-
Notifications
You must be signed in to change notification settings - Fork 29.4k
Working Copies
For quite some time, VSCode only provided an editor for textual edits on files or virtual documents. All logic for how to save, revert, backup these text models was not made for reuse. We then added other editor capabilities, such as custom editors and notebooks and felt the need for a unified API to deal with edits to documents: working copies. The goal was to have the same editing experience in these new editors that users got used to when editing text files.
The API shape (IWorkingCopy
) of a working copy provides the minimal structure for participating as an editable "thing" in the workbench. In most cases the user would edit it using an editor. If you build a new component that allows for editing, you can introduce your working copy into the workbench to benefit from:
- unified (auto)save/revert handling
- backups ("hot-exit")
- dirty tracking
The fundamental pieces of a working copy are:
-
resource
: to identify the working copy among others -
typeId
: to tell apart your working copy from others with the same resource -
capabilities
: to tell if a working copy is untitled or not (e.g. untitled working copies are never auto saved) - events (
onDidChangeDirty
,onDidChangeContent
,onDidSave
): to provide dirty tracking and backup support - methods (
backup
,save
,revert
): to actually provide the necessary implementation for the workbench to call on
Note: you will likely already have a model in your component that can map directly onto the working copy shape. For example, the ITextFileModel
that is used for every textual edit in files implements IWorkingCopy
so that essentially the text file model IS the working copy that gets registered.
As soon as your working copy comes to existence, call IWorkingCopyService.registerWorkingCopy
. From that moment on, the workbench will observe the working copy for changes using its events.
As soon as your working copy becomes void, dispose the registration again!
Note: it is an error to register the same working copy more than once. The combination of resource
and typeId
is used for comparing working copies with each other.
As soon as the working copy reports a content change, the workbench will call the backup
method to store a backup of the unsaved state in the backup location. You can:
- provide the
content
to backup as raw buffer - provide some
meta
information as object associated with the backup
Note: backups are automatically deleted once the working copy is saved and no longer reports as dirty.
The workbench is not in charge of resolving your working copy, it is up to you as provider. In order to restore potential backups that might be present from a previous session, you have to use IWorkingCopyBackupService.resolve
with the identifier (resource
and typeId
) and set the contents of the working copy to that when resolving. In addition, you should mark your working copy as dirty.
The contract of these methods is that after the operation succeeded, the working copy is no longer dirty. In addition, the onDidSave
and onDidChangeDirty
should have fired.
In most ways, working copies are editor agnostic, but there are cases where the workbench needs to convert a working copy to a functional editor. For example: when there are backups of working copies stored on disk and the workbench is asked to open, we eagerly restore all such working copies as editors so that the user is always aware of unsaved changes. However, this requires a bit of glue code so that the workbench knows how to open an editor from a working copy: IWorkingCopyEditorService
When you introduce a working copy to the workbench, make sure to call IWorkingCopyEditorService.registerHandler
:
-
handles
: given a working copy, return whether it is yours or not -
isOpen
: given a working copy, return whether you know the working copy is opened in the provided editor -
createEditor
: given a working copy, return an editor that can edit the working copy
Note: you will not be asked to createEditor
for a working copy you do not handle
.
Note: section is TBD and work in progress
Stored working copies are a variant of working copies that have a known representation on disk. For example, working copies of notebooks all provide a resource
that denotes to the location of a file on disk. If you have working copies like that, you can provide a stored working copy to benefit from even more capabilities, such as fail save file writing/reading and automated backup handling.
Project Management
- Roadmap
- Iteration Plans
- Development Process
- Issue Tracking
- Build Champion
- Release Process
- Running the Endgame
- Related Projects
Contributing
- How to Contribute
- Submitting Bugs and Suggestions
- Feedback Channels
- Source Code Organization
- Coding Guidelines
- Testing
- Dealing with Test Flakiness
- Contributor License Agreement
- Extension API Guidelines
- Accessibility Guidelines
Documentation