Adding WriteZipper::remove_subtrie

[luketpeterson]

Adding WriteZipper::remove_subtrie as this is a more fundamental operation than remove_branches, but remove_branches is useful and correct in its own right.

This new method balances the API, helping define and explain the consistency fixes proposed in #104 Also, in the future this will be the lowest-cost way to just chop a subtrie, although the current implementation doesn't offer much perf benefit vs. remove_branches.

I also updated the lean model to include the new method. However, this shifted the number of opcodes and threatened to invalidate the existing .bin files. So I made a change to the .bin header to include the opcode count, rather than hard-coding the %56 that was in there.

I pondered going further and actually redoing the .bin format entirely, storing opcodes + args instead of the raw fuzz signal, so decoding only happens once before saving a reproduction - but I didn't go that far because it seemed unnecessary and like I'd be over-stepping.

Thoughts, @imlvts?

[github-actions]

Bench A/B vs base: done

job log · 2026-10-09 03:17:28 UTC

base cb4e1e4 → head 0c9e6b7, 3 round(s), median of each run averaged; negative is faster

bench	cases	geomean	largest gain	largest loss	>5% faster	>5% slower
shakespeare	7	+3.4%	-0.3% shakespeare/shakespeare_words_insert	+12.2% shakespeare/shakespeare_sentences_get	0	2
cities	5	-0.8%	-2.2% cities/cities_val_count_act	+0.3% cities/cities_val_count	0	0
sparse_keys	96	-0.7%	-3.9% sparse_zipper_cursor/1600	+1.2% sparse_k_path_iter/50	0	0
binary_keys	77	-1.7%	-12.5% binary_descend_until_max_bytes/1000	+2.2% binary_insert/1600	3	0
superdense_keys	104	-0.8%	-6.3% superdense_val_count_bench/200	+6.7% superdense_drop_bench/100	2	1
act_paths	46	-1.1%	-10.6% shape_map_to_act_cata/random_long	+2.5% size_dense_paths_to_act/25000	2	0
zipper_head_owned	16	+1.7%	-1.5% zipper_head_owned/map_reader_is_val	+11.3% zipper_head_owned/shared_parent_reader_val	0	2
product_zipper	4	+0.1%	-2.0% product_zipper/generic_act_act	+1.8% product_zipper/introspecting_pathmap_pathmap	0	0

12 case(s) moved more than 5%

bench	case	base	head	change
binary_keys	binary_descend_until_max_bytes/1000	69.18 µs	60.50 µs	-12.5%
shakespeare	shakespeare/shakespeare_sentences_get	46.59 ms	52.27 ms	+12.2%
zipper_head_owned	zipper_head_owned/shared_parent_reader_val	1.26 µs	1.41 µs	+11.3%
act_paths	shape_map_to_act_cata/random_long	44.83 ms	40.08 ms	-10.6%
shakespeare	shakespeare/shakespeare_words_val_count	304.97 µs	327.70 µs	+7.5%
binary_keys	binary_set_val_at_short_replace/1	13 ns	12 ns	-6.8%
superdense_keys	superdense_drop_bench/100	50 ns	53 ns	+6.7%
act_paths	shape_map_to_act_cata/small_alphabet	10.76 ms	10.07 ms	-6.4%
binary_keys	binary_descend_until/500	33.86 µs	31.71 µs	-6.3%
superdense_keys	superdense_val_count_bench/200	100 ns	93 ns	-6.3%
superdense_keys	superdense_drop_head/4000	29.05 µs	27.43 µs	-5.6%
zipper_head_owned	zipper_head_owned/borrowed_head_read_value_in_shared_parent	13.52 µs	14.28 µs	+5.6%

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

[Adam-Vandervorst]

I believe this is good iff it's upheld by a lot of laws.
Such as being equivalent to remove_branches ; remove val or restrict(PathMap::new())