* Update paste and drop proposals
Reworks the document paste and drop API proposals. Main highlights:
- Align more with code action api
- Allow a single paste provider to return multiple edits
- Allow resolving applied edits lazily
- Switch from using ids to scoped kinds like used for code actions
* Adding paste context
* Add context
* Update test
Move await from `priority` for drop/paste API proposals
For #179430, #30066
Switching to use `yieldTo` instead of `priority` to let an extension de-rank itself in the list of edits. `priority` was an arbitrary number while `yieldTo` gives more control over how the ranking takes place
For copy paste, I added logic to prefer using the text content if there's both `image/*` and `text/*` content in the clipboard
This however I also incorrectly applied this logic when dropping. In those cases, we instead want to prefer the image data (at least we do in the case of dragging and dropping from VS Code's explorer)
* Add widget to change how content is pasted
For #30066
This adds a widget that lets you change how content is pasted if there are multiple ways it could be pasted
To do this, I've made the post drop widget generic and reused it for pasting too
* Update types
* More code deduplication
* Allow dropping images into notebook to create attachments
Fixes#157577
This allows you to drag and drop image files or image data into a notebook cell to create an attachment
As part of this work, I also updated the paste attachment logic so that we can:
- Create multiple attachments in a single operation
- Create attachments of other mime types besides `image/png`
- Create attachments for images that have spaces in the filename
* Also allow pasting uri lists to create attachments
* Fix indent
* Add id
* Move attachment out of custom metadata, prep for attachment clean up in diff editor
* recover attachments from dirty notebook document
* Allow metadata to be restored when content changed/reverted in nb diff editor
* Compress notebook output streams before rendering
* OOps
* Combine the buffers manually
* Address code review
* oops
* Fixes
* We can have multiple stream mimes in an output
* oops
* cache and cleaner complete, needs debounce
* minor renaming and reformatting
* bugfix for paste into new cell
* cleaning functionality complete
* refer to metadata as copy of current cell's
* check undef before reading from cache
* working state, pending cache restructure
* dots -> brackets
* pre-class refactor
* massive cleaner refactor
* cache typing, closed nb check, workspaceEdit only if metadata is changed
* undefined access fix
* proper debouncer
* get it up to work again
* no need to loop
* cell metadata uri parsing regression
* diagnostic
* Show diagnostics on document open
* transfer cache before file renames
* disable word wrap in notebook diff editor
* Avoid early notebook cell metadata deep clone
* No special case empty cell
* rename
* better naming
* Quick fix for invalid image attachment
* cleanup
* Add code action metadata
Co-authored-by: rebornix <penn.lv@gmail.com>
`currentCell` and `notebookUri` are currently any types, which hides type errors. To fix this and clean up the code, I've extracted a new `getCellFromCellDocument` method
Aaron Munger
Don Jayamanne
Matt Bierner
Peng Lyu
Michael Lively
Michael Lively