Keep ZSTD_invalidateRepCodes responsible for extracting the private previous
block state and checking that the match window has no external dictionary.
Delegate only the fixed three-entry repcode reset through a narrow pointer ABI,
with a C assertion preserving the Rust side's ZSTD_REP_NUM assumption.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression zstd_compress::tests::invalidate_rep_codes_clears_all_entries` -- passed
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 fuzzer` and `./tests/fuzzer -s4560 -t47 -i48 -v` -- passed
- Required clippy, nightly fmt, and diff checks -- passed
ZSTD_copyCDictTableIntoCCtx only transformed tagged CDict match-table
entries or copied untagged entries; context-owned table selection remains in
C. Keep that policy and both call sites in C, forwarding the selected flag
through a narrow ABI. Rust removes the frozen eight-bit short-cache tag or
uses copy_nonoverlapping for C memcpy-equivalent copying without adding
validation.
Test Plan:
- Focused and full compression Rust tests passed (2 and 267 tests).
- `make -B -C lib -j2 lib` passed.
- `make -C tests -j2 fuzzer` passed.
- `./tests/fuzzer -s4560 -t47 -i48 -v` passed.
- Clippy normal/benches/tests before and after `cargo +nightly fmt` passed.
- `git diff --check` and `git diff --cached --check` passed.
Keep C responsible for extracting the cdict and CCtx scalar fields while
moving the attach-versus-copy mode decision into the Rust parameter module.
The Rust ABI short-circuits absent cdicts and reuses the established
attachment predicate, preserving unknown-size, force-attach, force-copy, and
force-window behavior.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features
--features compression
cparam_mode_preserves_cdict_presence_and_attachment_policy -- passed
- make -B -C lib -j2 lib -- passed
- make -C tests -j2 fuzzer -- passed
- ./tests/fuzzer -s4560 -t47 -i48 -v -- passed
- Required clippy normal/benches/tests, nightly fmt, and diff checks -- passed
The public streaming-size functions still calculated their values in the C
compressor entry point. Keep those public symbols and their size_t ABI in C,
but route the pure formulas through zstd_compress_api.rs. Rust uses usize for
the C size_t-equivalent constants and reuses its compressBound implementation;
the fixed block and header sizes make the output helper exactly 131591 bytes.
Test Plan:
- Focused Rust cstream size test -- passed (1 test)
- make -B -C lib -j2 lib -- passed
- make -C tests -j2 fuzzer -- passed
- ./tests/fuzzer -s4560 -t47 -i48 -v -- passed
- Required clippy, nightly fmt, and staged diff checks -- passed
Keep the private CDict and CCtx parameter layouts in C while passing only the
five scalar inputs needed by the dictionary-attachment heuristic. Rust now
owns the exact per-strategy cutoffs and preserves unknown-size handling,
dedicated-search short-circuiting, force-attach and force-copy precedence, and
the force-window prohibition. The existing C helper and reset/copy call paths
remain unchanged.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression dictionary_attachment` -- passed (2 tests)
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 fuzzer` -- passed
- `./tests/fuzzer -s4560 -t47 -i48 -v` -- passed
- Rust clippy for the library, benches, and tests before and after nightly formatting -- passed
- `cargo +nightly fmt --manifest-path rust/Cargo.toml` -- passed
- `git diff --cached --check` -- passed
Keep dedicated-dictionary strategy dispatch, parameter selection, and call-site
control flow in C while routing only the hashLog field transformations through
scalar Rust ABI helpers. The helpers use wrapping u32 arithmetic so overflow
and the existing subtract-then-clamp behavior remain identical to C.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression dedicated_dict_search_hash_log_adjustments_match_strategy_cases` -- passed
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 fuzzer` -- passed
- `./tests/fuzzer -s4560 -t47 -i48 -v` -- passed
- clippy normal, benches, and tests before and after formatting -- passed
- `cargo +nightly fmt --manifest-path rust/Cargo.toml` and `git diff --check` -- passed
Keep the private writeBlockHeader signature and its debug logging in C while
moving the pure block-type, size, and 24-bit little-endian encoding behind a
narrow Rust ABI. The Rust leaf preserves the RLE blockSize path for cSize == 1
and the compressed cSize path for every other value.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression block_header -- --nocapture` -- passed
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 fuzzer` -- passed
- `./tests/fuzzer -s4560 -t47 -i48 -v` -- passed
- Required clippy, nightly fmt, and diff checks -- passed
ZSTD_compressCCtx was selecting the Rust frame path for every compression
level. For source sizes that resolve to lazy or optimal strategies, that path
fell through to fast matching and could produce a different frame, including
the CCtx reuse regression caught by fuzzer test 56.
Query the source-size-dependent C parameters before preparing the context.
Keep fast and double-fast strategies on the existing Rust path, while routing
all other strategies through ZSTD_compress_usingDict with no dictionary so the
original simple API initializes requested parameters and C match state.
Test Plan:
- `make -B -C tests -j2 fuzzer` -- passed
- `./tests/fuzzer -s4560 -t56 -i56 -v` -- passed
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression` -- 253 passed
- required clippy normal, benches, and tests checks -- passed
- `cargo +nightly fmt --manifest-path rust/Cargo.toml` -- passed
- `git diff --check` and `git diff --cached --check` -- passed
ZSTD_resolveExternalSequenceValidation() is a pure int identity used while
resolving CCtx parameters. Route its existing C wrapper through a Rust
extern "C" entry point in the parameter module, preserving the int-to-int
ABI and every call-site result. Add boundary coverage for signed int modes.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression` -- passed (253 tests)
- `make -B -C lib -j2 lib` -- passed
- required clippy passes for the compression feature, benches, and tests -- passed
- `cargo +nightly fmt --manifest-path rust/Cargo.toml` and `git diff --check` -- passed
The compression-parameter override helper previously lived in C and selectively
copied each nonzero field from a caller-provided override structure. Move that
pure field-wise operation into the Rust compression-parameter module, while
keeping the existing repr(C) structure and pointer-based ABI at both C call
sites. Zero-valued overrides continue to leave the current parameter untouched.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression` -- passed (252 tests)
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 fuzzer` and `tests/fuzzer -s4560 -t47 -i48 -v` -- passed
- required clippy passes, nightly fmt, and `git diff --check` -- passed
Keep validation of the applied compression parameters in C and delegate the context-free minimum calculation through a narrow ABI leaf. Preserve the public getter and deprecated block-compression paths.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression
- cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression --all-targets
- cargo +nightly fmt --manifest-path rust/Cargo.toml
- make -B -C lib -j2 lib
- make -C tests -j2 fuzzer
- tests/fuzzer -s4560 -t47 -i48 -v
Keep C ownership of the block-state structure while delegating the confirmed-block pointer swap through a narrow two-pointer ABI leaf. Preserve every existing confirmation call site and add direct pointer-swap coverage.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression
- cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression --all-targets
- cargo +nightly fmt --manifest-path rust/Cargo.toml
- make -B -C lib -j2 lib
- tests/fuzzer -s4560 -t47 -i48 -v
- make -C tests -j2 test-zstream
Port the no-delimiter external-sequence block scanner, match splitting, repcode updates, validation, and sequence/literal storage behind a narrow C shim. Preserve trailing bytes as literals when the sequence buffer ends before the requested block, and remove the obsolete C validation wrapper.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (242 passed)
- cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression [--benches|--tests] (clean)
- cargo +nightly fmt --manifest-path rust/Cargo.toml (clean)
- make -B -C lib -j2 lib (passed)
- make -C tests -j2 test-zstream (84 deterministic; 5960 and 8307 randomized passed)
- make -C tests -j2 test-fuzzer (existing test 56 CCtx-reuse mismatch; also fails at 37eb75758)
- make -C tests -j2 test-zstd (existing Rust CLI --patch-from gap)
Port ZSTD_dictNCountRepeat and ZSTD_loadCEntropy to a dedicated Rust compression module. Preserve the existing C-facing ZSTD_loadCEntropy symbol through a narrow shim while keeping dictionary content and compression-context ownership in C.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression
- cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression --benches --tests
- make -B -C lib -j2 lib
- make -C tests -j2 test-invalidDictionaries
- make -C tests -j2 test-zstream
Move ZSTD_transferSequences_wBlockDelim's delimiter scan, literal copying, SeqDef storage, validation, and repcode bookkeeping into the Rust compression statistics module. Keep ZSTD_CCtx and block-state ownership in C through a narrow scalar/pointer shim, with ABI layout assertions for ZSTD_SequencePosition.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression
- cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression --benches --tests
- make -B -C lib -j2 lib
- make -C tests -j2 test-zstream
Block splitting still estimated literal and sequence section sizes through
three C helpers. Move the pure estimator into zstd_compress_stats.rs so its
histogram, HUF, and FSE cost calculations use the Rust implementations already
behind the compression ABI. Preserve the original literal headers, multi-stream
Huffman surcharge, sequence additional-bit costs, entropy-header accounting,
and fallback estimates; leave C responsible for split recursion and table
ownership.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression` -- 229 passed
- `cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression` plus benches/tests -- passed before and after formatting
- `cargo +nightly fmt --manifest-path rust/Cargo.toml --all` -- passed
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 test-zstream` -- passed, including 84 deterministic and 15,375 randomized cases
The explicit-delimiter sequence path still calculated each block summary in C,
with a separate AVX2 implementation and scalar fallback. Move that pure scan
to zstd_compress_stats.rs and keep a C wrapper for the existing orchestration.
The Rust leaf preserves delimiter inclusion, literal and match totals, and the
external-sequences error when no delimiter is present, while the C ABI is pinned
with a repr(C) layout check.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression` -- 227 passed
- `cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression` plus benches/tests -- passed before and after formatting
- `cargo +nightly fmt --manifest-path rust/Cargo.toml --all` -- passed
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 test-zstream` -- passed, including 84 deterministic and 15,464 randomized cases
The external-sequence block path still converted public ZSTD_Sequence values
in C, with separate scalar and AVX2 implementations. Move that pure format
conversion into zstd_compress_stats.rs so both native configurations share one
implementation. Preserve offBase encoding, u16 truncation, the rep-ignored
contract, and the long-length side-band marker; leave C responsible for block
state, repcode resolution, and long-length bookkeeping.
Test Plan:
- `cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression` -- 225 passed
- `cargo clippy --manifest-path rust/Cargo.toml --no-default-features --features compression` plus benches/tests -- passed before and after formatting
- `cargo +nightly fmt --manifest-path rust/Cargo.toml --all` -- passed
- `make -B -C lib -j2 lib` -- passed
- `make -C tests -j2 test-zstream` -- passed, including 84 deterministic and 14,768 randomized cases
Port ZSTD_seqStore_resolveOffCodes and its repcode-resolution helper to Rust. The C shim keeps Repcodes_t and SeqStore_t ownership in the compressor context while Rust preserves long-literal handling, mismatch materialization, and independent decoder/compressor history updates.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (208 tests)
- cargo clippy --manifest-path rust/Cargo.toml
- cargo clippy --manifest-path rust/Cargo.toml --benches
- cargo clippy --manifest-path rust/Cargo.toml --tests
- make -B -C lib -j2 lib
Replace the shared C raw and one-byte RLE fallback block serializers with Rust ABI leaves. Preserve the zstd block headers, payload copies, capacity errors, and all existing C compressor dispatch and context ownership; share the raw serializer with the Rust one-shot path.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (208 tests)
- cargo clippy --manifest-path rust/Cargo.toml
- cargo clippy --manifest-path rust/Cargo.toml --benches
- cargo clippy --manifest-path rust/Cargo.toml --tests
- make -B -C lib -j2 lib
Port the scalar frame-finalization policy and byte serialization into the Rust compression leaf while keeping checksum finalization, tracing, and the compression context in C. Preserve the init-to-ongoing transition on partial failure and reset the stage only after a successful epilogue.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression
- cargo clippy --manifest-path rust/Cargo.toml
- cargo clippy --manifest-path rust/Cargo.toml --benches
- cargo clippy --manifest-path rust/Cargo.toml --tests
- make -B -C lib -j2 lib
- make -B -C tests -j2 test-zstream (84 deterministic, 6737 and 9012 fuzzer cases)
Port the external-sequence window, offset, and minimum-match checks through a scalar C ABI leaf while retaining the C caller's control flow and diagnostics.
Test Plan: cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression; cargo clippy --manifest-path rust/Cargo.toml; cargo clippy --manifest-path rust/Cargo.toml --benches; cargo clippy --manifest-path rust/Cargo.toml --tests; make -B -C lib -j2 lib; make -B -C tests -j2 test-zstream (84 deterministic, 7063 and 8665 fuzzer cases).
Move ZSTD_isUpdateAuthorized behind the existing Rust parameter API while retaining the C context-stage checks, diagnostics, and mutation flow.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (187 tests)
- cargo clippy for library, benches, and tests
- make -B -C lib -j2 lib
Move hash-salt mixing and context-reset overflow predicates behind scalar Rust ABI shims. Keep the C-owned match state and window arithmetic in the C wrapper while preserving exact architecture-dependent thresholds.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (186 tests)
- cargo clippy for library, benches, tests, and CLI all-targets
- make -B -C lib -j2 lib
- make -B -C tests -j2 test-zstream (84 deterministic tests; fuzzers 6451 and 9879 cases)
Move target block-size and block-splitter policy predicates behind narrow Rust ABI shims while retaining the C debug logging, finalized-mode assertion, and surrounding control flow.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression
- cargo clippy --manifest-path rust/Cargo.toml and native library build
- make -B -C tests -j2 test-zstream (84 deterministic tests; fuzzers 6778 and 8299 cases)
Port the 16-cell match-table reduction leaf while preserving the window-index threshold, wrapping subtraction, and btlazy2 unsorted marker. Keep ZSTD_reduceIndex and its strategy-specific table selection in C behind the narrow Rust ABI.
Test Plan: cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (180 passed); root clippy, bench clippy, and test clippy; make -B -C lib -j2 lib; make -B -C tests -j2 test-zstream (84 named tests plus 6,845 and 9,628 fuzz cases passed).
Port repcode offBase finalization and explicit-delimiter block sizing through the Rust compression ABI. Preserve wrapped sequence arithmetic, external-sequence error codes, and the no-delimiter target-size policy while keeping the C control flow intact.
Test Plan: cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (176 passed); cargo clippy root, benches, and tests; make -B -C lib -j2 lib; make -B -C tests -j2 test-zstream (84 named tests plus 6,230 and 8,598 fuzz cases passed).
Move sequence-store byte accounting, chunk derivation, and repcode resolution into the Rust compression module. Move MT raw-sequence buffer conversions into Rust while preserving the existing C adapters and ABI layouts.
Test Plan:\n- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression\n- cargo clippy --manifest-path rust/Cargo.toml\n- cargo clippy --manifest-path rust/Cargo.toml --benches\n- cargo clippy --manifest-path rust/Cargo.toml --tests\n- make -B -C lib -j2 lib\n- make -B -C tests -j2 test-cli-tests
Route public CCtx parameter bounds through the Rust policy implementation while keeping build-specific MT limits supplied by the C configuration. Preserve clamping and validation behavior for migrated parameters and add focused Rust coverage.
Test Plan: cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression; cargo clippy --manifest-path rust/Cargo.toml; cargo clippy --manifest-path rust/Cargo.toml --benches; cargo clippy --manifest-path rust/Cargo.toml --tests; make -B -C lib -j2 lib; make -B -C tests -j2 test-cli-tests; make -B -C tests -j2 test-zstream
Move ZSTD_CCtxParams_init_advanced and its default-resolution policy into the Rust parameter API while retaining the C public ABI wrapper. Preserve validation, zeroing, and the existing row-match, block-splitter, LDM, and repcode defaults.
Test Plan:\n- cargo clippy --manifest-path rust/Cargo.toml\n- cargo clippy --manifest-path rust/Cargo.toml --benches\n- cargo clippy --manifest-path rust/Cargo.toml --tests\n- cargo +nightly fmt --manifest-path rust/Cargo.toml --all -- --check\n- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression zstd_compress_params_api\n- make -B -C lib -j2 lib
Move CCtx_params allocation, initialization, and release behind Rust ABI helpers while preserving the public C wrappers and custom allocator contract. Keep parameter bounds and policy in the existing C implementation.
Test Plan:\n- cargo clippy --manifest-path rust/Cargo.toml\n- cargo clippy --manifest-path rust/Cargo.toml --benches\n- cargo clippy --manifest-path rust/Cargo.toml --tests\n- cargo +nightly fmt --manifest-path rust/Cargo.toml --all\n- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression\n- make -B -C lib -j2 lib\n- make -B -C tests -j2 test-rust-lib-smoke\n- make -B -C tests -j2 test-zstream\n- make -B -C tests -j2 test-cli-tests
Match the original optimal-parser statistics for zero-possible symbols, complete the shortest-path fallback when the forward pass runs past its final position, and refresh the projected window state for the ultra2 second pass. Keep the simple Rust frame leaf restricted to the match finders it implements so lazy and optimal strategies retain their stateful lifecycle.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (135 passed)
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression zstd_opt (3 passed)
- make -B -C tests -j2 test-cli-tests (41 passed)
- make -B -C tests -j2 test-zstream (84 named tests and fuzzer suites passed)
- cargo clippy --manifest-path rust/Cargo.toml
- cargo clippy --manifest-path rust/Cargo.toml --benches
- cargo clippy --manifest-path rust/Cargo.toml --tests
Route ordinary complete-input ZSTD_compressStream2 calls through the Rust frame compressor while retaining the C state machine for partial, advanced, and dictionary-backed streams. Track explicit maxBlockSize requests so the migrated path preserves the original byte-identical default behavior, and extend the C archive smoke test across context reuse and fallback cases.
Test Plan:
- cargo clippy --manifest-path rust/Cargo.toml
- cargo clippy --manifest-path rust/Cargo.toml --benches
- cargo clippy --manifest-path rust/Cargo.toml --tests
- cargo test --manifest-path rust/Cargo.toml --no-default-features --features compression (133 passed)
- make -B -C tests -j2 test-rust-lib-smoke and ./tests/rustLibSmoke
- make -B -C tests -j2 test-zstream (passed, including both APIs and maxBlockSize case 78)
Dispatch ordinary dictionary-free ZSTD_compress2 calls through the Rust frame
compressor while retaining the original C implementation for advanced
contexts. Prepare the C workspace first so allocation downsizing and
static-context limits remain observable.
Mark successful Rust frames so a following one-shot
ZSTD_compressStream2(..., ZSTD_e_end) call uses the same output path. Other
streaming directives and advanced parameters remain on the C implementation.
Test Plan:
- RUSTC_WRAPPER= CARGO_BUILD_RUSTC_WRAPPER= cargo test --manifest-path
rust/Cargo.toml --no-default-features --features compression
zstd_compress::tests
- RUSTC_WRAPPER= CARGO_BUILD_RUSTC_WRAPPER= cargo clippy --manifest-path
rust/Cargo.toml --no-default-features --features compression -- -D warnings
- cargo +nightly fmt --manifest-path rust/Cargo.toml -- --check
- make -B -C tests fuzzer
- ./tests/fuzzer -t114 -i1000 --no-big-tests -s1
The basic fuzzer run now reaches the existing Rust one-shot compression-ratio
limit at test 113; the new stateful reuse tests pass before that point.
Move ZSTD_compressCCtx() onto the Rust frame-compression path while
retaining only the C-owned reset needed for its private context layout.
The simple API ignores advanced context parameters by contract, so the
Rust entry point can share the one-shot implementation after reset.
Preserve valid frames when the Rust superblock leaf declines a compressed
block by emitting a raw block instead. Match the fast-strategy literal
policy for negative levels so compatibility behavior remains intact.
Streaming, stateful ZSTD_compress2(), and multithreaded context
orchestration remain C-owned for a later slice. The full fuzzer currently
reaches an existing compress2() superblock-expansion failure after these
simple-API checks; that stateful path is intentionally out of scope here.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml --no-default-features
--features compression zstd_compress::tests
- cargo clippy --manifest-path rust/Cargo.toml -- -D warnings
(library, benches, and tests)
- cargo +nightly fmt --manifest-path rust/Cargo.toml -- --check
- make -B -C programs zstd V=1
- make -B -C examples multiple_simple_compression
- ./multiple_simple_compression README.md ../rust/README.md
Implement ZSTD_compress in Rust using the migrated parameter, match-finder,
sequence, frame, and superblock leaves. Keep the private context and streaming
entry points C-backed until their configuration-dependent context projection is
ported, and remove only the duplicate C one-shot definition.
Test Plan:
- cargo test --manifest-path rust/Cargo.toml
- cargo clippy --manifest-path rust/Cargo.toml --all-targets -- -D warnings
- cargo +nightly fmt --manifest-path rust/Cargo.toml --all -- --check
- make -B -C tests test-rust-lib-smoke V=1
- git diff --check
Refs: rust/src/zstd_compress.rs, lib/compress/zstd_compress.c
Continue the incremental zstd_compress.c migration with its sequence
statistics and seqStore entropy-compression layer. The following now
live in rust/src/zstd_compress_stats.rs:
- ZSTD_seqToCodes(), exported under its original name because the C
dictionary builder (zdict.c) and decodecorpus link against it,
- ZSTD_buildSequencesStatistics() and its dummy variant, whose result
struct no longer crosses the language boundary,
- ZSTD_entropyCompressSeqStore_internal(), _wExtLitBuffer(), and
ZSTD_entropyCompressSeqStore(),
- ZSTD_buildBlockEntropyStats() with its literals/sequences helpers,
- ZSTD_copyBlockSequences() and the ZSTD_updateRep() rules it shares
with the superblock writer.
Boundary: ZSTD_CCtx and ZSTD_CCtx_params stay private to C. These
paths read exactly two parameter fields, so the C shims keep the
original static/exported function names and forward the strategy and
ZSTD_literalsCompressionIsDisabled() as int scalars alongside the
seqStore and entropy-table leaves. The leaf layouts (SeqDef,
SeqStore_t, ZSTD_hufCTables_t, ZSTD_fseCTables_t, entropy metadata,
SeqCollector, ZSTD_Sequence) are pinned by compile-time asserts in
zstd_compress.c and by both-pointer-width layout tests in Rust.
ZSTD_buildSeqStore, block dispatch/splitting, and the block-size
estimation helpers remain C for a later slice.
The superblock module previously round-tripped through the C export of
ZSTD_buildBlockEntropyStats and mirrored the entropy leaf structs
privately. It now calls the crate-internal builder directly, and the
shared struct definitions moved to zstd_compress_stats; consequently
ZSTD_rust_compressSuperBlock() takes the two parameter scalars instead
of an opaque ZSTD_CCtx_params pointer, extracted by its C shim. Its
layout and repcode tests moved with the definitions.
One C helper family gets no shim: the static
ZSTD_entropyCompressSeqStore_wExtLitBuffer() had a single caller and
was folded into the Rust implementation.
Byte-identity was verified against the pre-change compressor: COPYING,
datagen -g5000000 -s7, and datagen -g300000 -s21 -P90, each at levels
1/3/9/19 and --fast=5, plus a superblock-heavy pass at level 19 with
--target-compressed-block-size=1024; all 18 frames are byte-identical,
covering the repeat-mode state machine, longOffsets, RLE/raw fallback,
and dstSize_tooSmall paths through the block splitter and superblock.
Test plan:
- cd rust && cargo fmt --check && cargo clippy --all-targets -- -D
warnings && cargo test --all-targets (131 tests: new reference
vectors for seqToCodes, RLE table headers, empty-seqStore repeat
copies, literal-stats type selection, and repcode resolution in
copyBlockSequences)
- cargo build --release --no-default-features --features compression
and --features decompression
- make -C tests fuzzer && ./tests/fuzzer -i1 --no-big-tests (covers
ZSTD_generateSequences and ZDICT training over the Rust seqToCodes)
- make -C tests test-rust-lib-smoke
- ./tests/zstreamtest -i1 and ./tests/zstreamtest --newapi -t1 -i1
- ./tests/decodecorpus -t -T1s (decodecorpus links Rust seqToCodes)
- /tmp/ref vs /tmp/got frame diff as described above: byte-identical
Move the context-free compression-parameter logic of zstd_compress.c to
rust/src/zstd_compress_params.rs: the compression-level tables (formerly
clevels.h), parameter bounds/checking/clamping, cycle log, level-table
selection, source/dictionary parameter adjustment, default frame
parameters, and the match-state/CDict size estimators.
The boundary follows the module's design notes: Rust owns only leaves
whose behavior is independent of C preprocessor configuration. C keeps
the public ZSTD_* symbols and feeds the leaves everything that is
configuration-owned as explicit scalars:
- The ZSTD_EXCLUDE_*_BLOCK_COMPRESSOR strategy cascade stays in
ZSTD_adjustCParams_internal() ahead of the Rust adjustment leaf, so
reduced builds keep their fallback policy.
- Workspace estimation receives struct sizes (ZSTD_CDict, ZSTD_match_t,
ZSTD_optimal_t), HUF workspace size, and the sanitizer redzone size,
because those depend on private layouts and ASAN configuration.
- ZSTD_cParam_getBounds() forwards only the compression level and the
seven core parameters; all other parameter bounds remain C.
- Frozen private constants the leaves hardcode (short-cache and row-hash
tag widths, MaxML/MaxLL/MaxOff/Litbits, ZSTD_OPT_SIZE, cwksp alignment)
are pinned by ZSTD_STATIC_ASSERTs at the C call sites.
Two latent 32-bit bugs in the previously unwired module were fixed
before integration: dictAndWindowLog's max window size and the
window-resize threshold were hardcoded to the 64-bit constants
(1<<31, 1<<30) instead of deriving from ZSTD_WINDOWLOG_MAX, which is
30 on 32-bit targets.
clevels.h is no longer included anywhere but stays in-tree as the
reference for mechanical comparison against the Rust table.
Test plan:
- cd rust && cargo fmt --check && cargo clippy --all-targets -- -D
warnings && cargo test --all-targets (125 tests)
- make -C tests fuzzer && ./tests/fuzzer -i1 --no-big-tests
- make -C tests test-rust-lib-smoke
- Byte-identity: zstd CLI frames for COPYING, a 250 KB C source, and a
5 MB datagen sample at levels 1/3/9/19/--fast=5 are identical between
this change and its parent commit.
ddidderr