Add a cli option to use the high_bitdepth codepath for 8bit content

[lu-zero]

No description provided.

[coveralls]

Coverage decreased (-0.4%) to 83.388% when pulling 03517f5 on rust-av:8in16 into 245f0d3 on xiph:master.

[barrbrain]

I was able to find desyncs by adapting the fuzzer to run 8-in-16. Here is one small such config:

DecodeTestParameters {
    w: 16,
    h: 31,
    speed: 6,
    q: 0,
    limit: 2,
    bit_depth: 8,
    chroma_sampling: Cs420,
    min_keyint: 1,
    max_keyint: 4,
    switch_frame_interval: 0,
    low_latency: false,
    error_resilient: false,
    bitrate: 11523,
    tile_cols_log2: 1,
    tile_rows_log2: 0,
    still_picture: false,
}

I collected 535 failure examples by enabling 8-in-16 in the encode_decode fuzzer:
crashes.tar.gz

For reference, this was the complete corpus I filtered for the above crashes:
corpus.tar.gz

[lu-zero]

Shall we refresh it and try it again ?

[lu-zero]

Let's refresh this and see how much breaks in the process :)

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (1f4dd91) 88.24% compared to head (c8b380d) 88.49%.

Files	Patch %	Lines
src/bin/common.rs	50.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2746      +/-   ##
==========================================
+ Coverage   88.24%   88.49%   +0.25%     
==========================================
  Files          88       88              
  Lines       28210    28212       +2     
==========================================
+ Hits        24893    24967      +74     
+ Misses       3317     3245      -72     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[barrbrain]

I would somewhat prefer it be a hidden flag or have a "Here there be dragons" warning attached. We can fuzz and run through AWCY to verify.

[lu-zero]

Rebased and now with the cli tests wired in.