feat(rust): migrate high-level runtime paths

Move long-distance matching and high-level decompression from C shims into
Rust. The decoder now owns context, dictionary, parameter, one-shot, and
buffered streaming state while C retains allocation/configuration, legacy,
and trace leaves.

Move CLI parsing, safety policy, and dispatch into a separate Rust static
archive. Keeping it separate prevents library builds from retaining FIO
symbols, while C continues to own file opening, replacement, and I/O.
Program targets now select matching compression/decompression archives.

The remaining C boundary is intentional: high-level compression, optimal
parsing, dictionary building, legacy callbacks, and CLI file I/O still need
migration.

Test Plan:
- cargo test --all-targets (native and i686)
- cargo test --all-targets in rust/cli (native and i686)
- CLI crate compression-only and decompression-only feature tests
- native and i686 fuzzer/zstreamtest runs, plus legacy and dictionary tests
- ZSTD_C_PREDICT and ZSTD_HEAPMODE=0 fuzzer coverage
- library, dynamic-link, and program-target build/round-trip matrix

Refs: rust/README.md
This commit is contained in:
2026-07-11 09:03:41 +02:00
parent 031592da1e
commit fe7e24c770
14 changed files with 6886 additions and 4651 deletions
+76 -8
View File
@@ -27,8 +27,12 @@ include $(LIBZSTD_MK_DIR)/libzstd.mk
CARGO ?= cargo
RUST_DIR := ../rust
RUST_MANIFEST := $(RUST_DIR)/Cargo.toml
RUST_CLI_DIR := $(RUST_DIR)/cli
RUST_CLI_MANIFEST := $(RUST_CLI_DIR)/Cargo.toml
RUST_SOURCES := $(RUST_MANIFEST) $(RUST_DIR)/Cargo.lock \
$(shell find $(RUST_DIR)/src -type f -name '*.rs' -print)
RUST_CLI_SOURCES := $(RUST_CLI_MANIFEST) $(RUST_CLI_DIR)/Cargo.lock \
$(RUST_CLI_DIR)/src/lib.rs $(RUST_DIR)/src/zstd_cli.rs
# Keep Rust's HUF implementation in lockstep with libzstd.mk's C selection.
# Forced modes may arrive as libzstd.mk variables or as direct -D flags in
@@ -66,7 +70,8 @@ RUST_STATICLIB := $(RUST_TARGET_DIR)/release/libzstd_rs.a
RUST_TARGET_32 ?= i686-unknown-linux-gnu
RUST_STATICLIB_32 := $(RUST_TARGET_DIR)/$(RUST_TARGET_32)/release/libzstd_rs.a
RUST_CARGO_FLAGS := --manifest-path $(RUST_MANIFEST) --release \
--target-dir $(RUST_TARGET_DIR)
--target-dir $(RUST_TARGET_DIR) --no-default-features
RUST_CARGO_FLAGS += --features compression,decompression
ifneq ($(RUST_HUF_FEATURE),)
RUST_CARGO_FLAGS += --features $(RUST_HUF_FEATURE)
endif
@@ -77,6 +82,63 @@ $(RUST_STATICLIB): $(RUST_SOURCES)
$(RUST_STATICLIB_32): $(RUST_SOURCES)
$(CARGO) build $(RUST_CARGO_FLAGS) --target $(RUST_TARGET_32)
RUST_CLI_BUILD_CONFIG := cli-c1-d1-$(RUST_BUILD_CONFIG)
RUST_CLI_TARGET_DIR := $(RUST_DIR)/target/$(RUST_CLI_BUILD_CONFIG)
RUST_CLI_STATICLIB := $(RUST_CLI_TARGET_DIR)/release/libzstd_cli_rs.a
RUST_CLI_STATICLIB_32 := $(RUST_CLI_TARGET_DIR)/$(RUST_TARGET_32)/release/libzstd_cli_rs.a
RUST_CLI_CARGO_FLAGS := --manifest-path $(RUST_CLI_MANIFEST) --release \
--target-dir $(RUST_CLI_TARGET_DIR) \
--no-default-features --features compression,decompression
$(RUST_CLI_STATICLIB): $(RUST_CLI_SOURCES)
$(CARGO) build $(RUST_CLI_CARGO_FLAGS)
$(RUST_CLI_STATICLIB_32): $(RUST_CLI_SOURCES)
$(CARGO) build $(RUST_CLI_CARGO_FLAGS) --target $(RUST_TARGET_32)
RUST_DECOMPRESS_BUILD_CONFIG := lib-c0-d1-$(RUST_BUILD_CONFIG)
RUST_DECOMPRESS_TARGET_DIR := $(RUST_DIR)/target/$(RUST_DECOMPRESS_BUILD_CONFIG)
RUST_DECOMPRESS_STATICLIB := $(RUST_DECOMPRESS_TARGET_DIR)/release/libzstd_rs.a
RUST_DECOMPRESS_CARGO_FLAGS := --manifest-path $(RUST_MANIFEST) --release \
--target-dir $(RUST_DECOMPRESS_TARGET_DIR) \
--no-default-features --features decompression
ifneq ($(RUST_HUF_FEATURE),)
RUST_DECOMPRESS_CARGO_FLAGS += --features $(RUST_HUF_FEATURE)
endif
$(RUST_DECOMPRESS_STATICLIB): $(RUST_SOURCES)
$(CARGO) build $(RUST_DECOMPRESS_CARGO_FLAGS)
RUST_DECOMPRESS_CLI_BUILD_CONFIG := cli-c0-d1-$(RUST_BUILD_CONFIG)
RUST_DECOMPRESS_CLI_TARGET_DIR := $(RUST_DIR)/target/$(RUST_DECOMPRESS_CLI_BUILD_CONFIG)
RUST_DECOMPRESS_CLI_STATICLIB := $(RUST_DECOMPRESS_CLI_TARGET_DIR)/release/libzstd_cli_rs.a
RUST_DECOMPRESS_CLI_CARGO_FLAGS := --manifest-path $(RUST_CLI_MANIFEST) --release \
--target-dir $(RUST_DECOMPRESS_CLI_TARGET_DIR) \
--no-default-features --features decompression
$(RUST_DECOMPRESS_CLI_STATICLIB): $(RUST_CLI_SOURCES)
$(CARGO) build $(RUST_DECOMPRESS_CLI_CARGO_FLAGS)
RUST_COMPRESS_BUILD_CONFIG := lib-c1-d0-$(RUST_BUILD_CONFIG)
RUST_COMPRESS_TARGET_DIR := $(RUST_DIR)/target/$(RUST_COMPRESS_BUILD_CONFIG)
RUST_COMPRESS_STATICLIB := $(RUST_COMPRESS_TARGET_DIR)/release/libzstd_rs.a
RUST_COMPRESS_CARGO_FLAGS := --manifest-path $(RUST_MANIFEST) --release \
--target-dir $(RUST_COMPRESS_TARGET_DIR) \
--no-default-features --features compression
$(RUST_COMPRESS_STATICLIB): $(RUST_SOURCES)
$(CARGO) build $(RUST_COMPRESS_CARGO_FLAGS)
RUST_COMPRESS_CLI_BUILD_CONFIG := cli-c1-d0-$(RUST_BUILD_CONFIG)
RUST_COMPRESS_CLI_TARGET_DIR := $(RUST_DIR)/target/$(RUST_COMPRESS_CLI_BUILD_CONFIG)
RUST_COMPRESS_CLI_STATICLIB := $(RUST_COMPRESS_CLI_TARGET_DIR)/release/libzstd_cli_rs.a
RUST_COMPRESS_CLI_CARGO_FLAGS := --manifest-path $(RUST_CLI_MANIFEST) --release \
--target-dir $(RUST_COMPRESS_CLI_TARGET_DIR) \
--no-default-features --features compression
$(RUST_COMPRESS_CLI_STATICLIB): $(RUST_CLI_SOURCES)
$(CARGO) build $(RUST_COMPRESS_CLI_CARGO_FLAGS)
# Most program objects use a configuration-hashed directory, but the compact
# direct-source variants below do not. Give their C outputs an independent
# mode stamp so they cannot outlive a different Rust HUF archive selection.
@@ -192,6 +254,9 @@ ifeq ($(BACKTRACE), 1)
endif
endif
RUST_LIBRARY_LINK ?= $(RUST_STATICLIB)
RUST_CLI_LINK ?= $(RUST_CLI_STATICLIB)
SET_CACHE_DIRECTORY = \
+$(MAKE) --no-print-directory $@ \
BUILD_DIR=obj/$(HASH_DIR) \
@@ -199,6 +264,8 @@ SET_CACHE_DIRECTORY = \
CFLAGS="$(CFLAGS)" \
LDFLAGS="$(LDFLAGS)" \
LDLIBS="$(LDLIBS)" \
RUST_LIBRARY_LINK="$(RUST_LIBRARY_LINK)" \
RUST_CLI_LINK="$(RUST_CLI_LINK)" \
ZSTD_ALL_SRC="$(ZSTD_ALL_SRC)"
@@ -227,7 +294,7 @@ else
# BUILD_DIR is defined
ZSTD_OBJ := $(addprefix $(BUILD_DIR)/, $(ZSTD_ALL_OBJ))
$(BUILD_DIR)/zstd : $(ZSTD_OBJ) $(RUST_STATICLIB)
$(BUILD_DIR)/zstd : $(ZSTD_OBJ) $(RUST_LIBRARY_LINK) $(RUST_CLI_LINK)
@echo "$(THREAD_MSG)"
@echo "$(ZLIB_MSG)"
@echo "$(LZMA_MSG)"
@@ -267,14 +334,14 @@ zstd32 : CPPFLAGS += -DZSTD_LEGACY_SUPPORT=$(ZSTD_LEGACY_SUPPORT)
ifneq (,$(filter Windows%,$(OS)))
zstd32 : $(RES32_FILE)
endif
zstd32 : $(ZSTDLIB_FULL_SRC) $(ZSTD_CLI_SRC) $(RUST_STATICLIB_32)
zstd32 : $(ZSTDLIB_FULL_SRC) $(ZSTD_CLI_SRC) $(RUST_STATICLIB_32) $(RUST_CLI_STATICLIB_32)
$(CC) -m32 $(FLAGS) $^ -o $@$(EXT)
## zstd-nolegacy: same scope as zstd, with removed support of legacy formats
CLEAN += zstd-nolegacy
zstd-nolegacy : LDFLAGS += $(THREAD_LD) $(ZLIBLD) $(LZMALD) $(LZ4LD) $(DEBUGFLAGS_LD)
zstd-nolegacy : CPPFLAGS += -UZSTD_LEGACY_SUPPORT -DZSTD_LEGACY_SUPPORT=0
zstd-nolegacy : $(ZSTDLIB_CORE_SRC) $(ZDICT_SRC) $(ZSTD_CLI_OBJ) $(RUST_STATICLIB)
zstd-nolegacy : $(ZSTDLIB_CORE_SRC) $(ZDICT_SRC) $(ZSTD_CLI_OBJ) $(RUST_STATICLIB) $(RUST_CLI_STATICLIB)
$(CC) $(FLAGS) $^ -o $@$(EXT) $(LDFLAGS)
.PHONY: zstd-nomt
@@ -300,6 +367,7 @@ zstd-noxz : zstd
zstd-dll : LDFLAGS+= -L$(LIB_BINDIR)
zstd-dll : LDLIBS += -lzstd
zstd-dll : ZSTDLIB_LOCAL_SRC = xxhash.c pool.c threading.c
zstd-dll : RUST_LIBRARY_LINK =
zstd-dll : zstd
@@ -331,20 +399,20 @@ CLEAN += zstd-small zstd-frugal
# requested HUF decoder mode from the C sources while the Rust archive sees it.
ZSTD_SMALL_HUF_CFLAGS := $(filter -DHUF_FORCE_DECOMPRESS_X1% -DHUF_FORCE_DECOMPRESS_X2%,$(CFLAGS))
zstd-small: CFLAGS = -Os -Wl,-s $(ZSTD_SMALL_HUF_CFLAGS)
zstd-frugal zstd-small: $(ZSTDLIB_CORE_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c $(RUST_STATICLIB)
zstd-frugal zstd-small: $(ZSTDLIB_CORE_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c $(RUST_STATICLIB) $(RUST_CLI_STATICLIB)
$(CC) $(FLAGS) -DZSTD_NOBENCH -DZSTD_NODICT -DZSTD_NOTRACE -UZSTD_LEGACY_SUPPORT -DZSTD_LEGACY_SUPPORT=0 $^ -o $@$(EXT)
CLEAN += zstd-decompress
zstd-decompress: $(ZSTDLIB_COMMON_SRC) $(ZSTDLIB_DECOMPRESS_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c $(RUST_STATICLIB)
zstd-decompress: $(ZSTDLIB_COMMON_SRC) $(ZSTDLIB_DECOMPRESS_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c $(RUST_DECOMPRESS_STATICLIB) $(RUST_DECOMPRESS_CLI_STATICLIB)
$(CC) $(FLAGS) -DZSTD_NOBENCH -DZSTD_NODICT -DZSTD_NOCOMPRESS -DZSTD_NOTRACE -UZSTD_LEGACY_SUPPORT -DZSTD_LEGACY_SUPPORT=0 $^ -o $@$(EXT)
CLEAN += zstd-compress
zstd-compress: $(ZSTDLIB_COMMON_SRC) $(ZSTDLIB_COMPRESS_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c $(RUST_STATICLIB)
zstd-compress: $(ZSTDLIB_COMMON_SRC) $(ZSTDLIB_COMPRESS_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c $(RUST_COMPRESS_STATICLIB) $(RUST_COMPRESS_CLI_STATICLIB)
$(CC) $(FLAGS) -DZSTD_NOBENCH -DZSTD_NODICT -DZSTD_NODECOMPRESS -DZSTD_NOTRACE -UZSTD_LEGACY_SUPPORT -DZSTD_LEGACY_SUPPORT=0 $^ -o $@$(EXT)
## zstd-dictBuilder: executable supporting dictionary creation and compression (only)
CLEAN += zstd-dictBuilder
zstd-dictBuilder: $(ZSTDLIB_COMMON_SRC) $(ZSTDLIB_COMPRESS_SRC) $(ZDICT_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c dibio.c $(RUST_STATICLIB)
zstd-dictBuilder: $(ZSTDLIB_COMMON_SRC) $(ZSTDLIB_COMPRESS_SRC) $(ZDICT_SRC) zstdcli.c util.c timefn.c fileio.c fileio_asyncio.c dibio.c $(RUST_COMPRESS_STATICLIB) $(RUST_COMPRESS_CLI_STATICLIB)
$(CC) $(FLAGS) -DZSTD_NOBENCH -DZSTD_NODECOMPRESS -DZSTD_NOTRACE $^ -o $@$(EXT)
RUST_DIRECT_LINK_TARGETS := zstd32 zstd-nolegacy zstd-small zstd-frugal \