chore(deps): bump teeny submodule to main (dispatch_values, scan_, subsample, slice_along)

[balbasty]

Agent:claude-fastfields-to-teeny

Closesfastfields/fastfields-lib#106. Prerequisite for #50 (the regulariser tap-table engine). Part of the teeny-refactor umbrella (fastfields-lib#21).

What changed

Two lines:

• external/teeny: dcd591c → 5c46bd4 (teeny main tip, +125 commits).
• .gitmodules: branch = claude/fastfields-teeny-refactor-js42id → branch = main.

The old pin was the tip of teeny's stale claude/fastfields-teeny-refactor-js42id
branch, which has not moved since the refactor began — git ls-remote confirms it
still points at exactly dcd591c. main is what this project tracks now, so
.gitmodules should say so.

What it brings in

Verified against teeny's actual git log dcd591c..main, not a summary:

landing	teeny PR
dispatch_values — product-form runtime→static dispatch	#465
forward+backward scan_ sweep idiom docs	#466
scan_/scan sequential fold along one axis	#254
unfold<Axis>(size, step) sliding/strided window view	#256
subsample<Axes...>(k, starts...) coloured-lattice sugar	#258
peel_zip<Axes...>(a,b[,c]) lock-step iteration	#327
index_select<Axis>(idx) gather	#326
in-place maximum_/minimum_	#325
sqdist/dist wrappers	#324
host-side atomic fetch_add_/sub_	#257
static-unrolled dot/sqdist fast path	#255

Plus a large amount of unrelated hardening (signedness-aware offset decode in
dot/sqdist/peel_zip, the -DTNY_NO_NEGATIVE_INDEX CI variant, the
out-of-place and in-place __restrict__ linear fast paths).

Breaking changes and downstream impact

Exactly two commits in the range are marked breaking:

• 3c4ec01 refactor!: rename take_along to slice_along (teeny#423)
• 8458067 refactor(iterate)!: move scan_/scan's axis<A>{} tag to trailing (teeny#348)

The scan_ tag move cannot break us: scan_ did not exist at dcd591c, so there
are no call sites to break.

For the rename, I grepped every fastfields repo (fastfields-kernels,
-cpu-impl, -cpu-lib, -cuda-impl, -cuda-lib, -lib, and the Python repos)
for take_along:

$ grep -rn "take_along" fastfields-* --include=*.h --include=*.hpp \
--include=*.cpp --include=*.cu --include=*.md
(no matches)

Zero call sites. The only hits anywhere for the #254–259 family of names are
two prose mentions of scan_/subsample/index_select in
TEENIFICATION_REVIEW.md (the document that requested those features), which
are descriptive text, not code. So no downstream fixups ride along in this PR.

Verification

make -C fastfields-cpu-lib clean test, full rebuild, both compilers, before and
after the bump. Check counts are byte-identical across all four runs:

suite	before (clang++/g++)	after (clang++/g++)
test_distance	2352	2352
test_distance_mesh	4622	4622
test_distance_spline	704	704
test_posdef	5092	5092
test_pushpull	326	326
test_reg_field	465	465
test_reg_flow	854	854
test_reg_op	186	186
test_resize	4130	4130
test_restrict	191	191
test_splinc	4577	4577

failures: 0 / PASSED on every line, exit=0 on every run. 23,499 checks total,
no coverage gained or lost — which is the expected shape for a pin bump that adds
new API surface without changing any API we already call.

CUDA is compile/link-only in this environment (no nvcc here, no GPU in CI), so the
cuda-impl/cuda-lib path is unexercised by this PR beyond the fact that it
touches no CUDA source.

---

Generated by Claude Code

[claude]

Agent:claude-fastfields-to-teeny

About the red ❌ on test-via-cpu-lib

That check is not reporting on this diff. The workflow hardcodes ref: main when it checks out the cpu-lib harness, then pins only the nested kernels submodule to the commit under test — so any PR based on the teeny branch is built by the pre-teeny-refactor cpu-lib and cpu-impl, which cannot compile teeny-branch kernels at all.

Running the workflow's own recipe against teeny's tip e579add (which predates this PR) fails identically:

./impl/kernels/distance/spline.h:3:10: fatal error: 'teeny/teeny.h' file not found

Filed as fastfields/fastfields-lib#104, deliberately not fixed here (unrelated concern, own branch).

The gate that was actually run

make -C fastfields-cpu-lib clean test against the dev-tree checkout, on both compilers, before and after:

test_distance 2352 | test_distance_mesh 4622 | test_distance_spline 704
test_posdef 5092 | test_pushpull 326 | test_reg_field 465
test_reg_flow 854 | test_reg_op 186 | test_resize 4130
test_restrict 191 | test_splinc 4577

failures: 0, PASSED on every line, exit=0 — 23,499 checks, identical counts to the pre-change baseline.

---

Generated by Claude Code