Tests and fixes for a number of prune-flag bugs

[luketpeterson]

No description provided.

[github-actions]

Bench A/B vs base: done

job log · 2026-10-03 05:42:44 UTC

base b4a6abd → head 34e708d, 3 round(s), median of each run averaged; negative is faster

bench	cases	geomean	largest gain	largest loss	>5% faster	>5% slower
shakespeare	7	+1.2%	-4.4% shakespeare/shakespeare_words_insert	+10.0% shakespeare/shakespeare_sentences_get	0	2
cities	5	-1.0%	-4.8% cities/cities_val_count_act	+0.5% cities/cities_val_count	0	0
sparse_keys	96	-0.6%	-9.9% sparse_meet/50	+9.1% sparse_zipper_cursor/200	10	3
binary_keys	77	-2.0%	-7.3% binary_zipper_iter/100	+12.1% binary_descend_until_max_bytes/2000	15	1
superdense_keys	104	-0.4%	-5.0% superdense_val_count_bench/100	+4.4% superdense_val_count_bench/20000	1	0
act_paths	46	-0.0%	-4.5% shape_map_to_act_cata/dense_narrow	+8.6% size_map_to_act_cata/50000	0	4
zipper_head_owned	6	+0.5%	-7.2% zipper_head_owned/owned_head_write_creation_cleanup	+9.8% zipper_head_owned/borrowed_head_read_creation	2	2
product_zipper	4	-1.5%	-2.1% product_zipper/introspecting_pathmap_pathmap	-0.6% product_zipper/generic_pathmap_act	0	0

40 case(s) moved more than 5%

bench	case	base	head	change
binary_keys	binary_descend_until_max_bytes/2000	145.33 µs	162.90 µs	+12.1%
shakespeare	shakespeare/shakespeare_sentences_get	48.43 ms	53.28 ms	+10.0%
sparse_keys	sparse_meet/50	1.00 µs	905 ns	-9.9%
zipper_head_owned	zipper_head_owned/borrowed_head_read_creation	11.03 µs	12.11 µs	+9.8%
sparse_keys	sparse_zipper_cursor/200	2.35 µs	2.56 µs	+9.1%
sparse_keys	sparse_zipper_cursor/100	1.18 µs	1.29 µs	+9.1%
sparse_keys	sparse_meet/100	2.03 µs	1.85 µs	-8.9%
act_paths	size_map_to_act_cata/50000	6.73 ms	7.31 ms	+8.6%
act_paths	size_map_to_act_cata/100000	14.84 ms	16.04 ms	+8.1%
shakespeare	shakespeare/shakespeare_words_val_count	317.13 µs	341.83 µs	+7.8%
zipper_head_owned	zipper_head_owned/owned_head_read_creation	11.94 µs	12.87 µs	+7.8%
sparse_keys	sparse_meet/200	4.87 µs	4.50 µs	-7.7%
sparse_keys	sparse_zipper_cursor/400	5.20 µs	5.60 µs	+7.6%
sparse_keys	binary_drop_head/50	3.59 µs	3.32 µs	-7.5%
binary_keys	binary_zipper_iter/100	2.43 µs	2.26 µs	-7.3%
zipper_head_owned	zipper_head_owned/owned_head_write_creation_cleanup	37.13 µs	34.47 µs	-7.2%
binary_keys	binary_zipper_iter/800	18.34 µs	17.03 µs	-7.1%
binary_keys	binary_insert/100	7.90 µs	7.36 µs	-6.8%
binary_keys	binary_insert/200	16.73 µs	15.59 µs	-6.8%
binary_keys	binary_zipper_iter/1600	36.15 µs	33.74 µs	-6.7%
binary_keys	binary_zipper_iter/400	9.31 µs	8.69 µs	-6.6%
binary_keys	binary_insert/50	3.68 µs	3.44 µs	-6.5%
act_paths	size_map_to_act_cata/25000	3.20 ms	3.41 ms	+6.4%
binary_keys	binary_insert/400	35.70 µs	33.44 µs	-6.3%
sparse_keys	sparse_k_path_iter/800	16.34 µs	15.32 µs	-6.2%
zipper_head_owned	zipper_head_owned/borrowed_head_write_creation_cleanup	35.28 µs	33.12 µs	-6.1%
binary_keys	binary_zipper_iter/200	4.81 µs	4.52 µs	-6.1%
binary_keys	binary_insert/1600	178.00 µs	167.23 µs	-6.0%
act_paths	size_dense_paths_to_act/25000	830.10 µs	879.90 µs	+6.0%
sparse_keys	sparse_k_path_iter/1600	30.79 µs	28.95 µs	-6.0%
binary_keys	binary_insert/800	75.95 µs	71.46 µs	-5.9%
sparse_keys	binary_drop_head/100	7.90 µs	7.44 µs	-5.8%
binary_keys	binary_zipper_iter/50	1.26 µs	1.19 µs	-5.8%
binary_keys	binary_set_val_at_long_replace/160	53 ns	50 ns	-5.8%
sparse_keys	sparse_meet/400	10.95 µs	10.35 µs	-5.5%
binary_keys	binary_descend_until/500	33.59 µs	31.83 µs	-5.2%
sparse_keys	sparse_k_path_iter/400	7.60 µs	7.20 µs	-5.2%
binary_keys	binary_descend_until/250	16.03 µs	15.20 µs	-5.2%
sparse_keys	sparse_k_path_iter/200	3.56 µs	3.38 µs	-5.1%
superdense_keys	superdense_val_count_bench/100	60 ns	57 ns	-5.0%

Full tables per bench are in the job log and the bench-out artifact.

[imlvts]

I definitely like these changes. if there are no perf concerns, I'd merge this.

[luketpeterson]

I definitely like these changes. if there are no perf concerns, I'd merge this.

I spent a lot of time on benchmarking them. In fact, 10x more time ensuring perf parity than on the changes themselves.

Merging.