADSNotebookController.
* Wait for controller languages to be ready before marking session manager as ready.
* Add correct promise.
* Add undefined return type for optional supportedLanguages property.
* Refine promise logic.
* Move vscode functionality back to ExtHostNotebook, since the NotebookService can't be passed back over RPC (some kind of circular reference error).
* Fix remaining issues from last commit.
* Replace "not implemented" methods with placeholder return types in order to enable testing.
* Also wait for execution handler to be set before marking session manager as ready.
* Fix usage of NotebookRegistry when updating provider description languages.
* Refine file extension conversion.
* Fix file extension conversion to match ADS extension behavior.
* Emit new provider registration event when adding supported languages.
* Remove checks for duplicate file providers and kernels.
* Fix a test failure.
* Fix file extension parsing.
* Use default executeManager if one isn't defined for provider in notebookModel.
* Add descriptors for waiting on standardKernels registration.
* Increase timeout
* Add an error message.
* Start working on retrieving default kernel from registered providers, rather than always falling back to SQL.
* Revert "Start working on retrieving default kernel from registered providers, rather than always falling back to SQL."
This reverts commit 1916ea1ce3a0072f51bec683116dc7bb6c7aefdc.
* Emit activation events after provider registration.
* Wait on standard kernels availability when getting an execute provider.
* Throw an error if session manager isn't ready yet.
* Actually resolve language promise correctly.
* Add some checks for undefined notebook data objects.
* Create kernel spec data correctly.
* Add extension changes for local testing only.
* Clean up test class.
* Add a reminder comment.
* Undo commented out notebook stuff
* Temporarily hard code default kernel.
* Retrieve default kernel in notebookModel if it's not already provided.
* Revert an import change.
* Remove unnecessary method from extHostNotebook.
* Move an interface around.
* wip
* Check for proposed API for some VSCode extHost methods.
* Remove a comment.
* Fix notebookUtils tests.
* Fix notebookModel tests.
* Fix notebookFindModel tests.
* Fix notebookViewsExtension tests.
* Fix remaining notebookView tests.
* Refactor output conversion functionality into separate methods.
* Update some unit tests for output conversion.
* Move a method.
* Rename conversion methods to fit acronym styling.
* Add another conversion test case.
* Revert local testing changes.
* Remove old method.
* cleanup
* Remove some comments.
* Move localized string to locConstants.
* Add a space to loc string.
* Add comments to new SQL Carbon Edit tags.
* Create constants for nbformat and nbformat_minor.
* Move some vscode-only fields to proposed APIs.
* Check for valid state
* Properly null check
* Adding logging for provider wait timeouts.
* wip update
* Fix compile
* Switch to cell edits
* Update docs
* Remove custom output type
* cleanup
* fix
* cleanup
* more cleanup
* Fixes
* Fix tests and lint errors
Co-authored-by: Cory Rivera <corivera@microsoft.com>
Fixes#5231
- Add stdin handling. Has to be at UI level so add plumb through handling
- Add unit tests
- Add new StdIn component.
Testing:
Unit Tests and manual testing of following:
- Prompt for password using `getpass` in python.
- Password prompt is hidden since "password" is true.
- Hit enter, it completes
- prompt, stop cell running, StdIn disappears
- prompt, hit escape, stdIn disappears and stdIn request is handled.
Issues: focus isn't always set to the input even though we call focus.
Will investigate this further.
This is a partial fix that lays groundwork for full "Prompt to connect" if a kernel needs a connection.
I am waiting on Yurong's refactoring of connection handling before doing any of the prompt work.
- Adds kernel metadata about whether a connection is required.
- For Jupyter, only Spark kernels are listed as requiring a connection
- If this is true and there's no active connection, will show notification and not call execute
In the future, this path will still be used if user is prompted to connect and cancels out.
The future change will be to inject a "connect" handler from notebook.component to the cell callback and use to set connection context
