Repository navigation
DDIR: mechanical tidies (stale code and docs, one-off results, console.html) - #1011
Merged
frankmcsherry merged 10 commits intoOct 24, 2026
Merged
Conversation
- ir::Id was unreferenced. - lib.rs and explain referred to the retired `folded` module. - backend/corgi.rs said CorgiBackend mirrors VecBackend, which is deleted. - projection.rs pointed at examples/self_opt, which isn't merged. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
The section described a `vec` row backend that is deleted, and didn't mention shapes, projection or diagnostics. It now walks the four parts (language, analyses, runtime, host), and says where an expression language and a backend plug in. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
- hash_u64 and gen_field_seeded move beside the `random:` generated source in server.rs, their one library user; a test and an example name them as interactive::server::. - load_program, a read_to_string wrapper, is inlined at its four uses. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
…r drops generated.txt dropped the source by hand after its consumers, which now fails with "no program": the server reclaims it when the last importer drops. The session and the server examples README say so. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
Its tests query result edges by one-field keys, where scc.ddp keys bare integers and counts its result, so an include_str! would need edits throughout. The comment now says so. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
The release job now runs authored/check.py on the fixture graph at four workers, once each, so the 41 hand-written plans can't drift from the generated plans as the language and runtime change. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
The browser console was last changed in August, isn't run or tested, and has likely drifted from the protocol. A future console would read the diag.* collections and the current protocol, so it would be rebuilt rather than kept up to date. The server README no longer points to it. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
TimelyDataflow#1009 replaced the WebSocket diagnostics with the diag.* collections, so the server no longer reads DDIR_DIAG_PORT. The two-session demo, the txn bench and the multi-worker test still set it. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
Recorded results with no lineage go; anything here should be
reproducible from the scripts and their method docs.
- observations/sf1-memory-2026-10-21.{csv,json}: one profiling run.
- baselines/RESULTS.md and results-mini03-sf1.csv: one SQL comparison run.
- MEMORY.md keeps how profile_memory.py measures and how to run it; its
table of one run's observations, and the notes on that run's plans and
binary, go.
- UPDATES.md drops a note about one revision's SF1 run; METHOD.md no
longer points to RESULTS.md.
Kept: the bindings (sf1-memory.json and the rest), tiny.json, METHOD.md,
and every script.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
examples/import_bench.rs and examples/triangles_bench.rs were one-off Rust benches over the library, built by CI but run by hand. Frank: "Those two Rust programs can probably go." With import_bench gone, hash_u64's only user is gen_field_seeded, so it is private to server.rs. Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]> Claude-Session: https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Mechanical tidies from the read-only map of
interactive/. Nothing here changes behaviour. 10 commits, 28 files, +75/−1,478.Deleted
ir::Id, which nothing referenced.server/console.html, 687 lines, untouched since August and untested. It's retired for now; a console can be rebuilt on newer mechanisms, such as thediag.*traces.examples/import_bench.rsandexamples/triangles_bench.rs.Leftover
DDIR_DIAG_PORTsettings, in the two-session demo,bench_txn.pyandserver/tests/multiworker.rs.One-off recorded results, removed because anything worth knowing should be reproducible:
bench/ldbc/observations/sf1-memory-2026-10-21.{csv,json};baselines/RESULTS.mdandresults-mini03-sf1.csv;MEMORY.md,UPDATES.mdandMETHOD.md.Method docs, bindings,
tiny.jsonand every script stay.Fixed
lib.rson the retiredfoldedmodule;explain/mod.rs:412;backend/corgi.rsonVecBackend;projection.rs:1-3, which pointed at an unmerged example.examples/server/sessions/generated.txt: a generated source is now reclaimed when its last importer drops, so the manualdropis gone.Rewritten: the README's Architecture section.
Backend.Moved:
lib.rsis now just its module list.hash_u64andgen_field_seededmove to the server's generated-source code, where they're used.load_programis replaced byread_to_stringat its four call sites.Kept, with a reason:
tests/explain.rs'sSCC_ROWcopy now says why it differs fromexamples/programs/scc.ddp: its tests key edges by one-field tuples.CI: the release job now runs
bench/ldbc/authored/check.py --fixture. It checks that the 41 authored plans give the generated plans' answers.Checks:
cargo test -p interactive -p ddir-server: 199 passed.check.pystep.check.py: 41/41.examples/aoc2023/run.sh: 33/33.server/demo/two_sessions.pypasses.🤖 Generated with Claude Code
https://claude.ai/code/session_019XsakwFGwtESwZ1gSYwG4o