The main reason for this is that we need to include information about
a buffer's UndoMap into its protobuf representation. But it's a bit
complex to correctly incorporate this information into the current
protobuf representation.
If we want to continue reusing `Buffer::apply_remote_edit` for
incorporating the historical operations, we need to either make
that method capable of incorporating already-undone edits, or
serialize the UndoMap into undo *operations*, so that we can apply
these undo operations after the fact when deserializing. But this is
not trivial, because an UndoOperation requires information about
the full offset ranges that were undone.
Use downcasting instead for accessing worktree-specific state of the Files.
This will allow us to introduce a WorktreeId type and use that everywhere
for identifying worktrees. It also just removes some unnecessary coupling
between the language crate and the worktree.
In the case of the new Next.js app, the app will follow a redirect
from 'zed.dev/rpc' to the subdomain where the rust service is hosted.
Until then, the app will connect directly to zed.dev/rpc.
This allows us to respect the bias on anchors we use to create excerpt headers so that they always remain above any content inserted at the start of an excerpt.
Because multi-buffers can contain the same content multiple times, we need to use anchors to track our desired insertion and removal locations when moving lines. This is because deleting a line in order to move it might end up deleting *multiple* lines.
Since these methods take buffer points instead of display points, this adjusts
the logic for retrieving the visible selections, so that they are initially returned
in terms of buffer points.
This is still a bit weird because we can't remove the last line of an excerpt but we still move it into another buffer. There also seem to be issues with undo.
We added clipping of points against the buffer when excerpt headers were in the buffer, but now that they're just blocks, I think we can avoid the potential to panic in these methods by going back to not clipping.
Previously, cursors at column 0 had to be explicitly moved when those lines
were autoindented. This behavior was lost when we moved selections from
the buffer to the editor. Now, with the right bias, we get this behavior automatically.
Co-Authored-By: Nathan Sobo <nathan@zed.dev>
Antonio Scandurra