Review follow-ups: PR #672 — perf(import): mint a commit chunk's GCNs in one lookup, not one query per card #904

New Issue

2026-09-30T20:52:42Z

gambit-admin commented

2026-09-30 20:52:42 +00:00

Non-blocking findings from the auto code review of PR #672 (perf(import): mint a commit chunk's GCNs in one lookup, not one query per card) — captured for batch cleanup rather than blocking the PR.
Commit 8d238c9 · https://github.com/Gambit-Inc/gambit/pull/672

SHOULD-FIX:

backend/app/routes/imports.py:3601-3602: the comment justifying COMMIT_SLICE_UNITS ("commit mints one GCN (a query) per card, so 50 rows x qty 200 would be one 10k-query slice") is now false — this PR is exactly the change that eliminates the one-GCN-query-per-card cost. The slice-sizing rationale documented there no longer matches reality, which will mislead whoever next tunes COMMIT_SLICE_UNITS or investigates commit latency.

NITS:
none

Filed from a Claude Code session

Non-blocking findings from the auto code review of **PR #672** (perf(import): mint a commit chunk's GCNs in one lookup, not one query per card) — captured for batch cleanup rather than blocking the PR. Commit `8d238c9` · https://github.com/Gambit-Inc/gambit/pull/672 SHOULD-FIX: - `backend/app/routes/imports.py:3601-3602`: the comment justifying `COMMIT_SLICE_UNITS` ("commit mints one GCN (a query) per card, so 50 rows x qty 200 would be one 10k-query slice") is now false — this PR is exactly the change that eliminates the one-GCN-query-per-card cost. The slice-sizing rationale documented there no longer matches reality, which will mislead whoever next tunes `COMMIT_SLICE_UNITS` or investigates commit latency. NITS: none --- Filed from a Claude Code session

gambit-admin added the enhancement label 2026-09-30 20:52:42 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: gambit/gambit#904