Address part of #31

[Shnatsel]

Fixes one occurrence of unsoundness described in #31

This change actually makes decoding 5% faster - as measured with hyperfine:

hyperfine -m 25 --warmup=3 'target/release/examples/flate -i enwiki-latest-all-titles-in-ns0.gz -o /dev/null gzip-decode'

On my machine the time goes down from 2.250s to 2.150s

Tests still pass.

[codecov-io]

Codecov Report

Merging #34 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master      #34   +/-   ##
=======================================
  Coverage   91.46%   91.46%           
=======================================
  Files          14       14           
  Lines        1301     1301           
=======================================
  Hits         1190     1190           
  Misses        111      111

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 691454e...5a7bcd1. Read the comment docs.

[sile]

Great! Thank you so much!

[Stargateur]

This could be write:

self.bit_reader
    .as_inner_mut()
    .take(len.into())
    .read_to_end(&mut self.buffer)
    .and_then(|used| {
        if used != len.into() {
            Err(io::Error::new(
                io::ErrorKind::UnexpectedEof,
                "Say something here",
            ))
        } else {
            Ok(())
        }
    })

This should be both safe and fast. I do not really follow you about your claim your PR is more fast, just with one test on your machine.

[Shnatsel]

@Stargateur nice! Care to open a PR for that?

[Stargateur]

@Shnatsel Done