Some more nitpicking.

The term "coding unit" may need to be defined in section 3. (It is currently defined in section 6.3.)
In section 3, is "may" synonymous with "can"?
In section 8.6.1, in the description of the order in which edges are deblocked within each treeblock, clarify that the right edge and the bottom edge of each coding unit are excluded from filtering.
In section 8.6.1, replace "coinside" with "which coincide" in two places.
Both changes mentioned above for 8.4.3.1 also apply to 8.6.1.1.

8.6.1.3 seems to be saying that, whenever we find ourselves at the intersection of a horizontal edge and a vertical edge, we only apply the deblocking filter to the horizontal edge. Surely this is incorrect?

8.6.1.4 item 2 (p.152) "of size (2)x(nD)x(nD)," -> "of size (2)x(2)x(nD)x(nD), "
8.6.1.4 item 3 (p.152) ae -> are
8.6.1.4 (p.152): "Max" operation in 8-428 is redundant because bS[ 0 ][ xDk ][ yDm + i ] is not set for i>0 (note: 8-428 is duplicated, "Max" is redundant in both)
8.6.1.4 (p.152-153): dSample, bS_L, and t_C are undefined. (Is "dSample" the same as "dSam" defined near the end of p.154?)

8.6.3 (p.163-164): references to 8.6.2.1, 8.6.2.2, 8.6.2.3, and 8.6.2.4 should be 8.6.3.2, 8.6.3.3, 8.6.3.4, and 8.6.3.5.
8.6.3.2 (p.167): "( sum_k 2*alf_coeff_luma[ i ][ k ] ) >> 2" is equivalent to "( sum_k alf_coeff_luma[ i ][ k ] ) >> 1"

8.6.3.2-8.6.3.5: derivation of filter coefficients in 8.6.3.2 seems fundamentally inconsistent with filtering process in 8.6.3.3-8.6.3.5. For example, the sum in (8-476) always has at most 17 terms (judging by the figure 8-6 and from symmetry considerations, it appears that it is intended to have 17 elements at all times), but the array c_L constructed in 8.6.3.2 may have as many as 80 elements. Incomplete merge? As far as I can tell, sections 8.6.3.3-8.6.3.5 were mostly replaced in WD5 d5, but 8.6.3.2 and corresponding syntax sections (7.3.3.5 and 7.4.3.5) still describe the previous algorithm.

8.6.3.4 (p.169): alf_filter_shape_flag is undefined
8.6.3.4 (p.169): AlfNumCoeffLuma undefined
8.6.3.4 (p.169): references to c_L[fIdx[x][y][i]] should be c_L[fIdx[x][y]][i]
8.6.3.4 (p.169), step 4: "...specified in Table 8-14 and Table 8-15" should be "8-18 and 8-19"

8.6.3.5 (p.170): alf_filter_shape_chroma_flag is undefined
8.6.3.5 (p.171): AlfNumCoeffChroma undefined
8.6.3.5 (p.171): c_C is undefined

Some of the issues mentioned above are still present in the CD:

"RefPicListListA"/"RefPicListListB" instead of "RefPicListLA"/"RefPicListLB" in 3 places

8.5.3.1 (page 134): if InterTUSplitDirection is equal to 2, yB1 and xB2 are undefined

8.5.3.1 (page 135), 8.5.3.2 (page 136), 8.7.1.1 (page 144):

"– The variable log2TrafoWidth1 is set equal to (log2TrafoWidth – 2) * InterTUSplitDirection.
– The variable log2TrafoHeight1 is set equal to (log2TrafoHeight - 2) * (1 - InterTUSplitDirection)."

Should be:

"– The variable log2TrafoWidth1 is set equal to log2TrafoWidth – 2 * InterTUSplitDirection.
– The variable log2TrafoHeight1 is set equal to log2TrafoHeight - 2 * (1 - InterTUSplitDirection)."

8.7.1.4 (p.146) "ae" -> "are"