From 302afb7ce574b4c7b55e77062b64b49ada3777ca Mon Sep 17 00:00:00 2001 From: Nigel Tao Date: Thu, 28 May 2020 00:07:45 +1000 Subject: [PATCH] Add const to wuffs_base__pixel_swizzler__func arg --- internal/cgen/base/image-public.h | 10 +- internal/cgen/base/pixconv-submodule.c | 454 +++++++++++++----------- internal/cgen/data.go | 78 +++-- release/c/wuffs-unsupported-snapshot.c | 464 ++++++++++++++----------- 4 files changed, 572 insertions(+), 434 deletions(-) diff --git a/internal/cgen/base/image-public.h b/internal/cgen/base/image-public.h index 39c62ed8..596115ba 100644 --- a/internal/cgen/base/image-public.h +++ b/internal/cgen/base/image-public.h @@ -1287,10 +1287,12 @@ wuffs_base__pixel_palette__closest_element( // -------- // 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)(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len); typedef struct { // Do not access the private_impl's fields directly. There is no API/ABI diff --git a/internal/cgen/base/pixconv-submodule.c b/internal/cgen/base/pixconv-submodule.c index e041e5c8..d556db51 100644 --- a/internal/cgen/base/pixconv-submodule.c +++ b/internal/cgen/base/pixconv-submodule.c @@ -41,7 +41,7 @@ wuffs_base__pixel_buffer__color_u32_at(const wuffs_base__pixel_buffer* pb, } size_t stride = pb->private_impl.planes[0].stride; - uint8_t* row = pb->private_impl.planes[0].ptr + (stride * ((size_t)y)); + const uint8_t* row = pb->private_impl.planes[0].ptr + (stride * ((size_t)y)); switch (pb->pixcfg.private_impl.pixfmt.repr) { case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL: @@ -403,7 +403,7 @@ wuffs_base__pixel_swizzler__squash_bgr_565_888(wuffs_base__slice_u8 dst, wuffs_base__slice_u8 src) { size_t len4 = (dst.len < src.len ? dst.len : src.len) / 4; uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + const uint8_t* s = src.ptr; size_t n = len4; while (n--) { @@ -425,7 +425,7 @@ wuffs_base__pixel_swizzler__swap_rgbx_bgrx(wuffs_base__slice_u8 dst, wuffs_base__slice_u8 src) { size_t len4 = (dst.len < src.len ? dst.len : src.len) / 4; uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + const uint8_t* s = src.ptr; size_t n = len4; while (n--) { @@ -446,34 +446,47 @@ wuffs_base__pixel_swizzler__swap_rgbx_bgrx(wuffs_base__slice_u8 dst, // -------- static uint64_t // -wuffs_base__pixel_swizzler__copy_1_1(wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - 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__slice_u8 src) { - size_t dst_len3 = dst.len / 3; - size_t src_len3 = src.len / 3; - size_t len = dst_len3 < src_len3 ? dst_len3 : src_len3; +wuffs_base__pixel_swizzler__copy_1_1(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t len = (dst_len < src_len) ? dst_len : src_len; if (len > 0) { - memmove(dst.ptr, src.ptr, len * 3); + memmove(dst_ptr, src_ptr, len); } return len; } static uint64_t // -wuffs_base__pixel_swizzler__copy_4_4(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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; +wuffs_base__pixel_swizzler__copy_3_3(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len3 = dst_len / 3; + size_t src_len3 = src_len / 3; + size_t len = (dst_len3 < src_len3) ? dst_len3 : src_len3; if (len > 0) { - memmove(dst.ptr, src.ptr, len * 4); + memmove(dst_ptr, src_ptr, len * 3); + } + return len; +} + +static uint64_t // +wuffs_base__pixel_swizzler__copy_4_4(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + if (len > 0) { + memmove(dst_ptr, src_ptr, len * 4); } return len; } @@ -481,14 +494,17 @@ wuffs_base__pixel_swizzler__copy_4_4(wuffs_base__slice_u8 dst, // -------- static uint64_t // -wuffs_base__pixel_swizzler__bgr_565__bgr(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_len3 = src.len / 3; - size_t len = dst_len2 < src_len3 ? dst_len2 : src_len3; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__bgr_565__bgr(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t src_len3 = src_len / 3; + size_t len = (dst_len2 < src_len3) ? dst_len2 : src_len3; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -510,14 +526,17 @@ wuffs_base__pixel_swizzler__bgr_565__bgr(wuffs_base__slice_u8 dst, static uint64_t // wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src( - 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; - size_t len = dst_len2 < src_len4 ? dst_len2 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t src_len4 = src_len / 4; + size_t len = (dst_len2 < src_len4) ? dst_len2 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -539,14 +558,17 @@ wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src( static uint64_t // wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over( - 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; - size_t len = dst_len2 < src_len4 ? dst_len2 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t src_len4 = src_len / 4; + size_t len = (dst_len2 < src_len4) ? dst_len2 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -592,13 +614,16 @@ wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over( } static uint64_t // -wuffs_base__pixel_swizzler__bgr_565__y(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 len = dst_len2 < src.len ? dst_len2 : src.len; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__bgr_565__y(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t len = (dst_len2 < src_len) ? dst_len2 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -618,17 +643,19 @@ wuffs_base__pixel_swizzler__bgr_565__y(wuffs_base__slice_u8 dst, } 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) { +wuffs_base__pixel_swizzler__bgr_565__index__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len2 = dst_len / 2; + size_t len = (dst_len2 < src_len) ? dst_len2 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; const size_t loop_unroll_count = 4; @@ -636,16 +663,16 @@ wuffs_base__pixel_swizzler__bgr_565__index__src( 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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((size_t)s[3] * 4))); s += loop_unroll_count * 1; d += loop_unroll_count * 2; @@ -655,7 +682,7 @@ wuffs_base__pixel_swizzler__bgr_565__index__src( 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))); + dst_palette_ptr + ((size_t)s[0] * 4))); s += 1 * 1; d += 1 * 2; @@ -667,22 +694,25 @@ wuffs_base__pixel_swizzler__bgr_565__index__src( static uint64_t // wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over( - wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len2 = dst_len / 2; + size_t len = (dst_len2 < src_len) ? dst_len2 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // 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(dst_palette_ptr + ((size_t)s[0] * 4)); if (s0) { wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)s0); @@ -699,15 +729,17 @@ wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over( // -------- 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) { - 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; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -728,14 +760,17 @@ wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src( static uint64_t // wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over( - 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; - size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -775,14 +810,17 @@ wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over( static uint64_t // wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over( - 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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -806,14 +844,17 @@ wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over( static uint64_t // wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src( - 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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -834,14 +875,17 @@ wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src( static uint64_t // wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over( - 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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -863,16 +907,19 @@ wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over( // -------- static uint64_t // -wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { +wuffs_base__pixel_swizzler__xxx__index__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len3 = dst_len / 3; + size_t len = (dst_len3 < src_len) ? dst_len3 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; const size_t loop_unroll_count = 4; @@ -887,16 +934,16 @@ wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, 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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((size_t)s[3] * 4))); s += loop_unroll_count * 1; d += loop_unroll_count * 3; @@ -904,7 +951,7 @@ wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, } 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(dst_palette_ptr + ((size_t)s[0] * 4)); wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0); @@ -918,37 +965,40 @@ wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, static uint64_t // wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over( - wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len3 = dst_len / 3; + size_t len = (dst_len3 < src_len) ? dst_len3 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; 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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((size_t)s[3] * 4)); if (s3) { wuffs_base__store_u24le__no_bounds_check(d + (3 * 3), s3); @@ -960,7 +1010,7 @@ wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over( } 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(dst_palette_ptr + ((size_t)s[0] * 4)); if (s0) { wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0); @@ -975,13 +1025,16 @@ wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over( } static uint64_t // -wuffs_base__pixel_swizzler__xxx__y(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 len = dst_len3 < src.len ? dst_len3 : src.len; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__xxx__y(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len3 = dst_len / 3; + size_t len = (dst_len3 < src_len) ? dst_len3 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -1003,16 +1056,19 @@ wuffs_base__pixel_swizzler__xxx__y(wuffs_base__slice_u8 dst, // -------- static uint64_t // -wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { +wuffs_base__pixel_swizzler__xxxx__index__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len4 = dst_len / 4; + size_t len = (dst_len4 < src_len) ? dst_len4 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; const size_t loop_unroll_count = 4; @@ -1020,16 +1076,16 @@ wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, 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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((size_t)s[3] * 4))); s += loop_unroll_count * 1; d += loop_unroll_count * 4; @@ -1039,7 +1095,7 @@ wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, 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))); + dst_palette_ptr + ((size_t)s[0] * 4))); s += 1 * 1; d += 1 * 4; @@ -1051,37 +1107,40 @@ wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, static uint64_t // wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over( - wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len4 = dst_len / 4; + size_t len = (dst_len4 < src_len) ? dst_len4 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; 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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((size_t)s[3] * 4)); if (s3) { wuffs_base__store_u32le__no_bounds_check(d + (3 * 4), s3); @@ -1093,7 +1152,7 @@ wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over( } 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(dst_palette_ptr + ((size_t)s[0] * 4)); if (s0) { wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0); @@ -1108,14 +1167,17 @@ wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over( } static uint64_t // -wuffs_base__pixel_swizzler__xxxx__xxx(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_len3 = src.len / 3; - size_t len = dst_len4 < src_len3 ? dst_len4 : src_len3; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__xxxx__xxx(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len3 = src_len / 3; + size_t len = (dst_len4 < src_len3) ? dst_len4 : src_len3; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -1134,13 +1196,16 @@ wuffs_base__pixel_swizzler__xxxx__xxx(wuffs_base__slice_u8 dst, } static uint64_t // -wuffs_base__pixel_swizzler__xxxx__y(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 len = dst_len4 < src.len ? dst_len4 : src.len; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__xxxx__y(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t len = (dst_len4 < src_len) ? dst_len4 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -1420,7 +1485,8 @@ wuffs_base__pixel_swizzler__swizzle_interleaved( 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); + return (*p->private_impl.func)(dst.ptr, dst.len, dst_palette.ptr, + dst_palette.len, src.ptr, src.len); } return 0; } diff --git a/internal/cgen/data.go b/internal/cgen/data.go index 4e6f069a..a0020045 100644 --- a/internal/cgen/data.go +++ b/internal/cgen/data.go @@ -115,13 +115,13 @@ const baseF64ConvSubmoduleC = "" + const basePixConvSubmoduleC = "" + "// ---------------- Pixel Swizzler\n\nstatic inline uint32_t //\nwuffs_base__swap_u32_argb_abgr(uint32_t u) {\n uint32_t o = u & 0xFF00FF00;\n uint32_t r = u & 0x00FF0000;\n uint32_t b = u & 0x000000FF;\n return o | (r >> 16) | (b << 16);\n}\n\n" + "" + - "// --------\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__color_u32_argb_premul //\nwuffs_base__pixel_buffer__color_u32_at(const wuffs_base__pixel_buffer* pb,\n uint32_t x,\n uint32_t y) {\n if (!pb || (x >= pb->pixcfg.private_impl.width) ||\n (y >= pb->pixcfg.private_impl.height)) {\n return 0;\n }\n\n if (wuffs_base__pixel_format__is_planar(&pb->pixcfg.private_impl.pixfmt)) {\n // TODO: support planar formats.\n return 0;\n }\n\n size_t stride = pb->private_impl.planes[0].stride;\n uint8_t* row = pb->private_impl.planes[0].ptr + (stride * ((size_t)y));\n\n switch (pb->pixcfg.private_impl.pixfmt.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n return wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x)));\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY: {\n uint8_t* palette = pb->private_impl.plane" + - "s[3].ptr;\n return wuffs_base__load_u32le__no_bounds_check(palette +\n (4 * ((size_t)row[x])));\n }\n\n // Common formats above. Rarer formats below.\n\n case WUFFS_BASE__PIXEL_FORMAT__Y:\n return 0xFF000000 | (0x00010101 * ((uint32_t)(row[x])));\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_NONPREMUL: {\n uint8_t* palette = pb->private_impl.planes[3].ptr;\n return wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(palette +\n (4 * ((size_t)row[x]))));\n }\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n return wuffs_base__color_u16_rgb_565__as__color_u32_argb_premul(\n wuffs_base__load_u16le__no_bounds_check(row + (2 * ((size_t)x))));\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n return 0xFF000000 |\n wuffs_base__load_u24le__no_bounds_check(row + (3 * ((size_t)x)));\n case WUFFS_BASE__PIXEL_FORMAT_" + - "_BGRA_NONPREMUL:\n return wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x))));\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n return 0xFF000000 |\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x)));\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n return wuffs_base__swap_u32_argb_abgr(\n 0xFF000000 |\n wuffs_base__load_u24le__no_bounds_check(row + (3 * ((size_t)x))));\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n return wuffs_base__swap_u32_argb_abgr(\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(row +\n (4 * ((size_t)x)))));\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n return wuffs_base__swap_u32_argb_abgr(\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x)" + - ")));\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n return wuffs_base__swap_u32_argb_abgr(\n 0xFF000000 |\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x))));\n\n default:\n // TODO: support more formats.\n break;\n }\n\n return 0;\n}\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__status //\nwuffs_base__pixel_buffer__set_color_u32_at(\n wuffs_base__pixel_buffer* pb,\n uint32_t x,\n uint32_t y,\n wuffs_base__color_u32_argb_premul color) {\n if (!pb) {\n return wuffs_base__make_status(wuffs_base__error__bad_receiver);\n }\n if ((x >= pb->pixcfg.private_impl.width) ||\n (y >= pb->pixcfg.private_impl.height)) {\n return wuffs_base__make_status(wuffs_base__error__bad_argument);\n }\n\n if (wuffs_base__pixel_format__is_planar(&pb->pixcfg.private_impl.pixfmt)) {\n // TODO: support planar formats.\n return wuffs_base__make_status(wuffs_base__error__unsupported_option);\n }\n\n size_t stride = pb->private_impl.planes[0].stride;\n uint8_t* row = pb->private_impl.planes[0]" + - ".ptr + (stride * ((size_t)y));\n\n switch (pb->pixcfg.private_impl.pixfmt.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n wuffs_base__store_u32le__no_bounds_check(row + (4 * ((size_t)x)), color);\n break;\n\n // Common formats above. Rarer formats below.\n\n case WUFFS_BASE__PIXEL_FORMAT__Y:\n wuffs_base__store_u8__no_bounds_check(\n row + ((size_t)x),\n wuffs_base__color_u32_argb_premul__as__color_u8_gray(color));\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:\n wuffs_base__store_u8__no_bounds_check(\n row + ((size_t)x), wuffs_base__pixel_palette__closest_element(\n wuffs_base__pixel_buffer__palette(pb),\n pb->pixcfg.private_impl.pixfmt, color));\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n wuffs_base__store_u16le__no_bounds_check(\n row + (2 * ((size_t)x)),\n wuffs_base__color_u32_argb_premul__as" + - "__color_u16_rgb_565(color));\n break;\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n wuffs_base__store_u24le__no_bounds_check(row + (3 * ((size_t)x)), color);\n break;\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n wuffs_base__store_u32le__no_bounds_check(\n row + (4 * ((size_t)x)),\n wuffs_base__color_u32_argb_premul__as__color_u32_argb_nonpremul(\n color));\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n wuffs_base__store_u24le__no_bounds_check(\n row + (3 * ((size_t)x)), wuffs_base__swap_u32_argb_abgr(color));\n break;\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n wuffs_base__store_u32le__no_bounds_check(\n row + (4 * ((size_t)x)),\n wuffs_base__color_u32_argb_premul__as__color_u32_argb_nonpremul(\n wuffs_base__swap_u32_argb_abgr(color)));\n break;\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n wuffs_base__store_u32le__no_bounds_check(\n row" + - " + (4 * ((size_t)x)), wuffs_base__swap_u32_argb_abgr(color));\n break;\n\n default:\n // TODO: support more formats.\n return wuffs_base__make_status(wuffs_base__error__unsupported_option);\n }\n\n return wuffs_base__make_status(NULL);\n}\n\n" + + "// --------\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__color_u32_argb_premul //\nwuffs_base__pixel_buffer__color_u32_at(const wuffs_base__pixel_buffer* pb,\n uint32_t x,\n uint32_t y) {\n if (!pb || (x >= pb->pixcfg.private_impl.width) ||\n (y >= pb->pixcfg.private_impl.height)) {\n return 0;\n }\n\n if (wuffs_base__pixel_format__is_planar(&pb->pixcfg.private_impl.pixfmt)) {\n // TODO: support planar formats.\n return 0;\n }\n\n size_t stride = pb->private_impl.planes[0].stride;\n const uint8_t* row = pb->private_impl.planes[0].ptr + (stride * ((size_t)y));\n\n switch (pb->pixcfg.private_impl.pixfmt.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_BINARY:\n return wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x)));\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY: {\n uint8_t* palette = pb->private_impl" + + ".planes[3].ptr;\n return wuffs_base__load_u32le__no_bounds_check(palette +\n (4 * ((size_t)row[x])));\n }\n\n // Common formats above. Rarer formats below.\n\n case WUFFS_BASE__PIXEL_FORMAT__Y:\n return 0xFF000000 | (0x00010101 * ((uint32_t)(row[x])));\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_NONPREMUL: {\n uint8_t* palette = pb->private_impl.planes[3].ptr;\n return wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(palette +\n (4 * ((size_t)row[x]))));\n }\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n return wuffs_base__color_u16_rgb_565__as__color_u32_argb_premul(\n wuffs_base__load_u16le__no_bounds_check(row + (2 * ((size_t)x))));\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n return 0xFF000000 |\n wuffs_base__load_u24le__no_bounds_check(row + (3 * ((size_t)x)));\n case WUFFS_BASE__PIXEL_F" + + "ORMAT__BGRA_NONPREMUL:\n return wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x))));\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n return 0xFF000000 |\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x)));\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n return wuffs_base__swap_u32_argb_abgr(\n 0xFF000000 |\n wuffs_base__load_u24le__no_bounds_check(row + (3 * ((size_t)x))));\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n return wuffs_base__swap_u32_argb_abgr(\n wuffs_base__color_u32_argb_nonpremul__as__color_u32_argb_premul(\n wuffs_base__load_u32le__no_bounds_check(row +\n (4 * ((size_t)x)))));\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_BINARY:\n return wuffs_base__swap_u32_argb_abgr(\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((siz" + + "e_t)x))));\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n return wuffs_base__swap_u32_argb_abgr(\n 0xFF000000 |\n wuffs_base__load_u32le__no_bounds_check(row + (4 * ((size_t)x))));\n\n default:\n // TODO: support more formats.\n break;\n }\n\n return 0;\n}\n\nWUFFS_BASE__MAYBE_STATIC wuffs_base__status //\nwuffs_base__pixel_buffer__set_color_u32_at(\n wuffs_base__pixel_buffer* pb,\n uint32_t x,\n uint32_t y,\n wuffs_base__color_u32_argb_premul color) {\n if (!pb) {\n return wuffs_base__make_status(wuffs_base__error__bad_receiver);\n }\n if ((x >= pb->pixcfg.private_impl.width) ||\n (y >= pb->pixcfg.private_impl.height)) {\n return wuffs_base__make_status(wuffs_base__error__bad_argument);\n }\n\n if (wuffs_base__pixel_format__is_planar(&pb->pixcfg.private_impl.pixfmt)) {\n // TODO: support planar formats.\n return wuffs_base__make_status(wuffs_base__error__unsupported_option);\n }\n\n size_t stride = pb->private_impl.planes[0].stride;\n uint8_t* row = pb->private_impl.pla" + + "nes[0].ptr + (stride * ((size_t)y));\n\n switch (pb->pixcfg.private_impl.pixfmt.repr) {\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__BGRX:\n wuffs_base__store_u32le__no_bounds_check(row + (4 * ((size_t)x)), color);\n break;\n\n // Common formats above. Rarer formats below.\n\n case WUFFS_BASE__PIXEL_FORMAT__Y:\n wuffs_base__store_u8__no_bounds_check(\n row + ((size_t)x),\n wuffs_base__color_u32_argb_premul__as__color_u8_gray(color));\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_BINARY:\n wuffs_base__store_u8__no_bounds_check(\n row + ((size_t)x), wuffs_base__pixel_palette__closest_element(\n wuffs_base__pixel_buffer__palette(pb),\n pb->pixcfg.private_impl.pixfmt, color));\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__BGR_565:\n wuffs_base__store_u16le__no_bounds_check(\n row + (2 * ((size_t)x)),\n wuffs_base__color_u32_argb_prem" + + "ul__as__color_u16_rgb_565(color));\n break;\n case WUFFS_BASE__PIXEL_FORMAT__BGR:\n wuffs_base__store_u24le__no_bounds_check(row + (3 * ((size_t)x)), color);\n break;\n case WUFFS_BASE__PIXEL_FORMAT__BGRA_NONPREMUL:\n wuffs_base__store_u32le__no_bounds_check(\n row + (4 * ((size_t)x)),\n wuffs_base__color_u32_argb_premul__as__color_u32_argb_nonpremul(\n color));\n break;\n\n case WUFFS_BASE__PIXEL_FORMAT__RGB:\n wuffs_base__store_u24le__no_bounds_check(\n row + (3 * ((size_t)x)), wuffs_base__swap_u32_argb_abgr(color));\n break;\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_NONPREMUL:\n wuffs_base__store_u32le__no_bounds_check(\n row + (4 * ((size_t)x)),\n wuffs_base__color_u32_argb_premul__as__color_u32_argb_nonpremul(\n wuffs_base__swap_u32_argb_abgr(color)));\n break;\n case WUFFS_BASE__PIXEL_FORMAT__RGBA_PREMUL:\n case WUFFS_BASE__PIXEL_FORMAT__RGBX:\n wuffs_base__store_u32le__no_bounds_check(\n " + + " row + (4 * ((size_t)x)), wuffs_base__swap_u32_argb_abgr(color));\n break;\n\n default:\n // TODO: support more formats.\n return wuffs_base__make_status(wuffs_base__error__unsupported_option);\n }\n\n return wuffs_base__make_status(NULL);\n}\n\n" + "" + "// --------\n\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 size_t n = palette_slice.len / 4;\n if (n > 256) {\n n = 256;\n }\n size_t best_index = 0;\n uint64_t best_score = 0xFFFFFFFFFFFFFFFF;\n\n // Work in 16-bit color.\n uint32_t ca = 0x101 * (0xFF & (c >> 24));\n uint32_t cr = 0x101 * (0xFF & (c >> 16));\n uint32_t cg = 0x101 * (0xFF & (c >> 8));\n uint32_t cb = 0x101 * (0xFF & (c >> 0));\n\n switch (palette_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 bool nonpremul = palette_format.repr ==\n WUFFS_BASE__PIXEL_FORMAT__INDEXED__BGRA_NONPREMUL;\n\n size_t i;\n for (i = 0; i < n; i++) {\n // Work in 16-bit color.\n uint32_t pb = 0x101 * ((uint32_t)(palette" + "_slice.ptr[(4 * i) + 0]));\n uint32_t pg = 0x101 * ((uint32_t)(palette_slice.ptr[(4 * i) + 1]));\n uint32_t pr = 0x101 * ((uint32_t)(palette_slice.ptr[(4 * i) + 2]));\n uint32_t pa = 0x101 * ((uint32_t)(palette_slice.ptr[(4 * i) + 3]));\n\n // Convert to premultiplied alpha.\n if (nonpremul && (pa != 0xFFFF)) {\n pb = (pb * pa) / 0xFFFF;\n pg = (pg * pa) / 0xFFFF;\n pr = (pr * pa) / 0xFFFF;\n }\n\n // These deltas are conceptually int32_t (signed) but after squaring,\n // it's equivalent to work in uint32_t (unsigned).\n pb -= cb;\n pg -= cg;\n pr -= cr;\n pa -= ca;\n uint64_t score = ((uint64_t)(pb * pb)) + ((uint64_t)(pg * pg)) +\n ((uint64_t)(pr * pr)) + ((uint64_t)(pa * pa));\n if (best_score > score) {\n best_score = score;\n best_index = i;\n }\n }\n break;\n }\n }\n\n return (uint8_t)best_index;\n}\n\n" + @@ -132,39 +132,41 @@ const basePixConvSubmoduleC = "" + " uint32_t sa = 0x101 * (0xFF & (src_nonpremul >> 24));\n uint32_t sr = 0x101 * (0xFF & (src_nonpremul >> 16));\n uint32_t sg = 0x101 * (0xFF & (src_nonpremul >> 8));\n uint32_t sb = 0x101 * (0xFF & (src_nonpremul >> 0));\n uint32_t da = 0x101 * (0xFF & (dst_premul >> 24));\n uint32_t dr = 0x101 * (0xFF & (dst_premul >> 16));\n uint32_t dg = 0x101 * (0xFF & (dst_premul >> 8));\n uint32_t db = 0x101 * (0xFF & (dst_premul >> 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 da = sa + ((da * ia) / 0xFFFF);\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 and combine the components.\n da >>= 8;\n dr >>= 8;\n dg >>= 8;\n db >>= 8;\n return (db << 0) | (dg << 8) | (dr << 16) | (da << 24);\n}\n\nstatic inline uint32_t //\nwuffs_base__composite_premul_premul_u32_axxx(uint32_t dst_premul,\n " + " uint32_t src_premul) {\n // Convert from 8-bit color to 16-bit color.\n uint32_t sa = 0x101 * (0xFF & (src_premul >> 24));\n uint32_t sr = 0x101 * (0xFF & (src_premul >> 16));\n uint32_t sg = 0x101 * (0xFF & (src_premul >> 8));\n uint32_t sb = 0x101 * (0xFF & (src_premul >> 0));\n uint32_t da = 0x101 * (0xFF & (dst_premul >> 24));\n uint32_t dr = 0x101 * (0xFF & (dst_premul >> 16));\n uint32_t dg = 0x101 * (0xFF & (dst_premul >> 8));\n uint32_t db = 0x101 * (0xFF & (dst_premul >> 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 (premul) over dst (premul).\n da = sa + ((da * ia) / 0xFFFF);\n dr = sr + ((dr * ia) / 0xFFFF);\n dg = sg + ((dg * ia) / 0xFFFF);\n db = sb + ((db * ia) / 0xFFFF);\n\n // Convert from 16-bit color to 8-bit color and combine the components.\n da >>= 8;\n dr >>= 8;\n dg >>= 8;\n db >>= 8;\n return (db << 0) | (dg << 8) | (dr << 16) | (da << 24);\n}\n\n" + "" + - "// --------\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__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 const 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 const 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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\n size_t len = (dst_len < src_len) ? dst_len : src_len;\n if (len > 0) {\n memmove(dst_ptr, src_ptr, len);\n }\n return len;\n}\n\nstatic uint64_t //\nwuffs_base__pixel_swizzler__copy_3_3(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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__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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 uint8_t*" + + " dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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__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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 //\nw" + + "uffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over(\n uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 * s" + + "a) + (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 uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 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 ((siz" + + "e_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 uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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__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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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_palette_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 uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\n size_t dst_len4 = dst_len / 4;\n size_t src_le" + + "n3 = src_len / 3;\n size_t len = (dst_len4 < src_len3) ? dst_len4 : src_len3;\n uint8_t* d = dst_ptr;\n const 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(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len) {\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 const 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 " + @@ -176,7 +178,7 @@ const basePixConvSubmoduleC = "" + "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\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" + + " 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.ptr, dst.len, dst_palette.ptr,\n dst_palette.len, src.ptr, src.len);\n }\n return 0;\n}\n" + "" const baseFundamentalPrivateH = "" + @@ -318,10 +320,10 @@ const baseImagePublicH = "" + "" + "// --------\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)(uint8_t* dst_ptr,\n size_t dst_len,\n uint8_t* dst_palette_ptr,\n size_t dst_palette_len,\n const uint8_t* src_ptr,\n size_t src_len);\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#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 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__slice_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" + "" const baseIOPrivateH = "" + diff --git a/release/c/wuffs-unsupported-snapshot.c b/release/c/wuffs-unsupported-snapshot.c index 9415def0..3065fab7 100644 --- a/release/c/wuffs-unsupported-snapshot.c +++ b/release/c/wuffs-unsupported-snapshot.c @@ -3780,10 +3780,12 @@ wuffs_base__pixel_palette__closest_element( // -------- // 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)(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len); typedef struct { // Do not access the private_impl's fields directly. There is no API/ABI @@ -10398,7 +10400,7 @@ wuffs_base__pixel_buffer__color_u32_at(const wuffs_base__pixel_buffer* pb, } size_t stride = pb->private_impl.planes[0].stride; - uint8_t* row = pb->private_impl.planes[0].ptr + (stride * ((size_t)y)); + const uint8_t* row = pb->private_impl.planes[0].ptr + (stride * ((size_t)y)); switch (pb->pixcfg.private_impl.pixfmt.repr) { case WUFFS_BASE__PIXEL_FORMAT__BGRA_PREMUL: @@ -10760,7 +10762,7 @@ wuffs_base__pixel_swizzler__squash_bgr_565_888(wuffs_base__slice_u8 dst, wuffs_base__slice_u8 src) { size_t len4 = (dst.len < src.len ? dst.len : src.len) / 4; uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + const uint8_t* s = src.ptr; size_t n = len4; while (n--) { @@ -10782,7 +10784,7 @@ wuffs_base__pixel_swizzler__swap_rgbx_bgrx(wuffs_base__slice_u8 dst, wuffs_base__slice_u8 src) { size_t len4 = (dst.len < src.len ? dst.len : src.len) / 4; uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + const uint8_t* s = src.ptr; size_t n = len4; while (n--) { @@ -10803,34 +10805,47 @@ wuffs_base__pixel_swizzler__swap_rgbx_bgrx(wuffs_base__slice_u8 dst, // -------- static uint64_t // -wuffs_base__pixel_swizzler__copy_1_1(wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - 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__slice_u8 src) { - size_t dst_len3 = dst.len / 3; - size_t src_len3 = src.len / 3; - size_t len = dst_len3 < src_len3 ? dst_len3 : src_len3; +wuffs_base__pixel_swizzler__copy_1_1(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t len = (dst_len < src_len) ? dst_len : src_len; if (len > 0) { - memmove(dst.ptr, src.ptr, len * 3); + memmove(dst_ptr, src_ptr, len); } return len; } static uint64_t // -wuffs_base__pixel_swizzler__copy_4_4(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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; +wuffs_base__pixel_swizzler__copy_3_3(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len3 = dst_len / 3; + size_t src_len3 = src_len / 3; + size_t len = (dst_len3 < src_len3) ? dst_len3 : src_len3; if (len > 0) { - memmove(dst.ptr, src.ptr, len * 4); + memmove(dst_ptr, src_ptr, len * 3); + } + return len; +} + +static uint64_t // +wuffs_base__pixel_swizzler__copy_4_4(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + if (len > 0) { + memmove(dst_ptr, src_ptr, len * 4); } return len; } @@ -10838,14 +10853,17 @@ wuffs_base__pixel_swizzler__copy_4_4(wuffs_base__slice_u8 dst, // -------- static uint64_t // -wuffs_base__pixel_swizzler__bgr_565__bgr(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_len3 = src.len / 3; - size_t len = dst_len2 < src_len3 ? dst_len2 : src_len3; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__bgr_565__bgr(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t src_len3 = src_len / 3; + size_t len = (dst_len2 < src_len3) ? dst_len2 : src_len3; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -10867,14 +10885,17 @@ wuffs_base__pixel_swizzler__bgr_565__bgr(wuffs_base__slice_u8 dst, static uint64_t // wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src( - 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; - size_t len = dst_len2 < src_len4 ? dst_len2 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t src_len4 = src_len / 4; + size_t len = (dst_len2 < src_len4) ? dst_len2 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -10896,14 +10917,17 @@ wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src( static uint64_t // wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over( - 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; - size_t len = dst_len2 < src_len4 ? dst_len2 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t src_len4 = src_len / 4; + size_t len = (dst_len2 < src_len4) ? dst_len2 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -10949,13 +10973,16 @@ wuffs_base__pixel_swizzler__bgr_565__bgra_nonpremul__src_over( } static uint64_t // -wuffs_base__pixel_swizzler__bgr_565__y(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 len = dst_len2 < src.len ? dst_len2 : src.len; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__bgr_565__y(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len2 = dst_len / 2; + size_t len = (dst_len2 < src_len) ? dst_len2 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -10975,17 +11002,19 @@ wuffs_base__pixel_swizzler__bgr_565__y(wuffs_base__slice_u8 dst, } 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) { +wuffs_base__pixel_swizzler__bgr_565__index__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len2 = dst_len / 2; + size_t len = (dst_len2 < src_len) ? dst_len2 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; const size_t loop_unroll_count = 4; @@ -10993,16 +11022,16 @@ wuffs_base__pixel_swizzler__bgr_565__index__src( 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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((size_t)s[3] * 4))); s += loop_unroll_count * 1; d += loop_unroll_count * 2; @@ -11012,7 +11041,7 @@ wuffs_base__pixel_swizzler__bgr_565__index__src( 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))); + dst_palette_ptr + ((size_t)s[0] * 4))); s += 1 * 1; d += 1 * 2; @@ -11024,22 +11053,25 @@ wuffs_base__pixel_swizzler__bgr_565__index__src( static uint64_t // wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over( - wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len2 = dst_len / 2; + size_t len = (dst_len2 < src_len) ? dst_len2 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // 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(dst_palette_ptr + ((size_t)s[0] * 4)); if (s0) { wuffs_base__store_u16le__no_bounds_check(d + (0 * 2), (uint16_t)s0); @@ -11056,15 +11088,17 @@ wuffs_base__pixel_swizzler__bgr_565__index_binary_alpha__src_over( // -------- 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) { - 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; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11085,14 +11119,17 @@ wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src( static uint64_t // wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over( - 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; - size_t len = dst_len3 < src_len4 ? dst_len3 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11132,14 +11169,17 @@ wuffs_base__pixel_swizzler__bgr__bgra_nonpremul__src_over( static uint64_t // wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over( - 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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11163,14 +11203,17 @@ wuffs_base__pixel_swizzler__bgra_nonpremul__bgra_nonpremul__src_over( static uint64_t // wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src( - 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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11191,14 +11234,17 @@ wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src( static uint64_t // wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over( - 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; - size_t len = dst_len4 < src_len4 ? dst_len4 : src_len4; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len4 = src_len / 4; + size_t len = (dst_len4 < src_len4) ? dst_len4 : src_len4; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11220,16 +11266,19 @@ wuffs_base__pixel_swizzler__bgra_premul__bgra_nonpremul__src_over( // -------- static uint64_t // -wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { +wuffs_base__pixel_swizzler__xxx__index__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len3 = dst_len / 3; + size_t len = (dst_len3 < src_len) ? dst_len3 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; const size_t loop_unroll_count = 4; @@ -11244,16 +11293,16 @@ wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, 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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((size_t)s[3] * 4))); s += loop_unroll_count * 1; d += loop_unroll_count * 3; @@ -11261,7 +11310,7 @@ wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, } 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(dst_palette_ptr + ((size_t)s[0] * 4)); wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0); @@ -11275,37 +11324,40 @@ wuffs_base__pixel_swizzler__xxx__index__src(wuffs_base__slice_u8 dst, static uint64_t // wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over( - wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len3 = dst_len / 3; + size_t len = (dst_len3 < src_len) ? dst_len3 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; 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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((size_t)s[3] * 4)); if (s3) { wuffs_base__store_u24le__no_bounds_check(d + (3 * 3), s3); @@ -11317,7 +11369,7 @@ wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over( } 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(dst_palette_ptr + ((size_t)s[0] * 4)); if (s0) { wuffs_base__store_u24le__no_bounds_check(d + (0 * 3), s0); @@ -11332,13 +11384,16 @@ wuffs_base__pixel_swizzler__xxx__index_binary_alpha__src_over( } static uint64_t // -wuffs_base__pixel_swizzler__xxx__y(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 len = dst_len3 < src.len ? dst_len3 : src.len; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__xxx__y(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len3 = dst_len / 3; + size_t len = (dst_len3 < src_len) ? dst_len3 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11360,16 +11415,19 @@ wuffs_base__pixel_swizzler__xxx__y(wuffs_base__slice_u8 dst, // -------- static uint64_t // -wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { +wuffs_base__pixel_swizzler__xxxx__index__src(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len4 = dst_len / 4; + size_t len = (dst_len4 < src_len) ? dst_len4 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; const size_t loop_unroll_count = 4; @@ -11377,16 +11435,16 @@ wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, 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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((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))); + dst_palette_ptr + ((size_t)s[3] * 4))); s += loop_unroll_count * 1; d += loop_unroll_count * 4; @@ -11396,7 +11454,7 @@ wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, 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))); + dst_palette_ptr + ((size_t)s[0] * 4))); s += 1 * 1; d += 1 * 4; @@ -11408,37 +11466,40 @@ wuffs_base__pixel_swizzler__xxxx__index__src(wuffs_base__slice_u8 dst, static uint64_t // wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over( - wuffs_base__slice_u8 dst, - wuffs_base__slice_u8 dst_palette, - wuffs_base__slice_u8 src) { - if (dst_palette.len != 1024) { + uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + 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; - uint8_t* s = src.ptr; + size_t dst_len4 = dst_len / 4; + size_t len = (dst_len4 < src_len) ? dst_len4 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; 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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((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(dst_palette_ptr + ((size_t)s[3] * 4)); if (s3) { wuffs_base__store_u32le__no_bounds_check(d + (3 * 4), s3); @@ -11450,7 +11511,7 @@ wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over( } 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(dst_palette_ptr + ((size_t)s[0] * 4)); if (s0) { wuffs_base__store_u32le__no_bounds_check(d + (0 * 4), s0); @@ -11465,14 +11526,17 @@ wuffs_base__pixel_swizzler__xxxx__index_binary_alpha__src_over( } static uint64_t // -wuffs_base__pixel_swizzler__xxxx__xxx(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_len3 = src.len / 3; - size_t len = dst_len4 < src_len3 ? dst_len4 : src_len3; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__xxxx__xxx(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t src_len3 = src_len / 3; + size_t len = (dst_len4 < src_len3) ? dst_len4 : src_len3; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11491,13 +11555,16 @@ wuffs_base__pixel_swizzler__xxxx__xxx(wuffs_base__slice_u8 dst, } static uint64_t // -wuffs_base__pixel_swizzler__xxxx__y(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 len = dst_len4 < src.len ? dst_len4 : src.len; - uint8_t* d = dst.ptr; - uint8_t* s = src.ptr; +wuffs_base__pixel_swizzler__xxxx__y(uint8_t* dst_ptr, + size_t dst_len, + uint8_t* dst_palette_ptr, + size_t dst_palette_len, + const uint8_t* src_ptr, + size_t src_len) { + size_t dst_len4 = dst_len / 4; + size_t len = (dst_len4 < src_len) ? dst_len4 : src_len; + uint8_t* d = dst_ptr; + const uint8_t* s = src_ptr; size_t n = len; // TODO: unroll. @@ -11777,7 +11844,8 @@ wuffs_base__pixel_swizzler__swizzle_interleaved( 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); + return (*p->private_impl.func)(dst.ptr, dst.len, dst_palette.ptr, + dst_palette.len, src.ptr, src.len); } return 0; }