Add TIFF decoding support

codecs/tiff/.dockerignore

@@ -0,0 +1,2 @@
+**
+!/patches

codecs/tiff/.gitignore

@@ -0,0 +1,2 @@
+compile_flags.txt
+*.wat

codecs/tiff/Makefile

@@ -0,0 +1,160 @@
+CODEC_URL = https://download.osgeo.org/libtiff/tiff-4.6.0.tar.xz
+CODEC_PACKAGE = node_modules/tiff.tar.xz
+CODEC_PACKAGE_HASH = e178649607d1e22b51cf361dd20a3753f244f022eefab1f2f218fc62ebaf87d2
+
+ZLIB_URL = https://github.com/madler/zlib/releases/download/v1.3.1/zlib-1.3.1.tar.xz
+ZLIB_PACKAGE = node_modules/zlib.tar.xz
+ZLIB_PACKAGE_HASH = 38ef96b8dfe510d42707d9c781877914792541133e1870841463bfa73f883e32
+
+JPEG_URL = https://github.com/libjpeg-turbo/libjpeg-turbo/releases/download/3.0.2/libjpeg-turbo-3.0.2.tar.gz
+JPEG_PACKAGE = node_modules/libjpeg-turbo.tar.gz
+JPEG_PACKAGE_HASH = c2ce515a78d91b09023773ef2770d6b0df77d674e144de80d63e0389b3a15ca6
+
+export CODEC_DIR = node_modules/tiff
+export ZLIB_DIR = node_modules/zlib
+export JPEG_DIR = node_modules/jpeg
+
+.PHONY all: clean_binary dec/tiff_dec.wasm
+
+# ===============================================================
+# zlib
+
+$(ZLIB_PACKAGE):
+	mkdir -p "$(@D)"
+	curl -sSL "$(ZLIB_URL)" -o "$@"
+	echo "$(ZLIB_PACKAGE_HASH)  $(ZLIB_PACKAGE)" | sha256sum -c -
+
+$(ZLIB_DIR)/CMakeLists.txt: $(ZLIB_PACKAGE)
+	mkdir -p "$(@D)"
+	tar xJm --strip 1 -C "$(@D)" -f "$(ZLIB_PACKAGE)"
+# remove dummy Makefile
+	rm -f "$(ZLIB_DIR)/Makefile"
+
+$(ZLIB_DIR)/Makefile: $(ZLIB_DIR)/CMakeLists.txt
+	cmake -S "$(ZLIB_DIR)" -B "$(ZLIB_DIR)" \
+		-DCMAKE_BUILD_TYPE=MinSizeRel \
+		-DBUILD_SHARED_LIBS=OFF
+
+ZLIB_LIB=$(ZLIB_DIR)/libz.a
+
+$(ZLIB_LIB): $(ZLIB_DIR)/Makefile
+	$(MAKE) -C "$(ZLIB_DIR)" zlibstatic
+	mv -f "$(ZLIB_DIR)/libzlibstatic.a" "$(ZLIB_LIB)"
+
+# ===============================================================
+# libjpeg-turbo
+
+$(JPEG_PACKAGE):
+	mkdir -p "$(@D)"
+	curl -sSL "$(JPEG_URL)" -o "$@"
+	echo "$(JPEG_PACKAGE_HASH)  $(JPEG_PACKAGE)" | sha256sum -c -
+
+$(JPEG_DIR)/CMakeLists.txt: $(JPEG_PACKAGE)
+	mkdir -p "$(@D)"
+	tar xzm --strip 1 -C "$(@D)" -f "$(JPEG_PACKAGE)"
+	for i in ./patches/libjpeg-turbo/*.patch; do patch -d "$(@D)" -N -p1 < "$$i"; done
+
+$(JPEG_DIR)/Makefile: $(JPEG_DIR)/CMakeLists.txt
+	cmake -S "$(JPEG_DIR)" -B "$(JPEG_DIR)" \
+		-DCMAKE_BUILD_TYPE=MinSizeRel \
+		-DBUILD_SHARED_LIBS=OFF \
+		-DENABLE_SHARED=OFF \
+		-DWITH_SIMD=OFF \
+		-DWITH_TURBOJPEG=OFF
+
+JPEG_LIB=$(JPEG_DIR)/libjpeg.a
+
+$(JPEG_LIB): $(JPEG_DIR)/Makefile
+	$(MAKE) -C "$(JPEG_DIR)" jpeg-static
+
+# ===============================================================
+# libtiff
+
+$(CODEC_PACKAGE):
+	mkdir -p "$(@D)"
+	curl -sSL "$(CODEC_URL)" -o "$@"
+	echo "$(CODEC_PACKAGE_HASH)  $(CODEC_PACKAGE)" | sha256sum -c -
+
+$(CODEC_DIR)/configure: $(CODEC_PACKAGE)
+	mkdir -p "$(@D)"
+	tar xJm --strip 1 -C "$(@D)" -f "$(CODEC_PACKAGE)"
+	for i in ./patches/libtiff/*.patch; do patch -d "$(@D)" -N -p1 < "$$i"; done
+
+# It's a huge pain telling CMake to find our custom-built libraries, use autoconf instead...
+$(CODEC_DIR)/Makefile: $(CODEC_DIR)/configure
+	cd $(CODEC_DIR); \
+	./configure \
+		--host=wasm32-wasi \
+		--enable-shared=no \
+		--enable-cxx=no \
+		--disable-tools \
+		--disable-tests \
+		--disable-contrib \
+		--disable-docs \
+		--disable-thunder \
+		--disable-next \
+		--disable-mdi \
+		--disable-old-jpeg \
+		--disable-jbig \
+		--disable-lerc \
+		--disable-lzma \
+		--disable-zstd \
+		--disable-webp \
+		--disable-sphinx \
+		--with-zlib-include-dir="$(shell realpath $(ZLIB_DIR))" \
+		--with-zlib-lib-dir="$(shell realpath $(ZLIB_DIR))" \
+		--with-jpeg-include-dir="$(shell realpath $(JPEG_DIR))" \
+		--with-jpeg-lib-dir="$(shell realpath $(JPEG_DIR))"
+
+CODEC_LIB=$(CODEC_DIR)/libtiff/.libs/libtiff.a
+
+$(CODEC_LIB): $(CODEC_DIR)/Makefile
+	$(MAKE) -C "$(CODEC_DIR)"
+
+# ===============================================================
+
+dec/tiff_dec.wasm: $(ZLIB_LIB) $(JPEG_LIB) $(CODEC_LIB)
+	$(CC) \
+		$(CFLAGS) \
+		-Wall \
+		-Wextra \
+		-Wconversion \
+		-I"$(CODEC_DIR)/libtiff" \
+		-g \
+		-Wl,-z,stack-size=1048576 \
+		-Wl,--fatal-warnings \
+		-Wl,--no-entry \
+		-Wl,--export=malloc \
+		-Wl,--export=free \
+		-Wl,--trace-symbol=stderr \
+		-mexec-model=reactor \
+		-o dec/tiff_dec.wasm \
+		dec/tiff_dec.c \
+		"$(CODEC_LIB)" \
+		"$(JPEG_LIB)" \
+		"$(ZLIB_LIB)" \
+		-lm
+
+# ===============================================================
+
+distclean_zlib:
+	rm -rf "$(ZLIB_DIR)"
+
+distclean_jpeg:
+	rm -rf "$(JPEG_DIR)"
+
+distclean_codec:
+	rm -rf "$(CODEC_DIR)"
+
+.PHONY distclean: distclean_zlib distclean_jpeg distclean_codec
+
+clean_jpeg:
+	[ -f "$(JPEG_DIR)/Makefile" ] && $(MAKE) -C "$(JPEG_DIR)" clean
+
+clean_codec:
+	[ -f "$(CODEC_DIR)/Makefile" ] && $(MAKE) -C "$(CODEC_DIR)" clean
+
+clean_binary:
+	rm -f dec/tiff_dec.wasm
+
+.PHONY clean: clean_binary clean_codec clean_jpeg

codecs/tiff/build.sh

@@ -0,0 +1,8 @@
+#!/bin/bash
+set -euo pipefail
+
+docker build -t squoosh-wasi --network=host -f wasi.Dockerfile .
+docker run -it --rm -v "$(pwd)":/src -u "$(id -u):$(id -g)" squoosh-wasi "$@"
+
+# wasm2wat --enable-annotation --enable-code-metadata dec/tiff_dec.wasm > dec/tiff_dec.wat || true
+wasm-opt -O3 --strip-debug -o dec/tiff_dec.wasm dec/tiff_dec.wasm

codecs/tiff/dec/tiff_dec.c

@@ -0,0 +1,195 @@
+#include <errno.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+#include <tiffio.h>
+
+typedef struct cursor {
+  const char* data;
+  tmsize_t size;
+  tmsize_t off;
+} cursor_t;
+
+typedef union cd_as_cursor {
+  thandle_t client;
+  cursor_t* cursor;
+} cd_as_cursor_t;
+
+static tmsize_t _cursor_read_proc(thandle_t clientdata, void* buf, tmsize_t len) {
+  cd_as_cursor_t cdc;
+  cdc.client = clientdata;
+  cursor_t* c = cdc.cursor;
+
+  tmsize_t nremain = c->size - c->off;
+  tmsize_t nread = nremain > len ? len : nremain;
+
+  memcpy(buf, (const void*)(c->data + c->off), (size_t)nread);
+  c->off += nread;
+  return (tmsize_t)nread;
+}
+
+static tmsize_t _cursor_write_proc(thandle_t clientdata, void* buf, tmsize_t size) {
+  (void)clientdata, (void)buf, (void)size;
+  return 0;
+}
+
+static toff_t _cursor_seek_proc(thandle_t clientdata, toff_t off, int whence) {
+  cd_as_cursor_t cdc;
+  cdc.client = clientdata;
+  cursor_t* c = cdc.cursor;
+
+  tmsize_t off_m = (tmsize_t)off;
+  if ((toff_t)off_m != off) {
+    errno = EINVAL;
+    return (toff_t)-1;
+  }
+
+  switch (whence) {
+    case SEEK_SET:
+      c->off = off_m;
+      break;
+    case SEEK_CUR:
+      c->off += off_m;
+      break;
+    case SEEK_END:
+      if (off_m <= c->size) {
+        c->off = c->size - off_m;
+        break;
+      }
+      __attribute__((fallthrough));
+    default:
+      errno = EINVAL;
+      return (toff_t)-1;
+  }
+  return (toff_t)c->off;
+}
+
+static int _cursor_close_proc(thandle_t clientdata) {
+  cd_as_cursor_t cdc;
+  cdc.client = clientdata;
+  free(cdc.cursor);
+  return 0;
+}
+
+static uint64_t _cursor_size_proc(thandle_t clientdata) {
+  cd_as_cursor_t cdc;
+  cdc.client = clientdata;
+  return (uint64_t)cdc.cursor->size;
+}
+
+static int _cursor_map_proc(thandle_t clientdata, void** pbase, toff_t* psize) {
+  cd_as_cursor_t cdc;
+  cdc.client = clientdata;
+  cursor_t* c = cdc.cursor;
+
+  // `tif_unix.c` only passes `PROT_READ` to `mmap()`, so dropping `const` should be fine here.
+  *pbase = (void*)c->data;
+  *psize = (toff_t)c->size;
+  return 1;
+}
+
+static void _cursor_unmap_proc(thandle_t clientdata, void* base, toff_t size) {
+  (void)clientdata, (void)base, (void)size;
+}
+
+extern int _return_decoded_image(const uint32_t* raster, uint32_t width, uint32_t height)
+    __attribute__((import_name("return_decoded_image")));
+
+__attribute__((export_name("decode"))) int decode(const char* data, const size_t size) {
+  static const char* module_name = "tiff_dec";
+  if (!data)
+    return 0;
+
+  tmsize_t size_m = (tmsize_t)size;
+  if (((size_t)size_m) != size) {
+    TIFFErrorExtR(NULL, module_name, "Too large image buffer");
+    return 0;
+  }
+
+  cursor_t* c = malloc(sizeof(cursor_t));
+  if (!c)
+    return 0;
+  c->data = data;
+  c->size = size_m;
+  c->off = 0;
+
+  cd_as_cursor_t cdc;
+  cdc.cursor = c;
+  TIFF* tif = TIFFClientOpen("dummy.tif", "r", cdc.client, _cursor_read_proc, _cursor_write_proc,
+                             _cursor_seek_proc, _cursor_close_proc, _cursor_size_proc,
+                             _cursor_map_proc, _cursor_unmap_proc);
+  if (!tif) {
+    free(cdc.cursor);
+    return 0;
+  }
+
+  int ret = 0;
+
+  uint32_t width, height;
+  if (TIFFGetField(tif, TIFFTAG_IMAGEWIDTH, &width) == 0 ||
+      TIFFGetField(tif, TIFFTAG_IMAGELENGTH, &height) == 0) {
+    TIFFErrorExtR(tif, module_name, "Missing image dimension tag");
+    goto cleanup_tif;
+  }
+
+  uint32_t npixels;
+  if (__builtin_umul_overflow(width, height, &npixels)) {
+    TIFFErrorExtR(tif, module_name, "Too large image dimension");
+    goto cleanup_tif;
+  }
+
+  uint32_t* raster = (uint32_t*)malloc(npixels * sizeof(uint32_t));
+  if (!raster) {
+    TIFFErrorExtR(tif, module_name, "Failed to allocate memory of %d*%d*4 bytes", width, height);
+    goto cleanup_tif;
+  }
+
+  if (!TIFFReadRGBAImageOriented(tif, width, height, raster, 1, ORIENTATION_TOPLEFT)) {
+    goto cleanup_raster;
+  }
+
+  ret = _return_decoded_image(raster, width, height);
+
+cleanup_raster:
+  free(raster);
+cleanup_tif:
+  TIFFClose(tif);
+  return ret;
+}
+
+extern int _log_warning(const char* s, size_t len) __attribute__((import_name("log_warning")));
+
+extern int _log_error(const char* s, size_t len) __attribute__((import_name("log_error")));
+
+// Override unix error handlers from `tiff_unix.c`.
+static void _extern_warning_handler(const char* module, const char* fmt, va_list ap) {
+  (void)module;
+  char buf[256];
+  vsnprintf(buf, sizeof(buf), fmt, ap);
+  _log_warning(buf, strlen(buf));
+}
+TIFFErrorHandler _TIFFwarningHandler = _extern_warning_handler;
+
+static void _extern_error_handler(const char* module, const char* fmt, va_list ap) {
+  (void)module;
+  char buf[256];
+  vsnprintf(buf, sizeof(buf), fmt, ap);
+  _log_error(buf, strlen(buf));
+}
+TIFFErrorHandler _TIFFerrorHandler = _extern_error_handler;
+
+// Stubbing libc to remove `environ_get*` WASI imports.
+char* getenv(const char* name) {
+  (void)name;
+  return NULL;
+}
+
+// Stubbing libc to remove `fd_*` WASI imports.
+_Noreturn void __assert_fail(const char* expr, const char* file, int line, const char* func) {
+  // Modified from
+  // https://github.com/WebAssembly/wasi-libc/blob/wasi-sdk-20/libc-top-half/musl/src/exit/assert.c
+  char buf[256];
+  snprintf(buf, sizeof(buf), "Assertion failed: %s (%s: %s: %d)\n", expr, file, func, line);
+  _log_error(buf, strlen(buf));
+  abort();
+}

codecs/tiff/package.json

@@ -0,0 +1,6 @@
+{
+  "name": "tiff",
+  "scripts": {
+    "build": "./build.sh make"
+  }
+}