perf: read and write a run of adjacent bit fields in one call - #677

Open

Simon-Calbert-Aerospacelab wants to merge 12 commits into

sharksforarms:masterfrom

Simon-Calbert-Aerospacelab:perf/derive-batch-bit-runs

Commits on Sep 2, 2026

perf(derive): read a run of adjacent big-endian fields in one call
Simon-Calbert-Aerospacelab
committed
68d5ac0
perf(derive): write a run of adjacent big-endian fields in one call
Simon-Calbert-Aerospacelab
committed
b858da4
fix(writer): report the same overflow message as an unbatched write
Simon-Calbert-Aerospacelab
committed
75e2065
refactor(derive): decide bit-run eligibility by whitelist
Simon-Calbert-Aerospacelab
committed
0cce2f5
feat(derive): take bool bit fields into a run
Simon-Calbert-Aerospacelab
committed
fe2806f
docs(derive): trim the comments on the bit-run planner
Simon-Calbert-Aerospacelab
committed
b061a44
test: inspect deku errors via Debug so the no-std configs pass
Simon-Calbert-Aerospacelab
committed
6b33f7d
fix(writer): normalise a stale Lsb0 order flag before a batched write
Simon-Calbert-Aerospacelab
committed
268a9d9
fix(derive): fall back per field when a run cannot be written in one go
Simon-Calbert-Aerospacelab
committed
b34e915
fix(writer): drop the order flag when a write leaves no bits pending
Simon-Calbert-Aerospacelab
committed
b166b5f
fix(derive): trace every field a batched run serves
Simon-Calbert-Aerospacelab
committed
8fef3e4
test: pin that a runtime bit_order does not batch
Simon-Calbert-Aerospacelab
committed
0ad31f9