Skip to content

refactor: move the output layer of the Python bindings into sqlbuild-python-output - #1541

Merged
kvlonge merged 1 commit into
mainfrom
refactor/python-output
Nov 17, 2026
Merged

kvlonge merged 1 commit into
mainfrom
refactor/python-output

Conversation

@kvlonge

@kvlonge kvlonge commented Nov 17, 2026

Copy link
Copy Markdown
Contributor

Why

sqlbuild-python is still over the crate size limit. This slice moves the output layer into its own crate.

Changes

  • New crate sqlbuild-python-output, above sqlbuild-python-planner and beside sqlbuild-python-compile (neither depends on the other). It holds build progress, result output and documents, output capture, cost telemetry, compute log storage and the Recall session and stores. Callers use the new paths; no re-exports.
  • Staying in sqlbuild-python: each area's registration, the Recall query command and its rendering (commands layer), event exporting, and the cost report. The last two reach code that stays.
  • main/ entries hold the bodies they forwarded to. Compute log storage registers its own class, so its area has an entry.
  • fensu: the sqlbuild-python cap drops from 66,768 to 54,943 code lines. XSB072/XSB073 budgets are re-keyed with unchanged totals. [dupes] entries cover clusters main already had whose members moved. The native field fallback list follows the moved files.

Code lines across the sqlbuild-python* crates: +685 in total (main/ entries +902, use +405, mod +85, other item code -707).

No DuckShop case: this is a behaviour-neutral Rust move with no new project shape, value or lifecycle.

Verification

  • The _native surface is identical to main's build.
  • A dense 3,000-model compile gives identical JSON and identical files (11,175) to main.
  • Dense timing vs main, 7 alternating runs, same machine, load 3.0 on 8 cores (wall, CPU): cold 9.16→9.17 s, 26.60→26.35 s; warm 0.57→0.58 s, 0.59→0.60 s; one-model edit 1.63→1.66 s, 1.96→2.00 s.
  • Rebuild after a one-line edit in sqlbuild-python (cargo build --profile fast, warm, 3 alternating runs, medians): 22.87→22.56 s wall, 42.3→44.1 s CPU, unchanged.
  • A Recall build records the release version (0.236.3) in every version column.
  • make check-ci passed. make test: 12,312 passed; the 4 failures are publication tests that need Docker, as on main.

…python-output

sqlbuild-python is still over the crate size limit. The output layer
(build progress reporting, build result output and its documents,
output capture and invocation records, cost telemetry, compute log
storage and the Recall session and its stores) moves into a new crate
above sqlbuild-python-planner and beside sqlbuild-python-compile.

The Recall query command and its rendering stay in sqlbuild-python with
the other commands, as do event exporting and the cost report, which
reach code that stays there. Each moved area's registration stays too.

Old paths are deleted and callers updated; there are no re-exports.
main/ entries hold the bodies they used to forward to. The compute log
storage registers its own class, so its area has an entry, and the
Recall session takes the release version from its sqlbuild-python
callers, since CARGO_PKG_VERSION is the fixed workspace version outside
that crate. The
sqlbuild-python cap drops to its new size, XSB072/XSB073 budgets are
re-keyed with unchanged totals, and the native field fallback list
follows the moved files.
@kvlonge
kvlonge enabled auto-merge November 17, 2026 07:42
@kvlonge
kvlonge disabled auto-merge November 17, 2026 07:48
@kvlonge
kvlonge added this pull request to the merge queue Nov 17, 2026
@kvlonge
kvlonge removed this pull request from the merge queue due to a manual request Nov 17, 2026
@kvlonge
kvlonge added this pull request to the merge queue Nov 17, 2026
Merged via the queue into main with commit ca787ea Nov 17, 2026
32 checks passed
@kvlonge
kvlonge deleted the refactor/python-output branch November 17, 2026 09:28
kvlonge added a commit that referenced this pull request Nov 17, 2026
Moves the output layer from the bindings crate into the new
sqlbuild-python-output crate, as the slice did, keeping the
large-project tree's behaviour. The tree's registration bodies match
the slice's base, so the slice's registrations carry over unchanged;
main's statement-report and setup-context imports, which the tree does
not have, are not carried over.

Validation: cargo check --workspace --all-targets passes with no
warnings, the layout matches the slice, and every line the tree added
to a moved file is still present. Nothing is built or tested yet. fensu
reports 349 faults, unchanged from the previous slice.
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.

1 participant