Repository navigation

perf: skip redundant SQL normalisation and CTE scans during compile #1628

Job	Run time
Lockfile consistency	15s
Classify verification scope	10s
Build wheel (linux-x86_64)	7m 15s
Build wheel (macos-arm64)	9m 13s
Build wheel (macos-x86_64)	22m 26s
Build wheel (windows-x86_64)	9m 48s
Build wheel (linux-aarch64)	7m 38s
Static checks and unit + integration tests	19m 17s
Formatter performance guards	8m 10s
Native compiler on Windows	13m 40s
Compiler golden outputs	10m 1s
Native compiler on Python 3.13	9m 42s
DuckDB E2E (CLI commands)	23m 26s
DuckDB E2E (build core)	9m 41s
DuckDB E2E (integrations)	7m 8s
Dense all-rules compile (1000 models)	6m 51s
PostgreSQL integration and E2E tests	11m 11s
Dense all-rules compile (10000 models)	7m 42s
Dense all-rules compile (3000 models)	7m 35s
Native compiler on Python 3.14	7m 55s
Dense all-rules compile (5000 models)	10m 16s
DuckDB E2E (build incremental)	11m 20s
DuckDB E2E (CLI data)	9m 33s
Fresh-process compile (3000 models)	11m 10s
Fresh-process compile cache (3000 models)	10m 56s
Fresh-process compile (10000 models)	8m 29s
Same-runner compile ratio (dense 3000 models)	15m 14s
Same-runner compile ratio (fresh 3000 models)	14m 26s
Same-runner compile ratio (custom rules CPU vs base 3000 models)	18m 14s
DuckDB E2E (CLI inspection)	11m 33s
Fresh-process compile cache (10000 models)	9m 47s
Fresh-process compile (5000 models)	6m 19s
Same-runner compile ratio (model analysis CPU vs base 3000 models)	17m 40s
Fresh-process compile cache (5000 models)	10m 2s
Same-runner compile ratio (contracts CPU vs base 3000 models)	16m 54s
Rules performance guards	0s
Performance guards	0s
Verify	4s
	6h 21m 1s