Roll dst_palette into the pixel_swizzler struct
This commit will soon be followed by a rollback, but it is committed
anyway so that we can refer to these numbers in the git log.
name old speed new speed delta
wuffs_gif_decode_1k_bw/clang5 286MB/s ± 1% 251MB/s ± 0% -12.11% (p=0.000 n=10+9)
wuffs_gif_decode_1k_color_full_init/clang5 96.0MB/s ± 0% 88.9MB/s ± 0% -7.35% (p=0.000 n=10+10)
wuffs_gif_decode_1k_color_part_init/clang5 119MB/s ± 0% 108MB/s ± 0% -9.06% (p=0.000 n=10+8)
wuffs_gif_decode_10k_bgra/clang5 469MB/s ± 1% 438MB/s ± 0% -6.50% (p=0.000 n=10+8)
wuffs_gif_decode_10k_indexed/clang5 125MB/s ± 1% 117MB/s ± 1% -6.78% (p=0.000 n=9+9)
wuffs_gif_decode_20k/clang5 153MB/s ± 0% 139MB/s ± 0% -9.40% (p=0.000 n=9+9)
wuffs_gif_decode_100k_artificial/clang5 324MB/s ± 0% 304MB/s ± 0% -6.22% (p=0.000 n=10+9)
wuffs_gif_decode_100k_realistic/clang5 139MB/s ± 1% 125MB/s ± 1% -9.69% (p=0.000 n=9+10)
wuffs_gif_decode_1000k_full_init/clang5 141MB/s ± 0% 126MB/s ± 0% -10.40% (p=0.000 n=10+10)
wuffs_gif_decode_1000k_part_init/clang5 141MB/s ± 1% 126MB/s ± 0% -10.50% (p=0.000 n=10+8)
wuffs_gif_decode_anim_screencap/clang5 655MB/s ± 0% 617MB/s ± 1% -5.91% (p=0.000 n=9+9)
wuffs_gif_decode_1k_bw/gcc7 299MB/s ± 0% 295MB/s ± 0% -1.18% (p=0.000 n=10+8)
wuffs_gif_decode_1k_color_full_init/gcc7 85.2MB/s ± 1% 88.2MB/s ± 0% +3.58% (p=0.000 n=10+10)
wuffs_gif_decode_1k_color_part_init/gcc7 104MB/s ± 0% 103MB/s ± 0% -0.39% (p=0.001 n=9+10)
wuffs_gif_decode_10k_bgra/gcc7 349MB/s ± 2% 352MB/s ± 0% +0.77% (p=0.000 n=9+10)
wuffs_gif_decode_10k_indexed/gcc7 106MB/s ± 0% 106MB/s ± 2% ~ (p=0.762 n=8+10)
wuffs_gif_decode_20k/gcc7 133MB/s ± 0% 133MB/s ± 0% ~ (p=0.258 n=9+9)
wuffs_gif_decode_100k_artificial/gcc7 302MB/s ± 0% 301MB/s ± 1% -0.41% (p=0.035 n=10+10)
wuffs_gif_decode_100k_realistic/gcc7 118MB/s ± 1% 118MB/s ± 1% ~ (p=0.095 n=9+10)
wuffs_gif_decode_1000k_full_init/gcc7 120MB/s ± 1% 119MB/s ± 0% -0.34% (p=0.007 n=10+10)
wuffs_gif_decode_1000k_part_init/gcc7 119MB/s ± 1% 119MB/s ± 1% ~ (p=0.138 n=10+10)
wuffs_gif_decode_anim_screencap/gcc7 640MB/s ± 0% 640MB/s ± 0% ~ (p=0.905 n=9+10)
diff --git a/internal/cgen/base/image-public.h b/internal/cgen/base/image-public.h
index 39c62ed..52c7bad 100644
--- a/internal/cgen/base/image-public.h
+++ b/internal/cgen/base/image-public.h
@@ -1287,16 +1287,16 @@
// --------
// TODO: should the func type take restrict pointers?
-typedef uint64_t (*wuffs_base__pixel_swizzler__func)(
- wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
- wuffs_base__slice_u8 src);
+typedef uint64_t (*wuffs_base__pixel_swizzler__func)(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
+ wuffs_base__slice_u8 src);
typedef struct {
// Do not access the private_impl's fields directly. There is no API/ABI
// compatibility or safety guarantee if you do so.
struct {
wuffs_base__pixel_swizzler__func func;
+ uint8_t scratch1024[1024];
} private_impl;
#ifdef __cplusplus
@@ -1306,7 +1306,6 @@
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend);
inline uint64_t swizzle_interleaved(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) const;
#endif // __cplusplus
@@ -1336,7 +1335,6 @@
wuffs_base__pixel_swizzler__swizzle_interleaved(
const wuffs_base__pixel_swizzler* p,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src);
#ifdef __cplusplus
@@ -1354,10 +1352,8 @@
uint64_t //
wuffs_base__pixel_swizzler::swizzle_interleaved(
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) const {
- return wuffs_base__pixel_swizzler__swizzle_interleaved(this, dst, dst_palette,
- src);
+ return wuffs_base__pixel_swizzler__swizzle_interleaved(this, dst, src);
}
#endif // __cplusplus
diff --git a/internal/cgen/base/pixconv-submodule.c b/internal/cgen/base/pixconv-submodule.c
index e041e5c..2479a19 100644
--- a/internal/cgen/base/pixconv-submodule.c
+++ b/internal/cgen/base/pixconv-submodule.c
@@ -446,15 +446,15 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__copy_1_1(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__copy_1_1(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
return wuffs_base__slice_u8__copy_from_slice(dst, src);
}
static uint64_t //
-wuffs_base__pixel_swizzler__copy_3_3(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__copy_3_3(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t src_len3 = src.len / 3;
@@ -466,8 +466,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__copy_4_4(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__copy_4_4(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -481,8 +481,8 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__bgr_565__bgr(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__bgr_565__bgr(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t src_len3 = src.len / 3;
@@ -510,8 +510,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t src_len4 = src.len / 4;
@@ -539,8 +539,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t src_len4 = src.len / 4;
@@ -592,8 +592,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__bgr_565__y(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__bgr_565__y(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t len = dst_len2 < src.len ? dst_len2 : src.len;
@@ -618,13 +618,9 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__bgr_565__index__src(
- wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
- wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
+wuffs_base__pixel_swizzler__bgr_565__index__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
+ wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t len = dst_len2 < src.len ? dst_len2 : src.len;
uint8_t* d = dst.ptr;
@@ -636,16 +632,16 @@
while (n >= loop_unroll_count) {
wuffs_base__store_u16le__no_bounds_check(
d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
wuffs_base__store_u16le__no_bounds_check(
d + (1 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[1] * 4)));
+ scratch1024 + ((size_t)s[1] * 4)));
wuffs_base__store_u16le__no_bounds_check(
d + (2 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[2] * 4)));
+ scratch1024 + ((size_t)s[2] * 4)));
wuffs_base__store_u16le__no_bounds_check(
d + (3 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[3] * 4)));
+ scratch1024 + ((size_t)s[3] * 4)));
s += loop_unroll_count * 1;
d += loop_unroll_count * 2;
@@ -655,7 +651,7 @@
while (n >= 1) {
wuffs_base__store_u16le__no_bounds_check(
d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
s += 1 * 1;
d += 1 * 2;
@@ -667,12 +663,9 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len2 = dst.len / 2;
size_t len = dst_len2 < src.len ? dst_len2 : src.len;
uint8_t* d = dst.ptr;
@@ -682,7 +675,7 @@
// TODO: unroll.
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)s0);
@@ -699,10 +692,9 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(
- wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
- wuffs_base__slice_u8 src) {
+wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
+ wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t src_len4 = src.len / 4;
size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4;
@@ -728,8 +720,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t src_len4 = src.len / 4;
@@ -775,8 +767,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -806,8 +798,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -834,8 +826,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -863,12 +855,9 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxx__index__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len3 = dst.len / 3;
size_t len = dst_len3 < src.len ? dst_len3 : src.len;
uint8_t* d = dst.ptr;
@@ -887,16 +876,16 @@
while (n > loop_unroll_count) {
wuffs_base__store_u32le__no_bounds_check(
d + (0 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (1 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[1] * 4)));
+ scratch1024 + ((size_t)s[1] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (2 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[2] * 4)));
+ scratch1024 + ((size_t)s[2] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (3 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[3] * 4)));
+ scratch1024 + ((size_t)s[3] * 4)));
s += loop_unroll_count * 1;
d += loop_unroll_count * 3;
@@ -904,7 +893,7 @@
}
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);
@@ -918,12 +907,9 @@
static uint64_t //
wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len3 = dst.len / 3;
size_t len = dst_len3 < src.len ? dst_len3 : src.len;
uint8_t* d = dst.ptr;
@@ -933,22 +919,22 @@
const size_t loop_unroll_count = 4;
while (n >= loop_unroll_count) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);
}
- uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[1] * 4));
if (s1) {
wuffs_base__store_u24le__no_bounds_check(d + (1 * 3), s1);
}
- uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[2] * 4));
if (s2) {
wuffs_base__store_u24le__no_bounds_check(d + (2 * 3), s2);
}
- uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[3] * 4));
if (s3) {
wuffs_base__store_u24le__no_bounds_check(d + (3 * 3), s3);
@@ -960,7 +946,7 @@
}
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);
@@ -975,8 +961,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__xxx__y(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxx__y(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t len = dst_len3 < src.len ? dst_len3 : src.len;
@@ -1003,12 +989,9 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxxx__index__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len4 = dst.len / 4;
size_t len = dst_len4 < src.len ? dst_len4 : src.len;
uint8_t* d = dst.ptr;
@@ -1020,16 +1003,16 @@
while (n >= loop_unroll_count) {
wuffs_base__store_u32le__no_bounds_check(
d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (1 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[1] * 4)));
+ scratch1024 + ((size_t)s[1] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (2 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[2] * 4)));
+ scratch1024 + ((size_t)s[2] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (3 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[3] * 4)));
+ scratch1024 + ((size_t)s[3] * 4)));
s += loop_unroll_count * 1;
d += loop_unroll_count * 4;
@@ -1039,7 +1022,7 @@
while (n >= 1) {
wuffs_base__store_u32le__no_bounds_check(
d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
s += 1 * 1;
d += 1 * 4;
@@ -1051,12 +1034,9 @@
static uint64_t //
wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len4 = dst.len / 4;
size_t len = dst_len4 < src.len ? dst_len4 : src.len;
uint8_t* d = dst.ptr;
@@ -1066,22 +1046,22 @@
const size_t loop_unroll_count = 4;
while (n >= loop_unroll_count) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);
}
- uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[1] * 4));
if (s1) {
wuffs_base__store_u32le__no_bounds_check(d + (1 * 4), s1);
}
- uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[2] * 4));
if (s2) {
wuffs_base__store_u32le__no_bounds_check(d + (2 * 4), s2);
}
- uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[3] * 4));
if (s3) {
wuffs_base__store_u32le__no_bounds_check(d + (3 * 4), s3);
@@ -1093,7 +1073,7 @@
}
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);
@@ -1108,8 +1088,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__xxxx__xxx(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxxx__xxx(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len3 = src.len / 3;
@@ -1134,8 +1114,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__xxxx__y(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxxx__y(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t len = dst_len4 < src.len ? dst_len4 : src.len;
@@ -1162,7 +1142,6 @@
static wuffs_base__pixel_swizzler__func //
wuffs_base__pixel_swizzler__prepare__y(wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
switch (dst_format.repr) {
@@ -1190,15 +1169,15 @@
wuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(
wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
+ wuffs_base__slice_u8 scratch =
+ wuffs_base__make_slice_u8(&(p->private_impl.scratch1024[0]), 1024);
switch (dst_format.repr) {
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_NONPREMUL:
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_PREMUL:
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:
- if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=
- 1024) {
+ if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {
return NULL;
}
switch (blend) {
@@ -1208,7 +1187,7 @@
return NULL;
case WUFFS_BASE__PIXEL_FORMAT__BGR_565:
- if (wuffs_base__pixel_swizzler__squash_bgr_565_888(dst_palette,
+ if (wuffs_base__pixel_swizzler__squash_bgr_565_888(scratch,
src_palette) != 1024) {
return NULL;
}
@@ -1221,8 +1200,7 @@
return NULL;
case WUFFS_BASE__PIXEL_FORMAT__BGR:
- if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=
- 1024) {
+ if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {
return NULL;
}
switch (blend) {
@@ -1236,8 +1214,7 @@
case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:
case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:
case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:
- if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=
- 1024) {
+ if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {
return NULL;
}
switch (blend) {
@@ -1249,8 +1226,8 @@
return NULL;
case WUFFS_BASE__PIXEL_FORMAT__RGB:
- if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(dst_palette,
- src_palette) != 1024) {
+ if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(scratch, src_palette) !=
+ 1024) {
return NULL;
}
switch (blend) {
@@ -1264,8 +1241,8 @@
case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:
case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:
case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:
- if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(dst_palette,
- src_palette) != 1024) {
+ if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(scratch, src_palette) !=
+ 1024) {
return NULL;
}
switch (blend) {
@@ -1282,7 +1259,6 @@
static wuffs_base__pixel_swizzler__func //
wuffs_base__pixel_swizzler__prepare__bgr(wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
switch (dst_format.repr) {
@@ -1313,7 +1289,6 @@
wuffs_base__pixel_swizzler__prepare__bgra_nonpremul(
wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
switch (dst_format.repr) {
@@ -1388,39 +1363,46 @@
switch (src_format.repr) {
case WUFFS_BASE__PIXEL_FORMAT__Y:
- func = wuffs_base__pixel_swizzler__prepare__y(p, dst_format, dst_palette,
- src_palette, blend);
+ func = wuffs_base__pixel_swizzler__prepare__y(p, dst_format, src_palette,
+ blend);
break;
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:
func = wuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(
- p, dst_format, dst_palette, src_palette, blend);
+ p, dst_format, src_palette, blend);
break;
case WUFFS_BASE__PIXEL_FORMAT__BGR:
- func = wuffs_base__pixel_swizzler__prepare__bgr(
- p, dst_format, dst_palette, src_palette, blend);
+ func = wuffs_base__pixel_swizzler__prepare__bgr(p, dst_format,
+ src_palette, blend);
break;
case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:
func = wuffs_base__pixel_swizzler__prepare__bgra_nonpremul(
- p, dst_format, dst_palette, src_palette, blend);
+ p, dst_format, src_palette, blend);
break;
}
p->private_impl.func = func;
- return wuffs_base__make_status(
- func ? NULL : wuffs_base__error__unsupported_pixel_swizzler_option);
+ if (!func) {
+ return wuffs_base__make_status(
+ wuffs_base__error__unsupported_pixel_swizzler_option);
+ }
+ if (dst_palette.len == 1024) {
+ const uint8_t* scratch1024 = &(p->private_impl.scratch1024[0]);
+ memcpy(dst_palette.ptr, scratch1024, 1024);
+ }
+ return wuffs_base__make_status(NULL);
}
WUFFS_BASE__MAYBE_STATIC uint64_t //
wuffs_base__pixel_swizzler__swizzle_interleaved(
const wuffs_base__pixel_swizzler* p,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
if (p && p->private_impl.func) {
- return (*p->private_impl.func)(dst, dst_palette, src);
+ const uint8_t* scratch1024 = &(p->private_impl.scratch1024[0]);
+ return (*p->private_impl.func)(scratch1024, dst, src);
}
return 0;
}
diff --git a/internal/cgen/data.go b/internal/cgen/data.go
index 6be82bf..7e96e44 100644
--- a/internal/cgen/data.go
+++ b/internal/cgen/data.go
@@ -135,48 +135,50 @@
"// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__squash_bgr_565_888(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t len4 = (dst.len < src.len ? dst.len : src.len) / 4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n\n size_t n = len4;\n while (n--) {\n uint32_t argb = wuffs_base__load_u32le__no_bounds_check(s);\n uint32_t b5 = 0x1F & (argb >> (8 - 5));\n uint32_t g6 = 0x3F & (argb >> (16 - 6));\n uint32_t r5 = 0x1F & (argb >> (24 - 5));\n uint32_t alpha = argb & 0xFF000000;\n wuffs_base__store_u32le__no_bounds_check(\n d, alpha | (r5 << 11) | (g6 << 5) | (b5 << 0));\n s += 4;\n d += 4;\n }\n return len4 * 4;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__swap_rgbx_bgrx(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t len4 = (dst.len < src.len ? dst.len : src.len) / 4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n\n size_t n = len4;\n while (n--) {\n " +
" uint8_t b0 = s[0];\n uint8_t b1 = s[1];\n uint8_t b2 = s[2];\n uint8_t b3 = s[3];\n d[0] = b2;\n d[1] = b1;\n d[2] = b0;\n d[3] = b3;\n s += 4;\n d += 4;\n }\n return len4 * 4;\n}\n\n" +
"" +
- "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__copy_1_1(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n return wuffs_base__slice_u8__copy_from_slice(dst, src);\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__copy_3_3(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t src_len3 = src.len / 3;\n size_t len = dst_len3 < src_len3 ? dst_len3 : src_len3;\n if (len > 0) {\n memmove(dst.ptr, src.ptr, len * 3);\n }\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__copy_4_4(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 <" +
- " src_len4 ? dst_len4 : src_len4;\n if (len > 0) {\n memmove(dst.ptr, src.ptr, len * 4);\n }\n return len;\n}\n\n" +
+ "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__copy_1_1(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n return wuffs_base__slice_u8__copy_from_slice(dst, src);\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__copy_3_3(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t src_len3 = src.len / 3;\n size_t len = dst_len3 < src_len3 ? dst_len3 : src_len3;\n if (len > 0) {\n memmove(dst.ptr, src.ptr, len * 3);\n }\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__copy_4_4(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 < src_len4 ? dst_le" +
+ "n4 : src_len4;\n if (len > 0) {\n memmove(dst.ptr, src.ptr, len * 4);\n }\n return len;\n}\n\n" +
"" +
- "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__bgr(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t src_len3 = src.len / 3;\n size_t len = dst_len2 < src_len3 ? dst_len2 : src_len3;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t b5 = s[0] >> 3;\n uint32_t g6 = s[1] >> 2;\n uint32_t r5 = s[2] >> 3;\n uint32_t rgb_565 = (r5 << 11) | (g6 << 5) | (b5 << 0);\n wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)rgb_565);\n\n s += 1 * 3;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t src_len4 = src.len / 4;\n size_t len = " +
- "dst_len2 < src_len4 ? dst_len2 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n wuffs_base__store_u16le__no_bounds_check(\n d + (0 * 2),\n wuffs_base__color_u32_argb_premul__as__color_u16_rgb_565(\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(s + (0 * 4)))));\n\n s += 1 * 4;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len2 < src_len4 ? dst_len2 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n // Convert from 8-bit color to 16-bit color.\n uint32_t sa = 0x101 * ((uint32_t)s[3]);\n uint32_t sr = 0" +
- "x101 * ((uint32_t)s[2]);\n uint32_t sg = 0x101 * ((uint32_t)s[1]);\n uint32_t sb = 0x101 * ((uint32_t)s[0]);\n\n // Convert from 565 color to 16-bit color.\n uint32_t old_rgb_565 = wuffs_base__load_u16le__no_bounds_check(d + (0 * 2));\n uint32_t old_r5 = 0x1F & (old_rgb_565 >> 11);\n uint32_t dr = (0x8421 * old_r5) >> 4;\n uint32_t old_g6 = 0x3F & (old_rgb_565 >> 5);\n uint32_t dg = (0x1041 * old_g6) >> 2;\n uint32_t old_b5 = 0x1F & (old_rgb_565 >> 0);\n uint32_t db = (0x8421 * old_b5) >> 4;\n\n // Calculate the inverse of the src-alpha: how much of the dst to keep.\n uint32_t ia = 0xFFFF - sa;\n\n // Composite src (nonpremul) over dst (premul).\n dr = ((sr * sa) + (dr * ia)) / 0xFFFF;\n dg = ((sg * sa) + (dg * ia)) / 0xFFFF;\n db = ((sb * sa) + (db * ia)) / 0xFFFF;\n\n // Convert from 16-bit color to 565 color and combine the components.\n uint32_t new_r5 = 0x1F & (dr >> 11);\n uint32_t new_g6 = 0x3F & (dg >> 10);\n uint32_t new_b5 = 0x1F & (db >> 11);\n uint32_t new_rgb" +
- "_565 = (new_r5 << 11) | (new_g6 << 5) | (new_b5 << 0);\n wuffs_base__store_u16le__no_bounds_check(d + (0 * 2),\n (uint16_t)new_rgb_565);\n\n s += 1 * 4;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__y(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t len = dst_len2 < src.len ? dst_len2 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t y5 = s[0] >> 3;\n uint32_t y6 = s[0] >> 2;\n uint32_t rgb_565 = (y5 << 11) | (y6 << 5) | (y5 << 0);\n wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)rgb_565);\n\n s += 1 * 1;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__index__src(\n wuffs_base__slice_u8" +
- " dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n if (dst_palette.len != 1024) {\n return 0;\n }\n size_t dst_len2 = dst.len / 2;\n size_t len = dst_len2 < src.len ? dst_len2 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count) {\n wuffs_base__store_u16le__no_bounds_check(\n d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[0] * 4)));\n wuffs_base__store_u16le__no_bounds_check(\n d + (1 * 2), wuffs_base__load_u16le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[1] * 4)));\n wuffs_base__store_u16le__no_bounds_check(\n d + (2 * 2), wuffs_base__load_u16le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[2] * 4)));\n wuffs_base__store_u16le__no_bounds_check(\n d + (3 * 2), wuffs_base__load_u16le__no_bounds_check(\n dst_palette.ptr" +
- " + ((size_t)s[3] * 4)));\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 2;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n wuffs_base__store_u16le__no_bounds_check(\n d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[0] * 4)));\n\n s += 1 * 1;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n if (dst_palette.len != 1024) {\n return 0;\n }\n size_t dst_len2 = dst.len / 2;\n size_t len = dst_len2 < src.len ? dst_len2 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u16le__no_bounds" +
- "_check(d + (0 * 2), (uint16_t)s0);\n }\n\n s += 1 * 1;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\n" +
+ "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__bgr(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t src_len3 = src.len / 3;\n size_t len = dst_len2 < src_len3 ? dst_len2 : src_len3;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t b5 = s[0] >> 3;\n uint32_t g6 = s[1] >> 2;\n uint32_t r5 = s[2] >> 3;\n uint32_t rgb_565 = (r5 << 11) | (g6 << 5) | (b5 << 0);\n wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)rgb_565);\n\n s += 1 * 3;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len2 < s" +
+ "rc_len4 ? dst_len2 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n wuffs_base__store_u16le__no_bounds_check(\n d + (0 * 2),\n wuffs_base__color_u32_argb_premul__as__color_u16_rgb_565(\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(s + (0 * 4)))));\n\n s += 1 * 4;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len2 < src_len4 ? dst_len2 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n // Convert from 8-bit color to 16-bit color.\n uint32_t sa = 0x101 * ((uint32_t)s[3]);\n uint32_t sr = 0x101 * ((uint32_t)" +
+ "s[2]);\n uint32_t sg = 0x101 * ((uint32_t)s[1]);\n uint32_t sb = 0x101 * ((uint32_t)s[0]);\n\n // Convert from 565 color to 16-bit color.\n uint32_t old_rgb_565 = wuffs_base__load_u16le__no_bounds_check(d + (0 * 2));\n uint32_t old_r5 = 0x1F & (old_rgb_565 >> 11);\n uint32_t dr = (0x8421 * old_r5) >> 4;\n uint32_t old_g6 = 0x3F & (old_rgb_565 >> 5);\n uint32_t dg = (0x1041 * old_g6) >> 2;\n uint32_t old_b5 = 0x1F & (old_rgb_565 >> 0);\n uint32_t db = (0x8421 * old_b5) >> 4;\n\n // Calculate the inverse of the src-alpha: how much of the dst to keep.\n uint32_t ia = 0xFFFF - sa;\n\n // Composite src (nonpremul) over dst (premul).\n dr = ((sr * sa) + (dr * ia)) / 0xFFFF;\n dg = ((sg * sa) + (dg * ia)) / 0xFFFF;\n db = ((sb * sa) + (db * ia)) / 0xFFFF;\n\n // Convert from 16-bit color to 565 color and combine the components.\n uint32_t new_r5 = 0x1F & (dr >> 11);\n uint32_t new_g6 = 0x3F & (dg >> 10);\n uint32_t new_b5 = 0x1F & (db >> 11);\n uint32_t new_rgb_565 = (new_r5 << " +
+ "11) | (new_g6 << 5) | (new_b5 << 0);\n wuffs_base__store_u16le__no_bounds_check(d + (0 * 2),\n (uint16_t)new_rgb_565);\n\n s += 1 * 4;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__y(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t len = dst_len2 < src.len ? dst_len2 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t y5 = s[0] >> 3;\n uint32_t y6 = s[0] >> 2;\n uint32_t rgb_565 = (y5 << 11) | (y6 << 5) | (y5 << 0);\n wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)rgb_565);\n\n s += 1 * 1;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__index__src(const uint8_t* scratch1024,\n " +
+ " wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t len = dst_len2 < src.len ? dst_len2 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count) {\n wuffs_base__store_u16le__no_bounds_check(\n d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(\n scratch1024 + ((size_t)s[0] * 4)));\n wuffs_base__store_u16le__no_bounds_check(\n d + (1 * 2), wuffs_base__load_u16le__no_bounds_check(\n scratch1024 + ((size_t)s[1] * 4)));\n wuffs_base__store_u16le__no_bounds_check(\n d + (2 * 2), wuffs_base__load_u16le__no_bounds_check(\n scratch1024 + ((size_t)s[2] * 4)));\n wuffs_base__store_u16le__no_bounds_check(\n d + (3 * 2), wuffs_base__load_u16le__no_bounds_check(\n scratch1024 + ((size_t)s[" +
+ "3] * 4)));\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 2;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n wuffs_base__store_u16le__no_bounds_check(\n d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(\n scratch1024 + ((size_t)s[0] * 4)));\n\n s += 1 * 1;\n d += 1 * 2;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len2 = dst.len / 2;\n size_t len = dst_len2 < src.len ? dst_len2 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)s0);\n }\n\n s += 1 * 1;\n d += 1 * 2;\n n" +
+ " -= 1;\n }\n\n return len;\n}\n\n" +
"" +
- "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t s0 =\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(s + (0 * 4)));\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n\n s += 1 * 4;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4;\n uint8_t* d = d" +
- "st.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n // Convert from 8-bit color to 16-bit color.\n uint32_t sa = 0x101 * ((uint32_t)s[3]);\n uint32_t sr = 0x101 * ((uint32_t)s[2]);\n uint32_t sg = 0x101 * ((uint32_t)s[1]);\n uint32_t sb = 0x101 * ((uint32_t)s[0]);\n uint32_t dr = 0x101 * ((uint32_t)d[2]);\n uint32_t dg = 0x101 * ((uint32_t)d[1]);\n uint32_t db = 0x101 * ((uint32_t)d[0]);\n\n // Calculate the inverse of the src-alpha: how much of the dst to keep.\n uint32_t ia = 0xFFFF - sa;\n\n // Composite src (nonpremul) over dst (premul).\n dr = ((sr * sa) + (dr * ia)) / 0xFFFF;\n dg = ((sg * sa) + (dg * ia)) / 0xFFFF;\n db = ((sb * sa) + (db * ia)) / 0xFFFF;\n\n // Convert from 16-bit color to 8-bit color.\n d[0] = (uint8_t)(db >> 8);\n d[1] = (uint8_t)(dg >> 8);\n d[2] = (uint8_t)(dr >> 8);\n\n s += 1 * 4;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\n" +
+ "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t s0 =\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(s + (0 * 4)));\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n\n s += 1 * 4;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t src_len4 = src.le" +
+ "n / 4;\n size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n // Convert from 8-bit color to 16-bit color.\n uint32_t sa = 0x101 * ((uint32_t)s[3]);\n uint32_t sr = 0x101 * ((uint32_t)s[2]);\n uint32_t sg = 0x101 * ((uint32_t)s[1]);\n uint32_t sb = 0x101 * ((uint32_t)s[0]);\n uint32_t dr = 0x101 * ((uint32_t)d[2]);\n uint32_t dg = 0x101 * ((uint32_t)d[1]);\n uint32_t db = 0x101 * ((uint32_t)d[0]);\n\n // Calculate the inverse of the src-alpha: how much of the dst to keep.\n uint32_t ia = 0xFFFF - sa;\n\n // Composite src (nonpremul) over dst (premul).\n dr = ((sr * sa) + (dr * ia)) / 0xFFFF;\n dg = ((sg * sa) + (dg * ia)) / 0xFFFF;\n db = ((sb * sa) + (db * ia)) / 0xFFFF;\n\n // Convert from 16-bit color to 8-bit color.\n d[0] = (uint8_t)(db >> 8);\n d[1] = (uint8_t)(dg >> 8);\n d[2] = (uint8_t)(dr >> 8);\n\n s += 1 * 4;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;" +
+ "\n}\n\n" +
"" +
- "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t d0 = wuffs_base__load_u32le__no_bounds_check(d + (0 * 4));\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(s + (0 * 4));\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4),\n wuffs_base__composite_nonpremul_nonpremul_u32_axxx(d0, s0));\n\n s += 1 * 4;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\n" +
+ "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t d0 = wuffs_base__load_u32le__no_bounds_check(d + (0 * 4));\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(s + (0 * 4));\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4),\n wuffs_base__composite_nonpremul_nonpremul_u32_axxx(d0, s0));\n\n s += 1 * 4;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\n" +
"" +
- "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(s + (0 * 4));\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4),\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(s0));\n\n s += 1 * 4;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4;\n ui" +
- "nt8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t d0 = wuffs_base__load_u32le__no_bounds_check(d + (0 * 4));\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(s + (0 * 4));\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), wuffs_base__composite_premul_nonpremul_u32_axxx(d0, s0));\n\n s += 1 * 4;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\n" +
+ "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(s + (0 * 4));\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4),\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(s0));\n\n s += 1 * 4;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len4 = src.len / 4;\n size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4;\n uint8_t* d = d" +
+ "st.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint32_t d0 = wuffs_base__load_u32le__no_bounds_check(d + (0 * 4));\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(s + (0 * 4));\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), wuffs_base__composite_premul_nonpremul_u32_axxx(d0, s0));\n\n s += 1 * 4;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\n" +
"" +
- "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n if (dst_palette.len != 1024) {\n return 0;\n }\n size_t dst_len3 = dst.len / 3;\n size_t len = dst_len3 < src.len ? dst_len3 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n // The comparison in the while condition is \">\", not \">=\", because with\n // \">=\", the last 4-byte store could write past the end of the dst slice.\n //\n // Each 4-byte store writes one too many bytes, but a subsequent store\n // will overwrite that with the correct byte. There is always another\n // store, whether a 4-byte store in this loop or a 1-byte store in the\n // next loop.\n while (n > loop_unroll_count) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 3), wuffs_base__load_u32le__no_bounds_c" +
- "heck(\n dst_palette.ptr + ((size_t)s[0] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (1 * 3), wuffs_base__load_u32le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[1] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (2 * 3), wuffs_base__load_u32le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[2] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (3 * 3), wuffs_base__load_u32le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[3] * 4)));\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 3;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[0] * 4));\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n\n s += 1 * 1;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__" +
- "pixel_swizzler__xxx__index_binary_alpha__src_over(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n if (dst_palette.len != 1024) {\n return 0;\n }\n size_t dst_len3 = dst.len / 3;\n size_t len = dst_len3 < src.len ? dst_len3 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n }\n uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[1] * 4));\n if (s1) {\n wuffs_base__store_u24le__no_bounds_check(d + (1 * 3), s1);\n }\n uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n " +
- " ((size_t)s[2] * 4));\n if (s2) {\n wuffs_base__store_u24le__no_bounds_check(d + (2 * 3), s2);\n }\n uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[3] * 4));\n if (s3) {\n wuffs_base__store_u24le__no_bounds_check(d + (3 * 3), s3);\n }\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 3;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n }\n\n s += 1 * 1;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxx__y(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 =" +
- " dst.len / 3;\n size_t len = dst_len3 < src.len ? dst_len3 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n uint8_t s0 = s[0];\n d[0] = s0;\n d[1] = s0;\n d[2] = s0;\n\n s += 1 * 1;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\n" +
+ "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxx__index__src(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t len = dst_len3 < src.len ? dst_len3 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n // The comparison in the while condition is \">\", not \">=\", because with\n // \">=\", the last 4-byte store could write past the end of the dst slice.\n //\n // Each 4-byte store writes one too many bytes, but a subsequent store\n // will overwrite that with the correct byte. There is always another\n // store, whether a 4-byte store in this loop or a 1-byte store in the\n // next loop.\n while (n > loop_unroll_count) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 3), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[0" +
+ "] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (1 * 3), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[1] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (2 * 3), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[2] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (3 * 3), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[3] * 4)));\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 3;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[0] * 4));\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n\n s += 1 * 1;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over(\n const uint8_t* scratch" +
+ "1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t len = dst_len3 < src.len ? dst_len3 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n }\n uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[1] * 4));\n if (s1) {\n wuffs_base__store_u24le__no_bounds_check(d + (1 * 3), s1);\n }\n uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[2] * 4));\n if (s2) {\n wuffs_base__store_u24le__no_bounds_check(d + (2 * 3), s2);\n }\n uint32_t s3 = " +
+ "wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[3] * 4));\n if (s3) {\n wuffs_base__store_u24le__no_bounds_check(d + (3 * 3), s3);\n }\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 3;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);\n }\n\n s += 1 * 1;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxx__y(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len3 = dst.len / 3;\n size_t len = dst_len3 < src.len ? dst_len3 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n w" +
+ "hile (n >= 1) {\n uint8_t s0 = s[0];\n d[0] = s0;\n d[1] = s0;\n d[2] = s0;\n\n s += 1 * 1;\n d += 1 * 3;\n n -= 1;\n }\n\n return len;\n}\n\n" +
"" +
- "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n if (dst_palette.len != 1024) {\n return 0;\n }\n size_t dst_len4 = dst.len / 4;\n size_t len = dst_len4 < src.len ? dst_len4 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[0] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (1 * 4), wuffs_base__load_u32le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[1] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (2 * 4), wuffs_base__load_u32le__no_bounds_check(\n dst_pale" +
- "tte.ptr + ((size_t)s[2] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (3 * 4), wuffs_base__load_u32le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[3] * 4)));\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 4;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(\n dst_palette.ptr + ((size_t)s[0] * 4)));\n\n s += 1 * 1;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n if (dst_palette.len != 1024) {\n return 0;\n }\n size_t dst_len4 = dst.len / 4;\n size_t len = dst_len4 < src.len ? dst_len4 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count)" +
- " {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);\n }\n uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[1] * 4));\n if (s1) {\n wuffs_base__store_u32le__no_bounds_check(d + (1 * 4), s1);\n }\n uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[2] * 4));\n if (s2) {\n wuffs_base__store_u32le__no_bounds_check(d + (2 * 4), s2);\n }\n uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[3] * 4));\n if (s3) {\n wuffs_base__store_u32le__no_bounds_check(d + (3 * 4), s3);\n }\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count *" +
- " 4;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +\n ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);\n }\n\n s += 1 * 1;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__xxx(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len3 = src.len / 3;\n size_t len = dst_len4 < src_len3 ? dst_len4 : src_len3;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4),\n 0xFF000000 | wuffs_base__load_u24le__no_bounds_check(s + (0 * 3)));\n\n s += 1 * 3;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n" +
- "\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__y(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t len = dst_len4 < src.len ? dst_len4 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), 0xFF000000 | (0x010101 * (uint32_t)s[0]));\n\n s += 1 * 1;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\n" +
+ "// --------\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__index__src(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t len = dst_len4 < src.len ? dst_len4 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[0] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (1 * 4), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[1] * 4)));\n wuffs_base__store_u32le__no_bounds_check(\n d + (2 * 4), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[2] * 4)));\n wuffs_base__store_u32le__no_bound" +
+ "s_check(\n d + (3 * 4), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[3] * 4)));\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 4;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(\n scratch1024 + ((size_t)s[0] * 4)));\n\n s += 1 * 1;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over(\n const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t len = dst_len4 < src.len ? dst_len4 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n const size_t loop_unroll_count = 4;\n\n while (n >= loop_unroll_count) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n (" +
+ "(size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);\n }\n uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[1] * 4));\n if (s1) {\n wuffs_base__store_u32le__no_bounds_check(d + (1 * 4), s1);\n }\n uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[2] * 4));\n if (s2) {\n wuffs_base__store_u32le__no_bounds_check(d + (2 * 4), s2);\n }\n uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n ((size_t)s[3] * 4));\n if (s3) {\n wuffs_base__store_u32le__no_bounds_check(d + (3 * 4), s3);\n }\n\n s += loop_unroll_count * 1;\n d += loop_unroll_count * 4;\n n -= loop_unroll_count;\n }\n\n while (n >= 1) {\n uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +\n " +
+ " ((size_t)s[0] * 4));\n if (s0) {\n wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);\n }\n\n s += 1 * 1;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__xxx(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t src_len3 = src.len / 3;\n size_t len = dst_len4 < src_len3 ? dst_len4 : src_len3;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4),\n 0xFF000000 | wuffs_base__load_u24le__no_bounds_check(s + (0 * 3)));\n\n s += 1 * 3;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__xxxx__y(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n " +
+ " wuffs_base__slice_u8 src) {\n size_t dst_len4 = dst.len / 4;\n size_t len = dst_len4 < src.len ? dst_len4 : src.len;\n uint8_t* d = dst.ptr;\n uint8_t* s = src.ptr;\n size_t n = len;\n\n // TODO: unroll.\n\n while (n >= 1) {\n wuffs_base__store_u32le__no_bounds_check(\n d + (0 * 4), 0xFF000000 | (0x010101 * (uint32_t)s[0]));\n\n s += 1 * 1;\n d += 1 * 4;\n n -= 1;\n }\n\n return len;\n}\n\n" +
"" +
- "// --------\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__y(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n return wuffs_base__pixel_swizzler__bgr_565__y;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n return wuffs_base__pixel_swizzler__xxx__y;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n case WU" +
- "FFS_BASE__PIXEL_FORMAT__RGBX:\n return wuffs_base__pixel_swizzler__xxxx__y;\n }\n return NULL;\n}\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(\n wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:\n if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=\n 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__copy_1_1;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n if (wuffs_base__pixel_swizzler__squash_bgr_565_888(dst_palette,\n " +
- " src_palette) != 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__bgr_565__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=\n 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_pa" +
- "lette) !=\n 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(dst_palette,\n src_palette) != 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n if (wuffs_base__pixel_swizzler__" +
- "swap_rgbx_bgrx(dst_palette,\n src_palette) != 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over;\n }\n return NULL;\n }\n return NULL;\n}\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__bgr(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n return wuffs_base__pixel_swizzler__bgr_565__bgr;\n\n case WUFFS_BASE__PIXEL_FORMAT__B" +
- "GR:\n return wuffs_base__pixel_swizzler__copy_3_3;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n return wuffs_base__pixel_swizzler__xxxx__xxx;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n // TODO.\n break;\n }\n return NULL;\n}\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__bgra_nonpremul(\n wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_b" +
- "ase__pixel_swizzler__bgr_565__bgra_nonpremul__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__copy_4_4;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swiz" +
- "zler__bgra_premul__bgra_nonpremul__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n // TODO.\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n // TODO.\n break;\n }\n return NULL;\n}\n\n" +
+ "// --------\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__y(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n return wuffs_base__pixel_swizzler__bgr_565__y;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n return wuffs_base__pixel_swizzler__xxx__y;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n return wuffs_base__pixel_swizzler__xx" +
+ "xx__y;\n }\n return NULL;\n}\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(\n wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n wuffs_base__slice_u8 scratch =\n wuffs_base__make_slice_u8(&(p->private_impl.scratch1024[0]), 1024);\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:\n if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__copy_1_1;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n if (wuffs_base__pixel_swizzler__squash_bgr_565_888(scratch,\n src_p" +
+ "alette) != 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__bgr_565__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {\n return NULL;\n " +
+ " }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(scratch, src_palette) !=\n 1024) {\n return NULL;\n }\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(scratch, src_palette) !=\n 1024) {\n return NULL;\n }\n swit" +
+ "ch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__xxxx__index__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over;\n }\n return NULL;\n }\n return NULL;\n}\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__bgr(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n return wuffs_base__pixel_swizzler__bgr_565__bgr;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n return wuffs_base__pixel_swizzler__copy_3_3;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n case WUFFS_B" +
+ "ASE__PIXEL_FORMAT__BGRX:\n return wuffs_base__pixel_swizzler__xxxx__xxx;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n // TODO.\n break;\n }\n return NULL;\n}\n\nstatic wuffs_base__pixel_swizzler__func //\nwuffs_base__pixel_swizzler__prepare__bgra_nonpremul(\n wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n switch (dst_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n swit" +
+ "ch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__copy_4_4;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n switch (blend) {\n case WUFFS_BASE__PIXEL_BLEND__SRC:\n return wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src;\n case WUFFS_BASE__PIXEL_BLEND__SRC_OVER:\n return wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over;\n }\n return NULL;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n case " +
+ "WUFFS_BASE__PIXEL_FORMAT__BGRX:\n // TODO.\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n // TODO.\n break;\n }\n return NULL;\n}\n\n" +
"" +
- "// --------\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__status //\nwuffs_base__pixel_swizzler__prepare(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n if (!p) {\n return wuffs_base__make_status(wuffs_base__error__bad_receiver);\n }\n\n // TODO: support many more formats.\n\n wuffs_base__pixel_swizzler__func func = NULL;\n\n switch (src_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__Y:\n func = wuffs_base__pixel_swizzler__prepare__y(p, dst_format, dst_palette,\n src_palette, blend);\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:\n func = wuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(\n " +
- " p, dst_format, dst_palette, src_palette, blend);\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n func = wuffs_base__pixel_swizzler__prepare__bgr(\n p, dst_format, dst_palette, src_palette, blend);\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n func = wuffs_base__pixel_swizzler__prepare__bgra_nonpremul(\n p, dst_format, dst_palette, src_palette, blend);\n break;\n }\n\n p->private_impl.func = func;\n return wuffs_base__make_status(\n func ? NULL : wuffs_base__error__unsupported_pixel_swizzler_option);\n}\n\nWUFFS_BASE__MAYBE_STATIC uint64_t //\nwuffs_base__pixel_swizzler__swizzle_interleaved(\n const wuffs_base__pixel_swizzler* p,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) {\n if (p && p->private_impl.func) {\n return (*p->private_impl.func)(dst, dst_palette, src);\n }\n return 0;\n}\n" +
+ "// --------\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__status //\nwuffs_base__pixel_swizzler__prepare(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n if (!p) {\n return wuffs_base__make_status(wuffs_base__error__bad_receiver);\n }\n\n // TODO: support many more formats.\n\n wuffs_base__pixel_swizzler__func func = NULL;\n\n switch (src_format.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__Y:\n func = wuffs_base__pixel_swizzler__prepare__y(p, dst_format, src_palette,\n blend);\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:\n func = wuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(\n p, dst_" +
+ "format, src_palette, blend);\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n func = wuffs_base__pixel_swizzler__prepare__bgr(p, dst_format,\n src_palette, blend);\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n func = wuffs_base__pixel_swizzler__prepare__bgra_nonpremul(\n p, dst_format, src_palette, blend);\n break;\n }\n\n p->private_impl.func = func;\n if (!func) {\n return wuffs_base__make_status(\n wuffs_base__error__unsupported_pixel_swizzler_option);\n }\n if (dst_palette.len == 1024) {\n const uint8_t* scratch1024 = &(p->private_impl.scratch1024[0]);\n memcpy(dst_palette.ptr, scratch1024, 1024);\n }\n return wuffs_base__make_status(NULL);\n}\n\nWUFFS_BASE__MAYBE_STATIC uint64_t //\nwuffs_base__pixel_swizzler__swizzle_interleaved(\n const wuffs_base__pixel_swizzler* p,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src) {\n if (p && p->private_impl.func) {\n const uint8_t* scratch1024 =" +
+ " &(p->private_impl.scratch1024[0]);\n return (*p->private_impl.func)(scratch1024, dst, src);\n }\n return 0;\n}\n" +
""
const baseFundamentalPrivateH = "" +
@@ -318,10 +320,10 @@
"" +
"// --------\n\n// wuffs_base__pixel_palette__closest_element returns the index of the palette\n// element that minimizes the sum of squared differences of the four ARGB\n// channels, working in premultiplied alpha. Ties favor the smaller index.\n//\n// The palette_slice.len may equal (N*4), for N less than 256, which means that\n// only the first N palette elements are considered. It returns 0 when N is 0.\n//\n// Applying this function on a per-pixel basis will not produce whole-of-image\n// dithering.\nWUFFS_BASE__MAYBE_STATIC uint8_t //\nwuffs_base__pixel_palette__closest_element(\n wuffs_base__slice_u8 palette_slice,\n wuffs_base__pixel_format palette_format,\n wuffs_base__color_u32_argb_premul c);\n\n" +
"" +
- "// --------\n\n// TODO: should the func type take restrict pointers?\ntypedef uint64_t (*wuffs_base__pixel_swizzler__func)(\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src);\n\ntypedef struct {\n // Do not access the private_impl's fields directly. There is no API/ABI\n // compatibility or safety guarantee if you do so.\n struct {\n wuffs_base__pixel_swizzler__func func;\n } private_impl;\n\n#ifdef __cplusplus\n inline wuffs_base__status prepare(wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend);\n inline uint64_t swizzle_interleaved(wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src) const;\n#endi" +
- "f // __cplusplus\n\n} wuffs_base__pixel_swizzler;\n\n// wuffs_base__pixel_swizzler__prepare readies the pixel swizzler so that its\n// other methods may be called.\n//\n// For modular builds that divide the base module into sub-modules, using this\n// function requires the WUFFS_CONFIG__MODULE__BASE__PIXCONV sub-module, not\n// just WUFFS_CONFIG__MODULE__BASE__CORE.\nWUFFS_BASE__MAYBE_STATIC wuffs_base__status //\nwuffs_base__pixel_swizzler__prepare(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend);\n\n// wuffs_base__pixel_swizzler__swizzle_interleaved converts pixels from a\n// source format to a destination format.\n//\n// For modular builds that divide the base module into sub-modules, using " +
- "this\n// function requires the WUFFS_CONFIG__MODULE__BASE__PIXCONV sub-module, not\n// just WUFFS_CONFIG__MODULE__BASE__CORE.\nWUFFS_BASE__MAYBE_STATIC uint64_t //\nwuffs_base__pixel_swizzler__swizzle_interleaved(\n const wuffs_base__pixel_swizzler* p,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__slice_u8 src);\n\n#ifdef __cplusplus\n\ninline wuffs_base__status //\nwuffs_base__pixel_swizzler::prepare(wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n return wuffs_base__pixel_swizzler__prepare(this, dst_format, dst_palette,\n src_format, src_palette, blend);\n}\n\nuint64_t //\nwuffs_base__pixel_swizzler::swizzle_interleaved(\n wuffs_base__slice_u8 dst,\n wuffs_base__sl" +
- "ice_u8 dst_palette,\n wuffs_base__slice_u8 src) const {\n return wuffs_base__pixel_swizzler__swizzle_interleaved(this, dst, dst_palette,\n src);\n}\n\n#endif // __cplusplus\n" +
+ "// --------\n\n// TODO: should the func type take restrict pointers?\ntypedef uint64_t (*wuffs_base__pixel_swizzler__func)(const uint8_t* scratch1024,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src);\n\ntypedef struct {\n // Do not access the private_impl's fields directly. There is no API/ABI\n // compatibility or safety guarantee if you do so.\n struct {\n wuffs_base__pixel_swizzler__func func;\n uint8_t scratch1024[1024];\n } private_impl;\n\n#ifdef __cplusplus\n inline wuffs_base__status prepare(wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend);\n inline uint64_t swizzle_interleaved(wuffs_base__slice_u8 dst,\n " +
+ " wuffs_base__slice_u8 src) const;\n#endif // __cplusplus\n\n} wuffs_base__pixel_swizzler;\n\n// wuffs_base__pixel_swizzler__prepare readies the pixel swizzler so that its\n// other methods may be called.\n//\n// For modular builds that divide the base module into sub-modules, using this\n// function requires the WUFFS_CONFIG__MODULE__BASE__PIXCONV sub-module, not\n// just WUFFS_CONFIG__MODULE__BASE__CORE.\nWUFFS_BASE__MAYBE_STATIC wuffs_base__status //\nwuffs_base__pixel_swizzler__prepare(wuffs_base__pixel_swizzler* p,\n wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend);\n\n// wuffs_base__pixel_swizzler__swizzle_interleaved converts pixels from a\n// source format to a destination format.\n//\n// For modular builds that d" +
+ "ivide the base module into sub-modules, using this\n// function requires the WUFFS_CONFIG__MODULE__BASE__PIXCONV sub-module, not\n// just WUFFS_CONFIG__MODULE__BASE__CORE.\nWUFFS_BASE__MAYBE_STATIC uint64_t //\nwuffs_base__pixel_swizzler__swizzle_interleaved(\n const wuffs_base__pixel_swizzler* p,\n wuffs_base__slice_u8 dst,\n wuffs_base__slice_u8 src);\n\n#ifdef __cplusplus\n\ninline wuffs_base__status //\nwuffs_base__pixel_swizzler::prepare(wuffs_base__pixel_format dst_format,\n wuffs_base__slice_u8 dst_palette,\n wuffs_base__pixel_format src_format,\n wuffs_base__slice_u8 src_palette,\n wuffs_base__pixel_blend blend) {\n return wuffs_base__pixel_swizzler__prepare(this, dst_format, dst_palette,\n src_format, src_palette, blend);\n}\n\nuint64_t //\nwuffs_base__pixel_swizzler::swizzle_interleaved(\n wuffs_base__slice_u8 dst,\n wuffs_" +
+ "base__slice_u8 src) const {\n return wuffs_base__pixel_swizzler__swizzle_interleaved(this, dst, src);\n}\n\n#endif // __cplusplus\n" +
""
const baseIOPrivateH = "" +
diff --git a/lang/builtin/builtin.go b/lang/builtin/builtin.go
index cab8fd8..d2e2d35 100644
--- a/lang/builtin/builtin.go
+++ b/lang/builtin/builtin.go
@@ -381,7 +381,7 @@
"dst_pixfmt: pixel_format, dst_palette: slice u8," +
"src_pixfmt: pixel_format, src_palette: slice u8, blend: pixel_blend) status",
"pixel_swizzler.swizzle_interleaved!(" +
- "dst: slice u8, dst_palette: slice u8, src: slice u8) u64",
+ "dst: slice u8, src: slice u8) u64",
}
var Interfaces = []string{
diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c
index fecdc96..4562596 100644
--- a/release/c/wuffs-unsupported-snapshot.c
+++ b/release/c/wuffs-unsupported-snapshot.c
@@ -3773,16 +3773,16 @@
// --------
// TODO: should the func type take restrict pointers?
-typedef uint64_t (*wuffs_base__pixel_swizzler__func)(
- wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
- wuffs_base__slice_u8 src);
+typedef uint64_t (*wuffs_base__pixel_swizzler__func)(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
+ wuffs_base__slice_u8 src);
typedef struct {
// Do not access the private_impl's fields directly. There is no API/ABI
// compatibility or safety guarantee if you do so.
struct {
wuffs_base__pixel_swizzler__func func;
+ uint8_t scratch1024[1024];
} private_impl;
#ifdef __cplusplus
@@ -3792,7 +3792,6 @@
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend);
inline uint64_t swizzle_interleaved(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) const;
#endif // __cplusplus
@@ -3822,7 +3821,6 @@
wuffs_base__pixel_swizzler__swizzle_interleaved(
const wuffs_base__pixel_swizzler* p,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src);
#ifdef __cplusplus
@@ -3840,10 +3838,8 @@
uint64_t //
wuffs_base__pixel_swizzler::swizzle_interleaved(
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) const {
- return wuffs_base__pixel_swizzler__swizzle_interleaved(this, dst, dst_palette,
- src);
+ return wuffs_base__pixel_swizzler__swizzle_interleaved(this, dst, src);
}
#endif // __cplusplus
@@ -6193,7 +6189,6 @@
struct {
uint8_t f_compressed[4096];
uint8_t f_palettes[2][1024];
- uint8_t f_dst_palette[1024];
wuffs_lzw__decoder f_lzw;
struct {
@@ -10759,15 +10754,15 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__copy_1_1(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__copy_1_1(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
return wuffs_base__slice_u8__copy_from_slice(dst, src);
}
static uint64_t //
-wuffs_base__pixel_swizzler__copy_3_3(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__copy_3_3(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t src_len3 = src.len / 3;
@@ -10779,8 +10774,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__copy_4_4(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__copy_4_4(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -10794,8 +10789,8 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__bgr_565__bgr(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__bgr_565__bgr(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t src_len3 = src.len / 3;
@@ -10823,8 +10818,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t src_len4 = src.len / 4;
@@ -10852,8 +10847,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t src_len4 = src.len / 4;
@@ -10905,8 +10900,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__bgr_565__y(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__bgr_565__y(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t len = dst_len2 < src.len ? dst_len2 : src.len;
@@ -10931,13 +10926,9 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__bgr_565__index__src(
- wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
- wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
+wuffs_base__pixel_swizzler__bgr_565__index__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
+ wuffs_base__slice_u8 src) {
size_t dst_len2 = dst.len / 2;
size_t len = dst_len2 < src.len ? dst_len2 : src.len;
uint8_t* d = dst.ptr;
@@ -10949,16 +10940,16 @@
while (n >= loop_unroll_count) {
wuffs_base__store_u16le__no_bounds_check(
d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
wuffs_base__store_u16le__no_bounds_check(
d + (1 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[1] * 4)));
+ scratch1024 + ((size_t)s[1] * 4)));
wuffs_base__store_u16le__no_bounds_check(
d + (2 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[2] * 4)));
+ scratch1024 + ((size_t)s[2] * 4)));
wuffs_base__store_u16le__no_bounds_check(
d + (3 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[3] * 4)));
+ scratch1024 + ((size_t)s[3] * 4)));
s += loop_unroll_count * 1;
d += loop_unroll_count * 2;
@@ -10968,7 +10959,7 @@
while (n >= 1) {
wuffs_base__store_u16le__no_bounds_check(
d + (0 * 2), wuffs_base__load_u16le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
s += 1 * 1;
d += 1 * 2;
@@ -10980,12 +10971,9 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len2 = dst.len / 2;
size_t len = dst_len2 < src.len ? dst_len2 : src.len;
uint8_t* d = dst.ptr;
@@ -10995,7 +10983,7 @@
// TODO: unroll.
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)s0);
@@ -11012,10 +11000,9 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(
- wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
- wuffs_base__slice_u8 src) {
+wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
+ wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t src_len4 = src.len / 4;
size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4;
@@ -11041,8 +11028,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t src_len4 = src.len / 4;
@@ -11088,8 +11075,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -11119,8 +11106,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -11147,8 +11134,8 @@
static uint64_t //
wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len4 = src.len / 4;
@@ -11176,12 +11163,9 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxx__index__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len3 = dst.len / 3;
size_t len = dst_len3 < src.len ? dst_len3 : src.len;
uint8_t* d = dst.ptr;
@@ -11200,16 +11184,16 @@
while (n > loop_unroll_count) {
wuffs_base__store_u32le__no_bounds_check(
d + (0 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (1 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[1] * 4)));
+ scratch1024 + ((size_t)s[1] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (2 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[2] * 4)));
+ scratch1024 + ((size_t)s[2] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (3 * 3), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[3] * 4)));
+ scratch1024 + ((size_t)s[3] * 4)));
s += loop_unroll_count * 1;
d += loop_unroll_count * 3;
@@ -11217,7 +11201,7 @@
}
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);
@@ -11231,12 +11215,9 @@
static uint64_t //
wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len3 = dst.len / 3;
size_t len = dst_len3 < src.len ? dst_len3 : src.len;
uint8_t* d = dst.ptr;
@@ -11246,22 +11227,22 @@
const size_t loop_unroll_count = 4;
while (n >= loop_unroll_count) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);
}
- uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[1] * 4));
if (s1) {
wuffs_base__store_u24le__no_bounds_check(d + (1 * 3), s1);
}
- uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[2] * 4));
if (s2) {
wuffs_base__store_u24le__no_bounds_check(d + (2 * 3), s2);
}
- uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[3] * 4));
if (s3) {
wuffs_base__store_u24le__no_bounds_check(d + (3 * 3), s3);
@@ -11273,7 +11254,7 @@
}
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0);
@@ -11288,8 +11269,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__xxx__y(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxx__y(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len3 = dst.len / 3;
size_t len = dst_len3 < src.len ? dst_len3 : src.len;
@@ -11316,12 +11297,9 @@
// --------
static uint64_t //
-wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxxx__index__src(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len4 = dst.len / 4;
size_t len = dst_len4 < src.len ? dst_len4 : src.len;
uint8_t* d = dst.ptr;
@@ -11333,16 +11311,16 @@
while (n >= loop_unroll_count) {
wuffs_base__store_u32le__no_bounds_check(
d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (1 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[1] * 4)));
+ scratch1024 + ((size_t)s[1] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (2 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[2] * 4)));
+ scratch1024 + ((size_t)s[2] * 4)));
wuffs_base__store_u32le__no_bounds_check(
d + (3 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[3] * 4)));
+ scratch1024 + ((size_t)s[3] * 4)));
s += loop_unroll_count * 1;
d += loop_unroll_count * 4;
@@ -11352,7 +11330,7 @@
while (n >= 1) {
wuffs_base__store_u32le__no_bounds_check(
d + (0 * 4), wuffs_base__load_u32le__no_bounds_check(
- dst_palette.ptr + ((size_t)s[0] * 4)));
+ scratch1024 + ((size_t)s[0] * 4)));
s += 1 * 1;
d += 1 * 4;
@@ -11364,12 +11342,9 @@
static uint64_t //
wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over(
+ const uint8_t* scratch1024,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
- if (dst_palette.len != 1024) {
- return 0;
- }
size_t dst_len4 = dst.len / 4;
size_t len = dst_len4 < src.len ? dst_len4 : src.len;
uint8_t* d = dst.ptr;
@@ -11379,22 +11354,22 @@
const size_t loop_unroll_count = 4;
while (n >= loop_unroll_count) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);
}
- uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s1 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[1] * 4));
if (s1) {
wuffs_base__store_u32le__no_bounds_check(d + (1 * 4), s1);
}
- uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s2 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[2] * 4));
if (s2) {
wuffs_base__store_u32le__no_bounds_check(d + (2 * 4), s2);
}
- uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s3 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[3] * 4));
if (s3) {
wuffs_base__store_u32le__no_bounds_check(d + (3 * 4), s3);
@@ -11406,7 +11381,7 @@
}
while (n >= 1) {
- uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(dst_palette.ptr +
+ uint32_t s0 = wuffs_base__load_u32le__no_bounds_check(scratch1024 +
((size_t)s[0] * 4));
if (s0) {
wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0);
@@ -11421,8 +11396,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__xxxx__xxx(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxxx__xxx(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t src_len3 = src.len / 3;
@@ -11447,8 +11422,8 @@
}
static uint64_t //
-wuffs_base__pixel_swizzler__xxxx__y(wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
+wuffs_base__pixel_swizzler__xxxx__y(const uint8_t* scratch1024,
+ wuffs_base__slice_u8 dst,
wuffs_base__slice_u8 src) {
size_t dst_len4 = dst.len / 4;
size_t len = dst_len4 < src.len ? dst_len4 : src.len;
@@ -11475,7 +11450,6 @@
static wuffs_base__pixel_swizzler__func //
wuffs_base__pixel_swizzler__prepare__y(wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
switch (dst_format.repr) {
@@ -11503,15 +11477,15 @@
wuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(
wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
+ wuffs_base__slice_u8 scratch =
+ wuffs_base__make_slice_u8(&(p->private_impl.scratch1024[0]), 1024);
switch (dst_format.repr) {
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_NONPREMUL:
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_PREMUL:
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:
- if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=
- 1024) {
+ if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {
return NULL;
}
switch (blend) {
@@ -11521,7 +11495,7 @@
return NULL;
case WUFFS_BASE__PIXEL_FORMAT__BGR_565:
- if (wuffs_base__pixel_swizzler__squash_bgr_565_888(dst_palette,
+ if (wuffs_base__pixel_swizzler__squash_bgr_565_888(scratch,
src_palette) != 1024) {
return NULL;
}
@@ -11534,8 +11508,7 @@
return NULL;
case WUFFS_BASE__PIXEL_FORMAT__BGR:
- if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=
- 1024) {
+ if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {
return NULL;
}
switch (blend) {
@@ -11549,8 +11522,7 @@
case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:
case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:
case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:
- if (wuffs_base__slice_u8__copy_from_slice(dst_palette, src_palette) !=
- 1024) {
+ if (wuffs_base__slice_u8__copy_from_slice(scratch, src_palette) != 1024) {
return NULL;
}
switch (blend) {
@@ -11562,8 +11534,8 @@
return NULL;
case WUFFS_BASE__PIXEL_FORMAT__RGB:
- if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(dst_palette,
- src_palette) != 1024) {
+ if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(scratch, src_palette) !=
+ 1024) {
return NULL;
}
switch (blend) {
@@ -11577,8 +11549,8 @@
case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:
case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:
case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:
- if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(dst_palette,
- src_palette) != 1024) {
+ if (wuffs_base__pixel_swizzler__swap_rgbx_bgrx(scratch, src_palette) !=
+ 1024) {
return NULL;
}
switch (blend) {
@@ -11595,7 +11567,6 @@
static wuffs_base__pixel_swizzler__func //
wuffs_base__pixel_swizzler__prepare__bgr(wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
switch (dst_format.repr) {
@@ -11626,7 +11597,6 @@
wuffs_base__pixel_swizzler__prepare__bgra_nonpremul(
wuffs_base__pixel_swizzler* p,
wuffs_base__pixel_format dst_format,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src_palette,
wuffs_base__pixel_blend blend) {
switch (dst_format.repr) {
@@ -11701,39 +11671,46 @@
switch (src_format.repr) {
case WUFFS_BASE__PIXEL_FORMAT__Y:
- func = wuffs_base__pixel_swizzler__prepare__y(p, dst_format, dst_palette,
- src_palette, blend);
+ func = wuffs_base__pixel_swizzler__prepare__y(p, dst_format, src_palette,
+ blend);
break;
case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:
func = wuffs_base__pixel_swizzler__prepare__indexed__bgra_binary(
- p, dst_format, dst_palette, src_palette, blend);
+ p, dst_format, src_palette, blend);
break;
case WUFFS_BASE__PIXEL_FORMAT__BGR:
- func = wuffs_base__pixel_swizzler__prepare__bgr(
- p, dst_format, dst_palette, src_palette, blend);
+ func = wuffs_base__pixel_swizzler__prepare__bgr(p, dst_format,
+ src_palette, blend);
break;
case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:
func = wuffs_base__pixel_swizzler__prepare__bgra_nonpremul(
- p, dst_format, dst_palette, src_palette, blend);
+ p, dst_format, src_palette, blend);
break;
}
p->private_impl.func = func;
- return wuffs_base__make_status(
- func ? NULL : wuffs_base__error__unsupported_pixel_swizzler_option);
+ if (!func) {
+ return wuffs_base__make_status(
+ wuffs_base__error__unsupported_pixel_swizzler_option);
+ }
+ if (dst_palette.len == 1024) {
+ const uint8_t* scratch1024 = &(p->private_impl.scratch1024[0]);
+ memcpy(dst_palette.ptr, scratch1024, 1024);
+ }
+ return wuffs_base__make_status(NULL);
}
WUFFS_BASE__MAYBE_STATIC uint64_t //
wuffs_base__pixel_swizzler__swizzle_interleaved(
const wuffs_base__pixel_swizzler* p,
wuffs_base__slice_u8 dst,
- wuffs_base__slice_u8 dst_palette,
wuffs_base__slice_u8 src) {
if (p && p->private_impl.func) {
- return (*p->private_impl.func)(dst, dst_palette, src);
+ const uint8_t* scratch1024 = &(p->private_impl.scratch1024[0]);
+ return (*p->private_impl.func)(scratch1024, dst, src);
}
return 0;
}
@@ -13001,7 +12978,6 @@
wuffs_base__pixel_swizzler__swizzle_interleaved(
&self->private_impl.f_swizzler,
wuffs_base__slice_u8__subslice_i(v_dst, v_i),
- wuffs_base__utility__empty_slice_u8(),
wuffs_base__slice_u8__subslice_j(
wuffs_base__make_slice_u8(self->private_impl.f_stash, 4),
self->private_impl.f_num_stashed));
@@ -13035,8 +13011,7 @@
if (v_i < ((uint64_t)(v_dst.len))) {
v_n = wuffs_base__pixel_swizzler__swizzle_interleaved(
&self->private_impl.f_swizzler,
- wuffs_base__slice_u8__subslice_i(v_dst, v_i),
- wuffs_base__utility__empty_slice_u8(), a_src);
+ wuffs_base__slice_u8__subslice_i(v_dst, v_i), a_src);
wuffs_base__u32__sat_add_indirect(&self->private_impl.f_dst_x,
((uint32_t)((v_n & 4294967295))));
v_n = ((v_n & 4294967295) * ((uint64_t)(v_src_bytes_per_pixel)));
@@ -21019,7 +20994,6 @@
uint32_t v_num_palette_entries = 0;
uint32_t v_i = 0;
uint32_t v_argb = 0;
- wuffs_base__slice_u8 v_dst_palette = {0};
wuffs_base__status v_status = wuffs_base__make_status(NULL);
uint8_t v_lw = 0;
@@ -21138,14 +21112,10 @@
(4 * ((uint32_t)(self->private_impl.f_gc_transparent_index))) + 3)] =
0;
}
- v_dst_palette = wuffs_base__pixel_buffer__palette(a_dst);
- if (((uint64_t)(v_dst_palette.len)) == 0) {
- v_dst_palette =
- wuffs_base__make_slice_u8(self->private_data.f_dst_palette, 1024);
- }
v_status = wuffs_base__pixel_swizzler__prepare(
&self->private_impl.f_swizzler,
- wuffs_base__pixel_buffer__pixel_format(a_dst), v_dst_palette,
+ wuffs_base__pixel_buffer__pixel_format(a_dst),
+ wuffs_base__pixel_buffer__palette(a_dst),
wuffs_base__utility__make_pixel_format(2198077448),
wuffs_base__make_slice_u8(
self->private_data.f_palettes[v_which_palette], 1024),
@@ -21502,9 +21472,7 @@
v_dst = wuffs_base__slice_u8__subslice_i(v_dst, v_i);
}
v_n = wuffs_base__pixel_swizzler__swizzle_interleaved(
- &self->private_impl.f_swizzler, v_dst,
- wuffs_base__make_slice_u8(self->private_data.f_dst_palette, 1024),
- v_src);
+ &self->private_impl.f_swizzler, v_dst, v_src);
wuffs_base__u64__sat_add_indirect(&v_src_ri, v_n);
wuffs_base__u32__sat_add_indirect(&self->private_impl.f_dst_x,
((uint32_t)((v_n & 4294967295))));
@@ -24977,7 +24945,6 @@
v_c = ((uint8_t)(((((uint32_t)(v_c)) << 1) & 255)));
wuffs_base__pixel_swizzler__swizzle_interleaved(
&self->private_impl.f_swizzler, v_dst,
- wuffs_base__utility__empty_slice_u8(),
wuffs_base__make_slice_u8(v_src, 1));
if (v_dst_bytes_per_pixel <= ((uint64_t)(v_dst.len))) {
v_dst =
diff --git a/std/bmp/decode_bmp.wuffs b/std/bmp/decode_bmp.wuffs
index 5af01c6..106340e 100644
--- a/std/bmp/decode_bmp.wuffs
+++ b/std/bmp/decode_bmp.wuffs
@@ -341,7 +341,6 @@
if i < dst.length() {
this.swizzler.swizzle_interleaved!(
dst: dst[i ..],
- dst_palette: this.util.empty_slice_u8(),
src: this.stash[.. this.num_stashed])
this.dst_x ~sat+= 1
}
@@ -377,7 +376,6 @@
if i < dst.length() {
n = this.swizzler.swizzle_interleaved!(
dst: dst[i ..],
- dst_palette: this.util.empty_slice_u8(),
src: args.src)
this.dst_x ~sat+= (n & 0xFFFF_FFFF) as base.u32
n = (n & 0xFFFF_FFFF) * (src_bytes_per_pixel as base.u64)
diff --git a/std/gif/decode_gif.wuffs b/std/gif/decode_gif.wuffs
index 235a167..717012f 100644
--- a/std/gif/decode_gif.wuffs
+++ b/std/gif/decode_gif.wuffs
@@ -137,8 +137,6 @@
// palettes[0] and palettes[1] are the Global and Local Color Table.
palettes : array[2] array[4 * 256] base.u8,
- // dst_palette is the swizzled color table.
- dst_palette : array[4 * 256] base.u8,
lzw : lzw.decoder,
//#WHEN PREPROC101 decode_config.wuffs
@@ -821,7 +819,6 @@
var num_palette_entries : base.u32[..= 256]
var i : base.u32
var argb : base.u32
- var dst_palette : slice base.u8
var status : base.status
var lw : base.u8
@@ -873,16 +870,11 @@
this.palettes[1][(4 * (this.gc_transparent_index as base.u32)) + 3] = 0x00
}
- dst_palette = args.dst.palette()
- if dst_palette.length() == 0 {
- dst_palette = this.dst_palette[..]
- }
-
// TODO: a Wuffs (not just C) name for the
// WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY magic pixfmt constant.
status = this.swizzler.prepare!(
dst_pixfmt: args.dst.pixel_format(),
- dst_palette: dst_palette,
+ dst_palette: args.dst.palette(),
src_pixfmt: this.util.make_pixel_format(repr: 0x8304_0008),
src_palette: this.palettes[which_palette][..],
blend: args.blend)
@@ -1074,8 +1066,7 @@
} else {
dst = dst[i ..]
}
- n = this.swizzler.swizzle_interleaved!(
- dst: dst, dst_palette: this.dst_palette[..], src: src)
+ n = this.swizzler.swizzle_interleaved!(dst: dst, src: src)
src_ri ~sat+= n
this.dst_x ~sat+= (n & 0xFFFF_FFFF) as base.u32
diff --git a/std/wbmp/decode_wbmp.wuffs b/std/wbmp/decode_wbmp.wuffs
index b0be49c..d0a74c6 100644
--- a/std/wbmp/decode_wbmp.wuffs
+++ b/std/wbmp/decode_wbmp.wuffs
@@ -212,8 +212,7 @@
// v_c <<= 1;
c = (((c as base.u32) << 1) & 0xFF) as base.u8
- this.swizzler.swizzle_interleaved!(
- dst: dst, dst_palette: this.util.empty_slice_u8(), src: src[..])
+ this.swizzler.swizzle_interleaved!(dst: dst, src: src[..])
if dst_bytes_per_pixel <= dst.length() {
dst = dst[dst_bytes_per_pixel ..]
diff --git a/test/c/std/wbmp.c b/test/c/std/wbmp.c
index 0cb7ad4..bf65c11 100644
--- a/test/c/std/wbmp.c
+++ b/test/c/std/wbmp.c
@@ -107,7 +107,6 @@
const uint32_t width = 5;
const uint32_t height = 5;
- uint8_t dummy_palette_array[1024];
wuffs_base__pixel_swizzler swizzler;
const struct {
@@ -201,14 +200,6 @@
wuffs_base__pixel_alpha_transparency dst_transparency =
wuffs_base__pixel_format__transparency(&dst_pixfmt);
- wuffs_base__slice_u8 dst_palette =
- wuffs_base__pixel_buffer__palette(&dst_pixbuf);
- if (dst_palette.len == 0) {
- dst_palette = wuffs_base__make_slice_u8(
- &dummy_palette_array[0],
- WUFFS_TESTLIB_ARRAY_SIZE(dummy_palette_array));
- }
-
int b;
for (b = 0; b < WUFFS_TESTLIB_ARRAY_SIZE(blends); b++) {
// Set the middle dst pixel.
@@ -220,14 +211,14 @@
CHECK_STATUS(
"prepare",
wuffs_base__pixel_swizzler__prepare(
- &swizzler, dst_pixfmt, dst_palette,
+ &swizzler, dst_pixfmt,
+ wuffs_base__pixel_buffer__palette(&dst_pixbuf),
wuffs_base__make_pixel_format(srcs[s].pixfmt_repr),
wuffs_base__pixel_buffer__palette(&src_pixbuf), blends[b]));
wuffs_base__pixel_swizzler__swizzle_interleaved(
&swizzler,
wuffs_base__table_u8__row(
wuffs_base__pixel_buffer__plane(&dst_pixbuf, 0), height / 2),
- dst_palette,
wuffs_base__table_u8__row(
wuffs_base__pixel_buffer__plane(&src_pixbuf, 0), height / 2));
