mirror of
https://github.com/ddidderr/wuffs.git
synced 2026-11-22 00:47:28 +01:00
This is the first milestone of a byte-exact rewrite of brotli (reference implementation tag v1.2.1) in Wuffs. The decoder mirrors c/dec/decode.c function by function, including its validation checks: the same output for valid input and the same verdict, with each BROTLI_DECODER_ERROR_FORMAT_* code mapped to a Wuffs status (see std/brotli/README.md). The C state machine is replaced by Wuffs coroutines, so decoding works with arbitrarily small src and dst buffers. The history window lives in the caller's workbuf. Huffman table groups, context maps and the static dictionary scratch space live in the decoder struct. The static dictionary, transforms, context lookup tables and prefix code tables in common_tables.wuffs are generated from the brotli source by script/print-brotli-tables.go. Tests: golden tests on the upstream tests/testdata vectors and on streams produced by the v1.2.1 CLI at several qualities and window sizes, 1-byte I/O limits, truncation at every prefix, malformed inputs per error code, trailing data and workbuf boundaries. With -mimic, every vector and every malformed input is also compared against libbrotlidec, including the exact error code. A differential run over about a million mutated inputs found no divergence. Not yet supported (left as TODOs): large-window brotli, custom/shared dictionaries, metadata callbacks and fast paths. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_011NoosQNe7DKWnqDCBeqw1p
279 lines
10 KiB
Markdown
279 lines
10 KiB
Markdown
# Changelog
|
|
|
|
|
|
## Work In Progress
|
|
|
|
The LICENSE has changed from a single license (Apache 2) to a dual license
|
|
(Apache 2 or MIT, at your option).
|
|
|
|
- Added `base.bitvec256`.
|
|
- Added `base.optional_u63`.
|
|
- Added `base.hasher_bitvec256`.
|
|
- Added `base.hasher_u32` `update!` and `checksum_u32` methods.
|
|
- Added `base.hasher_u64`.
|
|
- Added `base.range_ie_i32`.
|
|
- Added `base.rect_ie_i32`.
|
|
- Added `compact_retaining` and `dst_history_retain_length`.
|
|
- Added `example/toy-aux-image`.
|
|
- Added `example/mzcat`.
|
|
- Added `get_quirk(key: u32) u64`.
|
|
- Added `std/brotli`.
|
|
- Added `std/crc64`.
|
|
- Added `std/etc2`.
|
|
- Added `std/handsum`.
|
|
- Added `std/jpeg`.
|
|
- Added `std/lzip`.
|
|
- Added `std/lzma`.
|
|
- Added `std/netpbm`.
|
|
- Added `std/qoi`.
|
|
- Added `std/sha256`.
|
|
- Added `std/thumbhash`.
|
|
- Added `std/vp8`.
|
|
- Added `std/webp`.
|
|
- Added `std/xxhash32`.
|
|
- Added `std/xxhash64`.
|
|
- Added `std/xz`.
|
|
- Added `WUFFS_BASE__QUIRK_QUALITY`.
|
|
- Added `WUFFS_CONFIG__DISABLE_MSVC_CPU_ARCH__X86_64_FAMILY`.
|
|
- Added `WUFFS_CONFIG__DST_PIXEL_FORMAT__ENABLE_ALLOWLIST`.
|
|
- Added `WUFFS_CONFIG__ENABLE_DROP_IN_REPLACEMENT__STB`.
|
|
- Added `WUFFS_CONFIG__ENABLE_MSVC_CPU_ARCH__X86_64_V2`.
|
|
- Added `WUFFS_CONFIG__ENABLE_MSVC_CPU_ARCH__X86_64_V3`.
|
|
- Added `wuffs_base__status__is_truncated_input_error`.
|
|
- Changed `lzw.set_literal_width` to `lzw.set_quirk`.
|
|
- Changed `set_quirk_enabled!(quirk: u32, enabled: bool)` to `set_quirk!(key:
|
|
u32, value: u64) status`.
|
|
- Deprecated `std/lzw.decoder.flush`.
|
|
- Fixed `PIXEL_FORMAT__YA_{NON,}PREMUL` constant values.
|
|
- Generated constants now default to unsigned.
|
|
- Halved the sizeof `wuffs_foo__bar::unique_ptr`.
|
|
- Let `std/png` decode PNG color type 4 to `PIXEL_FORMAT__YA_NONPREMUL` (two
|
|
channels) instead of `PIXEL_FORMAT__BGRA_NONPREMUL` (four channels).
|
|
- Reassigned `lib/base38` alphabet and numbers.
|
|
- Removed the `std/gif -> std/lzw` dependency.
|
|
- Removed `endwhile` keyword.
|
|
- Removed `example/bzcat`.
|
|
- Renamed `std/tga` to `std/targa`.
|
|
- Set image decoder pixel width and height inclusive maximum to `0xFF_FFFF`,
|
|
down from `0x7FFF_FFFF`.
|
|
|
|
The dot points below probably aren't of interest unless you're _writing_ Wuffs
|
|
code (instead of writing C/C++ code that _uses_ Wuffs' standard library).
|
|
|
|
- Added `if.likely` and `if.unlikely`.
|
|
- Added `io_forget_history`.
|
|
- Added `slice_var as nptr array[etc] etc` conversion.
|
|
- Added read-only type decorators: `roarray`, `roslice` and `rotable`.
|
|
- Banned recursive function calls.
|
|
- Renamed `base` `min/max` argument from `a` to `no_more/less_than`.
|
|
- Wuffs struct private data now needs a "+" between the "()" pairs.
|
|
- `wuffsfmt` double-indents hanging lines and each indent is now 4 spaces (not
|
|
a tab).
|
|
|
|
|
|
## 2026-10-20 version 0.3.6
|
|
|
|
- Fix some `parse_number_f64` and `parse_number_u64` bugs.
|
|
- Fix "go test" example names for newer Go toolchains.
|
|
|
|
|
|
## 2026-07-11 version 0.3.5
|
|
|
|
- No significant changes to `wuffs-v0.3.c` per se.
|
|
- Some test files were renamed, to work better with pkg.go.dev indexing.
|
|
- Some test code had #include lines added, to fix some third party builds.
|
|
|
|
|
|
## 2024-04-19 version 0.3.4
|
|
|
|
- Have `std/png` ignore tRNS chunks for color types 4 (YA) and 6 (RGBA).
|
|
|
|
|
|
## 2023-04-08 version 0.3.3
|
|
|
|
The `wuffs_base__parse_number_f64` function has been further optimized.
|
|
|
|
|
|
## 2023-04-07 version 0.3.2
|
|
|
|
The `std/bmp` and `std/nie` image decoders' `decode_frame` method now allow
|
|
decoding to a pixel buffer that's smaller than the source image. This makes
|
|
these two image decoders consistent with the other ones in `std`.
|
|
|
|
|
|
## 2023-03-04 version 0.3.1
|
|
|
|
For a *closed* `io_reader`, the standard library now returns `"#truncated
|
|
input"` instead of `"$short read"`. Importantly, this is an error, not a
|
|
suspension.
|
|
|
|
The `wuffs_base__parse_number_f64` function's Simple Decimal Conversion
|
|
fallback algorithm has been optimized.
|
|
|
|
|
|
## 2023-01-26 version 0.3.0
|
|
|
|
The headline feature is that we have a production quality PNG decoder. It's
|
|
also [the fastest, safest PNG decoder in the
|
|
world](https://nigeltao.github.io/blog/2021/fastest-safest-png-decoder.html).
|
|
There's also a [memory-safe, zero-allocation JSON
|
|
decoder](https://nigeltao.github.io/blog/2020/jsonptr.html).
|
|
|
|
The dot points below probably aren't of interest unless you're upgrading from
|
|
Wuffs version 0.2.
|
|
|
|
- Added `0b` prefixed binary numbers.
|
|
- Added `WUFFS_BASE__PIXEL_BLEND__SRC_OVER`.
|
|
- Added `WUFFS_BASE__PIXEL_FORMAT__BGR_565`.
|
|
- Added `WUFFS_CONFIG__MODULE__BASE__ETC` sub-modules.
|
|
- Added `auxiliary` code.
|
|
- Added `base` library support for UTF-8.
|
|
- Added `base` library support for `atoi`-like string conversion.
|
|
- Added `choose` and `choosy`.
|
|
- Added `cpu_arch`.
|
|
- Added `doc/logo`.
|
|
- Added `endwhile` syntax.
|
|
- Added `example/bzcat`.
|
|
- Added `example/cbor-to-json`.
|
|
- Added `example/convert-to-nia`.
|
|
- Added `example/imageviewer`.
|
|
- Added `example/json-to-cbor`.
|
|
- Added `example/jsonfindptrs`.
|
|
- Added `example/jsonptr`.
|
|
- Added `example/sdl-imageviewer`.
|
|
- Added `slice base.u8 peek/poke` methods.
|
|
- Added `std/bmp`.
|
|
- Added `std/bzip2`.
|
|
- Added `std/cbor`.
|
|
- Added `std/json`.
|
|
- Added `std/nie`.
|
|
- Added `std/png`.
|
|
- Added `std/tga`.
|
|
- Added `std/wbmp`.
|
|
- Added `tell_me_more?` mechanism.
|
|
- Added SIMD.
|
|
- Added alloc functions.
|
|
- Added colons to const syntax.
|
|
- Added double-curly blocks.
|
|
- Added interfaces.
|
|
- Added iterate advance parameter.
|
|
- Added single-quoted strings.
|
|
- Added slice `uintptr_low_12_bits` method.
|
|
- Added tokens.
|
|
- Changed `gif.decoder_workbuf_len_max_incl_worst_case` from 1 to 0.
|
|
- Changed default C compilers from `clang-5.0,gcc` to `clang,gcc`.
|
|
- Changed the C formatting style; removed the `-cformatter` flag.
|
|
- Changed what the `std/gif` benchmarks actually measure.
|
|
- Made `wuffs_base__pixel_format` a struct.
|
|
- Made `wuffs_base__pixel_subsampling` a struct.
|
|
- Made `wuffs_base__status` a struct.
|
|
- Prohibited iterate loops inside coroutines.
|
|
- Removed `ack_metadata_chunk?`.
|
|
- Removed `wuffs_base__frame_config__blend`.
|
|
- Renamed I/O `available` methods to `length`.
|
|
- Renamed `decode_io_writer?` methods to `transform_io?`.
|
|
- Renamed `example/library` to `example/toy-genlib`.
|
|
- Renamed `load` and `store` to `peek` and `poke`.
|
|
- Renamed `{read,writ}er_io_position` to `{read,writ}er_position`.
|
|
- Renamed `set_ignore_checksum!` as a quirk.
|
|
- Renamed `swizzle_interleaved!` to `swizzle_interleaved_from_slice!`.
|
|
- Renamed warnings to notes.
|
|
|
|
|
|
## 2019-12-19 version 0.2.0
|
|
|
|
The headline feature is that the GIF decoder is now of production quality.
|
|
There is now API for overall metadata (e.g. ICCP color profiles) and to
|
|
recreate each frame (width, height, BGRA pixels, timing, etc.) of a GIF
|
|
animation, instead of version 0.1's proof-of-concept GIF decoder API, which
|
|
just gave you a one-dimensional stream of palette indexes. It also now accepts
|
|
a variety of GIF images that are invalid, when strictly following the GIF
|
|
specifiction, but are nonetheless accepted by other real world GIF
|
|
implementations. The Wuffs GIF decoder has also been optimized to be about 1.5x
|
|
faster than Wuffs version 0.1 and about 2x faster than giflib (the C library).
|
|
|
|
The Wuffs GIF decoder is being trialled by Skia, the 2-D graphics library used
|
|
by both the Android operating system and the Chromium web browser.
|
|
|
|
Work also proceeds on the NIE and RAC file formats, but both are still
|
|
experimental and may change later in backwards incompatible ways.
|
|
|
|
The dot points below probably aren't of interest unless you're upgrading from
|
|
Wuffs version 0.1.
|
|
|
|
- Renamed Puffs to Wuffs.
|
|
- Ship as a "single file C library"
|
|
- Added a `skipgendeps` flag.
|
|
- Added a `nullptr` literal and `nptr T` type.
|
|
- Added `io_bind` and `io_limit` keywords.
|
|
- Added a `use` keyword.
|
|
- Added a `yield` keyword.
|
|
- Made the `return` value mandatory; added `ok` literal.
|
|
- Restricted `var` statements to the top of functions.
|
|
- Dropped the `= RHS` out of `var x T = RHS`.
|
|
- Changed func out-type from struct to bare type.
|
|
- Renamed the implicit `in` variable to `args`.
|
|
- Added an implicit `coroutine_resumed` variable.
|
|
- Added `std/adler32`, `std/crc32` and `std/gzip`.
|
|
- Added `std/gif` quirks.
|
|
- Spun `std/lzw` out of `std/gif`.
|
|
- Spun `std/zlib` out of `std/flate`.
|
|
- Let the `std/gzip` and `std/zlib` decoder ignore checksums.
|
|
- Renamed `std/flate` to `std/deflate`.
|
|
- Renamed `!=` to `<>`; `!` is now only for impure functions.
|
|
- Renamed `~+` to `~mod+`; added `~mod-`, `~sat+` and `~sat-`.
|
|
- Removed `&^`.
|
|
- Renamed `$(etc)` to `[etc]`.
|
|
- Renamed `[i..j]` to `[i ..= j]`, consistent with Rust syntax.
|
|
- Renamed `[i:j]` to `[i .. j]`, consistent with Rust syntax.
|
|
- Renamed `x T` to `x: T`, consistent with Rust syntax.
|
|
- Renamed `[N] T` and `[] T` types to `array[N] T` and `slice T`.
|
|
- Renamed `while:label` to `while.label`.
|
|
- Renamed `u32`, `buf1`, etc to `base.u32`, `base.io_buffer`, etc.
|
|
- Renamed `unread_u8?` to `undo_byte!`; added `can_undo_byte`.
|
|
- Renamed some `decode?` methods to `decode_io_writer?`.
|
|
- Replaced `= try foo` with `=? foo`.
|
|
- Prohibited effect-ful subexpressions.
|
|
- Redesigned iterate blocks.
|
|
- Added `{frame,image,pixel}_config` and `pixel_buffer` types.
|
|
- Added a `reset` method.
|
|
- Added `peek_uxx`, `skip_fast` and `write_fast_uxx` methods.
|
|
- Renamed some `read_uxx` methods as `read_uxx_as_uyy`.
|
|
- Report image metadata such as ICCP and XMP.
|
|
- Added I/O positions.
|
|
- Tweaked how I/O marks and limits work.
|
|
- Tweaked `io_buffer` / `io_reader` distinction in C and Wuffs.
|
|
- Added extra fields (uninitialized internal buffers) to structs.
|
|
- Supported animated (not just single frame) and interlaced GIFs.
|
|
- Marked some internal status codes as private.
|
|
- Removed closed-for-read/write built-in status codes.
|
|
- Changed the string messages for built-in status codes.
|
|
- Changed `error "foo"` to `"#foo"` or `base."#bar"`.
|
|
- Added warnings as another status code category.
|
|
- Made the status type a `const char *`, not an `int32_t`.
|
|
- Disallowed `__double_underscore` prefixed names.
|
|
- Added fuzz tests.
|
|
- Added `WUFFS_CONFIG__MODULES`.
|
|
- Added `WUFFS_CONFIG__STATIC_FUNCTIONS`.
|
|
- Added some C++ convenience methods.
|
|
- Added some Go and Rust benchmarks.
|
|
- Made struct implementations private (opaque pointers).
|
|
- Sped up the `mimic_deflate_xxx` benchmarks.
|
|
- Fix compile errors under MSVC (Microsoft Visual C/C++).
|
|
- Moved the base38 package from `lang/base38` to `lib/base38`.
|
|
- Added a stand-alone `lib/interval` package.
|
|
- Added NIE file format spec.
|
|
- Added RAC file format spec and Go implementation.
|
|
|
|
|
|
## 2017-11-16 version 0.1.0
|
|
|
|
- [Initial open source
|
|
release](https://groups.google.com/d/topic/puffslang/2z61mNTAMns/discussion),
|
|
under the "Puffs" name.
|
|
|
|
|
|
---
|
|
|
|
Updated on April 2023.
|