1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
// Copyright (c) the JPEG XL Project Authors. All rights reserved.
//
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
use RenderPipelineStageInfo;
use ;
use crate::;
pub use ;
/// Modifies channels in-place.
///
/// For these stages, `process_row_chunk` receives read-write access to each row in each channel.
/// Modifies data and writes it to a new buffer, of possibly different type.
///
/// BORDER_X and BORDER_Y represent the amount of padding required on the input side.
/// SHIFT_X and SHIFT_Y represent the base 2 log of the number of rows/columns produced
/// for each row/column of input.
///
/// For each channel:
/// - the input slice contains 1 + BORDER_Y * 2 slices, each of length
/// xsize + BORDER_X * 2, i.e. covering one input row and up to BORDER pixels of
/// padding on either side.
/// - the output slice contains 1 << SHIFT_Y slices, each of length xsize << SHIFT_X, the
/// corresponding output pixels.
///
/// `process_row_chunk` is passed a pair of q(input, output)` slices.
/// Does not directly modify the current image pixels, but extends the current image with
/// additional data.
///
/// `uses_channel` must always return true, and stages of this type should override
/// `new_size` and `original_data_origin`.
/// `process_row_chunk` will be called with the *new* image coordinates, and will only be called
/// on row chunks outside of the original image data.
/// After stages of this type, no stage can have a non-0 SHIFT_X or SHIFT_Y.
// TODO(veluca): figure out how to modify the interface for concurrent usage.