Fixed jpeg encoder last bytes flushing bug

[br3aker]

Prerequisites

• I have written a descriptive pull-request title
• I have verified that there are no overlapping pull-requests open
• I have verified that I am following the existing coding patterns and practice as demonstrated in the repository. These follow strict Stylecop rules 👮.
• I have provided test coverage for my change (where applicable)

Description

I've made quite a mistake during jpeg encoder optimization for last bytes flushing to the stream:

int valuableBytesCount = (int)Numerics.DivideCeil((uint)this.bitCount, 8);
uint packedBytes = this.accumulatedBits | (uint.MaxValue >> this.bitCount);
this.emitBuffer[--this.emitWriteIndex] = packedBytes;

int endIndex = (this.emitWriteIndex * 4) - 4 + valuableBytesCount;
this.FlushToStream(endIndex);

uint buffer size is 4 for simplicity. Let's say we have 3 bits to pad with 1's for full byte and write to the stream:

bufferSize = 4
emitWriteIndex = 4
bitCount = 3
accumulatedBits = 101|0_0000_...

int valuableBytesCount = 1
uint packedBytes = 101|1_1111_...
--emitWriteIndex = 3
int endIndex = 3 * 4 - 4 + 1 = 12 - 4 + 1 = 9

// bytes indices written to the stream:
15 14 13 12 11 10 9

So extra 6 bytes written on top of 1 byte with actual information. Same setup with fix from this PR:

bufferSize = 4
emitWriteIndex = 4
bitCount = 3
accumulatedBits = 101|0_0000_...

int valuableBytesCount = 1
uint packedBytes = 101|1_1111_...
int endIndex = 4 * 4 - 1 = 16 - 1 = 15

// bytes indices written to the stream:
15

This bug wasn't critical and actually isn't spottable by photoshop or JpegSnoop as extra written bytes do not corrupt resulting jpeg but it's still a problem introduced in one of my PRs (I'm really sorry).

I've tested a couple of images from test folder and got following binary size difference:

Calliphora.jpg:
master: 257.540b
PR: 257.537b

winter420_noninterleaved.jpg: -
master: 248.048b
PR: 248.042b

[codecov]

Codecov Report

Merging #1876 (e9b6b8a) into master (952af37) will not change coverage.
The diff coverage is 100%.

@@          Coverage Diff           @@
##           master   #1876   +/-   ##
======================================
  Coverage      87%     87%           
======================================
  Files         935     935           
  Lines       48939   48939           
  Branches     6086    6086           
======================================
  Hits        42814   42814           
  Misses       5115    5115           
  Partials     1010    1010           

Flag	Coverage Δ
unittests	87% <100%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...mats/Jpeg/Components/Encoder/HuffmanScanEncoder.cs	96% <100%> (ø)

---

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 952af37...e9b6b8a. Read the comment docs.

[JimBobSquarePants]

No need to apologize!! Nobody here caught it.