read coding tables in one go if there is sufficient input

[folkertdev]

especially for smaller files this is an improvement

Benchmark 2 (180 runs): target/relwithdebinfo/examples/decompress rs tests/input/bzip2-testfiles/dotnetzip/dancing-color.ps.bz2
  measurement          mean ± σ            min … max           outliers         delta
  wall_time          27.9ms ±  352us    27.2ms … 31.0ms          2 ( 1%)        ⚡-  2.3% ±  0.3%
  peak_rss           7.54MB ± 67.0KB    7.34MB … 7.60MB          0 ( 0%)          +  0.0% ±  0.2%
  cpu_cycles          113M  ± 1.08M      112M  …  125M          11 ( 6%)        ⚡-  2.6% ±  0.3%
  instructions        205M  ±  289       205M  …  205M           0 ( 0%)        ⚡-  3.0% ±  0.0%
  cache_references   2.49M  ± 17.8K     2.47M  … 2.61M          35 (19%)          +  0.9% ±  0.2%
  cache_misses        545K  ± 2.69K      538K  …  561K           4 ( 2%)        💩+  4.0% ±  0.2%
  branch_misses       948K  ± 1.99K      942K  …  957K           7 ( 4%)        ⚡- 15.5% ±  0.1%

(CI benchmarks just show noise...)

[codecov]

Codecov Report

Attention: Patch coverage is 95.83333% with 1 line in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
libbz2-rs-sys/src/decompress.rs	95.83%	1 Missing ⚠️

Files with missing lines	Coverage Δ
libbz2-rs-sys/src/decompress.rs	94.24% <95.83%> (+0.02%)	⬆️

... and 2 files with indirect coverage changes

[folkertdev]

it should be possible to bound this loop (I think it's 20 bits of code at most, so 40bits total?). But it's hard to be confident

[folkertdev]

I fuzzed this for many hours using this size of 20. So it holds at least for files encoded using the current version (which uses 17 bits at most, while earlier bzip2 versions used up to 20 bits).

[bjorn3]

What code is this block based on?

[folkertdev]

this bit of C code

libbzip2-rs/decompress.c

Lines 322 to 335 in d5d181b

	/*--- Now the coding tables ---*/
	for (t = 0; t < nGroups; t++) {
	GET_BITS(BZ_X_CODING_1, curr, 5);
	for (i = 0; i < alphaSize; i++) {
	while (True) {
	if (curr < 1 || curr > 20) RETURN(BZ_DATA_ERROR);
	GET_BIT(BZ_X_CODING_2, uc);
	if (uc == 0) break;
	GET_BIT(BZ_X_CODING_3, uc);
	if (uc == 0) curr++; else curr--;
	}
	s->len[t][i] = curr;
	}
	}

it's all scattered around in our implementation here, but if we assume sufficient input, we can basically copy the C code.