This method returns the anchor range associated with each edit. The
anchor ranges allow you to determine how each edit interacts with
an existing anchor range that the edit has touched.
As part of #1405, we changed the way we performed undo and redo to
support combining transactions that were not temporally adjacent for
IME purposes.
We introduced a bug with that release that caused divergence
when performing undo: the bug was caused by only changing the visibility
of fragments whose insertion id was contained in the undo operation. However,
an undo operation also affects deletions which we were mistakenly not
considering. Randomized tests caught this but I guess we didn't run enough
of them.
Now, instead of using these versioned offset ranges, we locate the
fragments associated with a transaction using the transaction's
edit ids. To make this possible, buffers now store a new map called
`insertion_slices`, which lets you look up the ranges of insertions
that were affected by a given edit.
Co-authored-by: Antonio Scandurra <antonio@zed.dev>
* Moving the logic from Rope to text::Buffer makes it easier
to keep the Rope in sync with the fragment tree.
* Removing carriage return characters is lossier, but is much
simpler than incrementally maintaining the invariant that
there are no carriage returns followed by newlines. We may
want to do something smarter in the future.
Co-authored-by: Keith Simmons <keith@zed.dev>
This is calculated in `Rope` and uses the `bromberg_sl2` homomorphic
hash function to determine the fingerprint of a single chunk and
compose each chunk fingerprint into a single fingerprint for the entire
rope that is equivalent to hashing all the rope's bytes at once.
* Add a `hard_tabs` setting that causes indentation to be performed
using a tab instead of multiple spaces.
* Change Buffer's indentation-related APIs to return an `IndentSize`
struct with a length and a kind, instead of just a single u32.
* Use hard tabs by default in Go.
Previously, buffer edits represented empty strings as None
variants of an Option. Now, the edit logic just explicitly
checks for empty strings.
Co-authored-by: Keith Simmons <keith@zed.dev>
We forgot to observe the footprint of the reconnecting replica's prior
undos into the local clock. This could cause the replica to generate
edits with a version strictly smaller than what other peers may have
observed. As such, those peers would think they had already seen those
edits and skip them.
These completions don't supply a range that should be overwritten, so
the client needs to infer it via substring matching.
Co-authored-by: Keith Simmons <keith@zed.dev>
This lets us use `next` or `prev` to decide whether to park the cursor
at the first or last filtered item.
Co-Authored-By: Nathan Sobo <nathan@zed.dev>
When handling this messages on the host, wait until the desired
version has been observed before performing the save.
Co-Authored-By: Nathan Sobo <nathan@zed.dev>
When undoing edits performed in the multi-buffer, we also undo subsequent edits that may have occurred outside of the multi-buffer. This commit makes us redo those edits as well.
Co-Authored-By: Antonio Scandurra <me@as-cii.com>
When undoing in the multi-buffer, don't preserve edits that occurred outside the multi-buffer after the edit being undone.
Co-Authored-By: Antonio Scandurra <me@as-cii.com>
This is required because, after joining, we want to be able to refer
to operations that have happened prior to joining, which are not
captured by the state. There is probably a way of reconstructing operations
from the state, but that seems unnecessary and we've already talked about
wanting to have the server store operations rather than state once we start
persisting worktrees.
Also, use env_logger consistently in the tests for each crate.
Only initiallize the logger at all if some RUST_LOG env var is set.
Co-Authored-By: Nathan Sobo <nathan@zed.dev>
Specifically, with this commit:
- We will now refresh the anchor if it escapes the boundaries of the excerpt by
using the `Excerpt::contains` method. This was not the case before, as we were
just checking if the excerpt id and buffer id of the anchors matched the ones
stored on the excerpt.
- We fixed a bug that was causing the anchor to be outside of the excerpt when
resetting it to one of the excerpt's endpoints after we couldn't keep its
position. This would happen because we were using `anchor_at`, which resolved
the anchor to an offset first and then converted it back into an anchor with
the given bias, which is a lossy operation. We now use `Anchor::bias` to
achieve the same goal: note that this could still lead to the anchor escaping
the excerpt's boundary when the bias doesn't match the endpoint's bias, so we
take extra care to avoid that and `min`/`max` the newly-produced anchor with
the other endpoint.
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.
