perf: decode rewrite, encode zeros optimization, iai-callgrind benchmarks (0.3.1)

Decode rewritten with shrinking-slice pattern (split_first/split_at)
instead of index arithmetic. Decode zeros 14% faster, decode nonzero
flips from 5% behind corncobs to 2% ahead, decode mixed closes to tied.

Encode zeros fast path moved before split_at — now tied with cobs
(was 6% behind). Code size reduced to 821B (was 869B).

Replaced criterion with iai-callgrind for deterministic instruction-count
benchmarks via Valgrind. Zero run-to-run variance. Expanded to 162
benchmarks across 9 sizes.

ucobs now wins or ties 16 of 18 benchmarks vs corncobs and cobs.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

justfile

@@ -4,8 +4,8 @@
 default:
     @just --list
 
-# run ALL quality gates (tests, clippy, fmt check, doc, fuzz, miri)
-test: test-unit test-clippy test-fmt test-doc test-fuzz test-miri
+# run ALL quality gates (tests, clippy, fmt check, doc, fuzz, miri, mutants)
+test: test-unit test-clippy test-fmt test-doc test-fuzz test-miri test-mutants
 
 # quick compile check (no tests)
 check:
@@ -39,10 +39,26 @@ test-fuzz:
 test-miri:
     cargo +nightly miri test -- --skip proptests
 
-# run criterion benchmarks (ucobs vs cobs vs corncobs)
-bench:
+# mutation testing (requires cargo-mutants)
+test-mutants:
+    cargo mutants --no-shuffle
+
+# run iai-callgrind benchmarks (deterministic instruction counts, requires valgrind)
+bench: bench-iai bench-size
+
+# run iai-callgrind benchmarks (ucobs vs cobs vs corncobs)
+bench-iai:
     cargo bench
 
+# measure compiled code size (encode + decode .text bytes)
+bench-size:
+    @cargo build --release 2>/dev/null
+    @echo ""
+    @echo "Code size (release, encode + decode .text):"
+    @nm -S target/release/libucobs.rlib 2>/dev/null \
+        | grep 'T _ZN5ucobs' \
+        | awk '{size=strtonum("0x"$2); name=$4; sub(/.*ucobs[0-9]*/, "", name); sub(/17h.*/, "", name); printf "  ucobs %-7s %4d B\n", name, size}'
+
 # run python cobs benchmarks (requires: pip install cobs)
 bench-python:
     python3 bench_python.py