Inline the fast path for send_bits #1079

Job	Run time
Clippy	21s
cargo-c dynamic library	10s
Clippy (--no-default-features --features="c-allocator")	35s
Build and test (stable, ubuntu-latest, x86_64-unknown-linux-gnu)	53s
Clippy (--no-default-features --features="rust-allocator")	19s
QEMU build & test (stable, ubuntu-latest, s390x-unknown-linux-gnu, s390x-linux-gnu-gcc)	41s
Build and test (msrv, ubuntu-latest, x86_64-unknown-linux-gnu)	1m 3s
QEMU build & test (stable, ubuntu-latest, i686-unknown-linux-gnu, i686-linux-gnu-gcc)	33s
Build and test (beta, ubuntu-latest, x86_64-unknown-linux-gnu)	1m 0s
Smoke-test fuzzing targets (default)	2m 33s
Build and test (stable, macos-latest, x86_64-apple-darwin)	1m 5s
Smoke-test fuzzing targets (c-allocator)	2m 41s
Smoke-test fuzzing targets (stable, macos-14, aarch64-apple-darwin)	2m 37s
Build and test (stable, macos-14, aarch64-apple-darwin)	59s
Smoke-test fuzzing targets (rust-allocator)	2m 31s
Build and test (stable-x86_64-gnu, windows-2022, x86_64-pc-windows-gnu)	2m 6s
vanilla dynamic library	22s
Miri	1m 40s
run flate2 test suite	18s
wasm32	37s
	23m 4s

Provide feedback