Skip to content

Integrate master-next into master#746

Merged
frankmcsherry merged 17 commits into
masterfrom
master-next
May 29, 2026
Merged

Integrate master-next into master#746
frankmcsherry merged 17 commits into
masterfrom
master-next

Conversation

@antiguru
Copy link
Copy Markdown
Member

Integrate master-next into master in preparation of a new release.

frankmcsherry and others added 16 commits April 19, 2026 15:58
* Remove ShutdownDeadmans

* Remove TraceRc

* Relocate TraceBox and remove rc.rs

* Remove redundant trait constraints

* Remove chainless_batcher.rs

* Remove dead and commented code

* Remove Clone constraint

* Convert ::max_value() to ::MAX

* Remove needless borrows and add Rc::clone

* Remove RHH

* Remove Huffman container
* Improve use of join_assign

* Use owned time in join

* Add missing PointStamp::meet_assign
Moves the columnar arrangement / container infrastructure from the
`examples/columnar/columnar_support/` tree into `src/columnar/` as a
public, experimental module. API and internals are explicitly marked as
unstable in the module-level docs; rough edges (`unimplemented!`
`ContainerBytes`, eager-consolidate `leave_dynamic`, single-`U`
`join_function`) are listed up front.

Also generalizes the dynamic-scope helpers: `DynTime<TOuter, T>` is now
parametric (was hardcoded to `Product<u64, PointStamp<u64>>`), and
`leave_dynamic` carries matching bounds plus a `level > 0` assert.

Updates the in-tree consumers: `examples/columnar/main.rs` and
`interactive/examples/ddir_col.rs` switch from path-mounted modules to
`use differential_dataflow::columnar`. `ddir_vec.rs` + `interactive/src/ir.rs`
gain a `RowLike` impl for `SmallVec<A>`, used by the vec-backed ddir
example for its row representation.

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Use meet_assign in ValueHistory::replay

* SCC example for comparative benchmarking
* Introduce Cursor::populate_key

* Update EditList, ValueHistory

* Unify paths in populate_key
* Restore pre-#725 spines.rs and inline EditList::load

Brings back the spines arrangement bake-off (deleted in #724 Spring
cleaning, then RHH-dependent) with three modes: `key` (OrdKeySpine),
`val` (OrdValSpine with Val=()), and `col` (columnar ValSpine via the
columnar module added in #730). All three feed the same Vec-shaped
input collections through one driver loop; `col` repacks via a small
in-dataflow `unary` (`ToRecorded`) that builds `RecordedUpdates`
containers before `arrange_core`.

Bisecting against the example exposed a regression introduced in #725:
EditList::load now delegates to populate_key, which seek_keys + checks
+ rewinds vals on every call. In the merge-join inner loop (join.rs
Ordering::Equal arm), the cursor is already positioned by the upstream
`match trace_key.cmp(&batch_key)` work, so the seek is redundant.
Repeated 1M times in the spines query phase, this added ~3s (+40%
queries time vs pre-#725 baseline).

Restoring EditList::load to its pre-#725 division of labor — assume
the cursor is positioned, walk vals inline — recovers performance.
populate_key and replay_key keep the seek for callers that legitimately
need it (reduce, ValueHistory). The Option-based meet API from #725
stays.

Measurements (1M keys, 1000 size, key mode):
- v0.23.0 baseline: 6.56s queries
- pre-#725 (f4e7550): 7.16s queries
- master HEAD before this commit: 10.12s queries
- this commit: 7.00s queries

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

* Tighten up spines examples

* Extract common target columnar size

* TrieChunker work

* De-penalize col in spiners.rs

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
* Introduce UpdatesView reader type

* Use UpdatesView reader

* Introduce UpdatesOwned for writeable typed allocations

* Re-introduce Updates as Stash-backed fields that can be serialized

* Pivot RecordedUpdates to Stash-backed Updates

* Migrate read uses to UpdatesView

* Move UpdatesOwned::iter() to UpdatesView::iter()

* Rename UpdatesOwned to UpdatesTyped

* impl ContainerBytes for RecordedUpdates

* Fixup ddir_col example
* Remove InternalMerge implementations

* Move VecMerger to its own independent module

* Remove InternalMerge*
* Bespoke columnar MergeBatcher

* Introduce spill traits

* Introduce fetching iteration

* Split large chain links

* Rotate spill files

* Compression before paging

* Tidy columnar_spill example
* V0 explanation infrastructure

* Min-val optimization

* Use invertible maps where possible

* Remove enforced monotonicity for internal demand nodes

* Remove variables for intermediate non-var demand

* Use invertible map to chase through var binding

* Convert more operators to 1:1 demand

* Notes of improved shared responsibility for operators

* Fix Leave bug, and improve things at the same time

* Modernize explain.rs
The chunker was part of the batcher and responsible for transforming input
data into the batcher's chunk format. Hence, the batcher needed to be aware
of its input types, although it would not otherwise use this information.

Drop the `Input` associated type and `push_container` method from the
`Batcher` trait; batchers now accept already-chunked input via
`PushInto<Self::Output>`. The vec `MergeBatcher` loses its `Input` and `C`
(chunker) type parameters, and the columnar `MergeBatcher` loses its internal
`TrieChunker`. Both now expose `PushInto` that inserts a chunk directly as a
chain.

Chunking moves into `arrange_core`, which gains a `Chu: ContainerBuilder`
type parameter so callers supply a chunker that maps the stream's input
container into the batcher's output container. The operator drives the
chunker (push, extract, and a `finish` drain before sealing) where the
batcher previously did.

The `Arrange` trait constrains `Ba::Output = C` and hardcodes
`ContainerChunker<C>` internally, so `.arrange::<Ba, Bu, Tr>()` callsites for
`Vec`-based collections are unchanged. Callers needing a cross-container
chunker (columnar layouts, interactive, spill) drop to `arrange_core`
directly and pass an explicit `ValChunker`.

Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
@antiguru antiguru requested a review from frankmcsherry May 29, 2026 20:49
timely 0.30 bumps its columnar dependency to 0.13, so the workspace
columnar pin moves to 0.13 in lockstep to keep a single shared columnar
crate (avoids `Product: Columnar` failing across crate-version skew).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
@frankmcsherry frankmcsherry merged commit e3bdcd5 into master May 29, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants