Fix dec artifacts w/cropped+smoothed prog DC scans

This commit modifies decompress_smooth_data(), adding missing MCU column
offsets to the prev_block_row and next_block_row indices that are used
for block rows other than the first and last.  Effectively, this
eliminates unexpected visual artifacts when using jpeg_crop_scanline()
along with interblock smoothing while decompressing the DC scan of a
progressive JPEG image.

Based on:
0227d4fb48

Fixes #456
Closes #457

ChangeLog.md

@@ -23,6 +23,10 @@ skipping past the end of an image.
 3. The ARM 64-bit (ARMv8) NEON SIMD extensions can now be built using MinGW
 toolchains targetting ARM64 (AArch64) Windows binaries.
 
+4. Fixed unexpected visual artifacts that occurred when using
+`jpeg_crop_scanline()` and interblock smoothing while decompressing only the
+DC scan of a progressive JPEG image.
+
 
 2.0.5
 =====

jdcoefct.c

@@ -6,7 +6,7 @@
  * libjpeg-turbo Modifications:
  * Copyright 2009 Pierre Ossman <ossman@cendio.se> for Cendio AB
  * Copyright (C) 2010, 2015-2016, D. R. Commander.
- * Copyright (C) 2015, Google, Inc.
+ * Copyright (C) 2015, 2020, Google, Inc.
  * For conditions of distribution and use, see the accompanying README.ijg
  * file.
  *
@@ -495,11 +495,13 @@ decompress_smooth_data(j_decompress_ptr cinfo, JSAMPIMAGE output_buf)
       if (first_row && block_row == 0)
         prev_block_row = buffer_ptr;
       else
-        prev_block_row = buffer[block_row - 1];
+        prev_block_row = buffer[block_row - 1] +
+                         cinfo->master->first_MCU_col[ci];
       if (last_row && block_row == block_rows - 1)
         next_block_row = buffer_ptr;
       else
-        next_block_row = buffer[block_row + 1];
+        next_block_row = buffer[block_row + 1] +
+                         cinfo->master->first_MCU_col[ci];
       /* We fetch the surrounding DC values using a sliding-register approach.
        * Initialize all nine here so as to do the right thing on narrow pics.
        */