zstd-native-ruby 1.0.1 → 1.0.2

data/{vendor → ext/zstd_native}/zstd/doc/zstd_manual.html

@@ -110,7 +110,6 @@ unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
            - ZSTD_CONTENTSIZE_UNKNOWN if the size cannot be determined
            - ZSTD_CONTENTSIZE_ERROR if an error occurred (e.g. invalid magic number, srcSize too small)
   note 1 : a 0 return value means the frame is valid but "empty".
-           When invoking this method on a skippable frame, it will return 0.
   note 2 : decompressed size is an optional field, it may not be present (typically in streaming mode).
            When `return==ZSTD_CONTENTSIZE_UNKNOWN`, data to decompress could be any size.
            In which case, it's necessary to use streaming mode to decompress data.
@@ -129,8 +128,9 @@ unsigned long long ZSTD_getFrameContentSize(const void *src, size_t srcSize);
 </p></pre><BR>
 
 <pre><b>ZSTD_DEPRECATED("Replaced by ZSTD_getFrameContentSize")
+ZSTDLIB_API
 unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);
-</b><p>  This function is now obsolete, in favor of ZSTD_getFrameContentSize().
+</b><p>  NOTE: This function is now obsolete, in favor of ZSTD_getFrameContentSize().
   Both functions work the same way, but ZSTD_getDecompressedSize() blends
   "empty", "unknown" and "error" results to the same return value (0),
   while ZSTD_getFrameContentSize() gives them separate return values.
@@ -142,12 +142,7 @@ unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);
  `srcSize` must be >= first frame size
  @return : the compressed size of the first frame starting at `src`,
            suitable to pass as `srcSize` to `ZSTD_decompress` or similar,
-           or an error code if input is invalid
-  Note 1: this method is called _find*() because it's not enough to read the header,
-          it may have to scan through the frame's content, to reach its end.
-  Note 2: this method also works with Skippable Frames. In which case,
-          it returns the size of the complete skippable frame,
-          which is always equal to its content size + 8 bytes for headers. 
+        or an error code if input is invalid 
 </p></pre><BR>
 
 <h3>Compression helper functions</h3><pre></pre><b><pre></pre></b><BR>
@@ -171,7 +166,8 @@ size_t ZSTD_compressBound(size_t srcSize); </b>/*!< maximum compressed size in w
  
 </p></pre><BR>
 
-<h3>Error helper functions</h3><pre></pre><b><pre></b>/* ZSTD_isError() :<b>
+<h3>Error helper functions</h3><pre></pre><b><pre>#include "zstd_errors.h" </b>/* list of errors */<b>
+</b>/* ZSTD_isError() :<b>
  * Most ZSTD_* functions returning a size_t value can be tested for error,
  * using ZSTD_isError().
  * @return 1 if error, 0 otherwise
@@ -1243,7 +1239,7 @@ size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
 </p></pre><BR>
 
 <pre><b>ZSTDLIB_STATIC_API size_t ZSTD_frameHeaderSize(const void* src, size_t srcSize);
-</b><p>  srcSize must be large enough, aka >= ZSTD_FRAMEHEADERSIZE_PREFIX.
+</b><p>  srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX.
  @return : size of the Frame Header,
            or an error code (if srcSize is too small) 
 </p></pre><BR>
@@ -1256,20 +1252,20 @@ size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
     unsigned blockSizeMax;
     ZSTD_FrameType_e frameType;          </b>/* if == ZSTD_skippableFrame, frameContentSize is the size of skippable content */<b>
     unsigned headerSize;
-    unsigned dictID;                     </b>/* for ZSTD_skippableFrame, contains the skippable magic variant [0-15] */<b>
+    unsigned dictID;
     unsigned checksumFlag;
     unsigned _reserved1;
     unsigned _reserved2;
-} ZSTD_FrameHeader;
+} ZSTD_frameHeader;
 </b></pre><BR>
-<pre><b>ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader(ZSTD_FrameHeader* zfhPtr, const void* src, size_t srcSize);
+<pre><b>ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader(ZSTD_FrameHeader* zfhPtr, const void* src, size_t srcSize);   </b>/**< doesn't consume input */<b>
 </b>/*! ZSTD_getFrameHeader_advanced() :<b>
  *  same as ZSTD_getFrameHeader(),
  *  with added capability to select a format (like ZSTD_f_zstd1_magicless) */
 ZSTDLIB_STATIC_API size_t ZSTD_getFrameHeader_advanced(ZSTD_FrameHeader* zfhPtr, const void* src, size_t srcSize, ZSTD_format_e format);
-</b><p>  decode Frame Header into `zfhPtr`, or requires larger `srcSize`.
- @return : 0 => header is complete, `zfhPtr` is correctly filled,
-          >0 => `srcSize` is too small, @return value is the wanted `srcSize` amount, `zfhPtr` is not filled,
+</b><p>  decode Frame Header, or requires larger `srcSize`.
+ @return : 0, `zfhPtr` is correctly filled,
+          >0, `srcSize` is too small, value is wanted `srcSize` amount,
            or an error code, which can be tested using ZSTD_isError() 
 </p></pre><BR>
 
@@ -1425,7 +1421,7 @@ ZSTD_compressSequences(ZSTD_CCtx* cctx,
 ZSTD_compressSequencesAndLiterals(ZSTD_CCtx* cctx,
                       void* dst, size_t dstCapacity,
                 const ZSTD_Sequence* inSeqs, size_t nbSequences,
-                const void* literals, size_t litSize, size_t litBufCapacity,
+                const void* literals, size_t litSize, size_t litCapacity,
                 size_t decompressedSize);
 </b><p> This is a variant of ZSTD_compressSequences() which,
  instead of receiving (src,srcSize) as input parameter, receives (literals,litSize),
@@ -1439,22 +1435,20 @@ ZSTD_compressSequencesAndLiterals(ZSTD_CCtx* cctx,
  - Not compatible with frame checksum, which must be disabled
  - If any block is incompressible, will fail and return an error
  - @litSize must be == sum of all @.litLength fields in @inSeqs. Any discrepancy will generate an error.
- - @litBufCapacity is the size of the underlying buffer into which literals are written, starting at address @literals.
-   @litBufCapacity must be at least 8 bytes larger than @litSize.
+ - the buffer @literals must have a size @litCapacity which is larger than @litSize by at least 8 bytes.
  - @decompressedSize must be correct, and correspond to the sum of all Sequences. Any discrepancy will generate an error.
  @return : final compressed size, or a ZSTD error code.
  
 </p></pre><BR>
 
 <pre><b>ZSTDLIB_STATIC_API size_t ZSTD_writeSkippableFrame(void* dst, size_t dstCapacity,
-                                 const void* src, size_t srcSize,
-                                       unsigned magicVariant);
+                                const void* src, size_t srcSize, unsigned magicVariant);
 </b><p> Generates a zstd skippable frame containing data given by src, and writes it to dst buffer.
 
  Skippable frames begin with a 4-byte magic number. There are 16 possible choices of magic number,
  ranging from ZSTD_MAGIC_SKIPPABLE_START to ZSTD_MAGIC_SKIPPABLE_START+15.
- As such, the parameter magicVariant controls the exact skippable frame magic number variant used,
- so the magic number used will be ZSTD_MAGIC_SKIPPABLE_START + magicVariant.
+ As such, the parameter magicVariant controls the exact skippable frame magic number variant used, so
+ the magic number used will be ZSTD_MAGIC_SKIPPABLE_START + magicVariant.
 
  Returns an error if destination buffer is not large enough, if the source size is not representable
  with a 4-byte unsigned int, or if the parameter magicVariant is greater than 15 (and therefore invalid).
@@ -1463,14 +1457,13 @@ ZSTD_compressSequencesAndLiterals(ZSTD_CCtx* cctx,
  
 </p></pre><BR>
 
-<pre><b>ZSTDLIB_STATIC_API size_t ZSTD_readSkippableFrame(void* dst, size_t dstCapacity,
-                                      unsigned* magicVariant,
-                                      const void* src, size_t srcSize);
-</b><p> Retrieves the content of a zstd skippable frame starting at @src, and writes it to @dst buffer.
+<pre><b>size_t ZSTD_readSkippableFrame(void* dst, size_t dstCapacity, unsigned* magicVariant,
+                                const void* src, size_t srcSize);
+</b><p> Retrieves a zstd skippable frame containing data given by src, and writes it to dst buffer.
 
- The parameter @magicVariant will receive the magicVariant that was supplied when the frame was written,
- i.e. magicNumber - ZSTD_MAGIC_SKIPPABLE_START.
- This can be NULL if the caller is not interested in the magicVariant.
+ The parameter magicVariant will receive the magicVariant that was supplied when the frame was written,
+ i.e. magicNumber - ZSTD_MAGIC_SKIPPABLE_START.  This can be NULL if the caller is not interested
+ in the magicVariant.
 
  Returns an error if destination buffer is not large enough, or if the frame is not skippable.
 
@@ -1478,7 +1471,7 @@ ZSTD_compressSequencesAndLiterals(ZSTD_CCtx* cctx,
  
 </p></pre><BR>
 
-<pre><b>ZSTDLIB_STATIC_API unsigned ZSTD_isSkippableFrame(const void* buffer, size_t size);
+<pre><b>unsigned ZSTD_isSkippableFrame(const void* buffer, size_t size);
 </b><p>  Tells if the content of `buffer` starts with a valid Frame Identifier for a skippable frame.
  
 </p></pre><BR>

data/{vendor → ext/zstd_native}/zstd/lib/Makefile

@@ -128,7 +128,7 @@ $(ZSTD_STATICLIB): $(ZSTD_STATICLIB_OBJ)
 	$(AR) $(ARFLAGS) $@ $^
 
 libzstd.a: $(ZSTD_STATICLIB)
-	$(CP) $< $@
+	cp -f $< $@
 
 endif
 
@@ -166,13 +166,13 @@ $(ZSTD_DYNLIB): $(ZSTD_DYNLIB_OBJ)
 	$(if $(filter -DZSTD_MULTITHREAD,$(CPPFLAGS)),\
     @echo compiling multi-threaded dynamic library $(LIBVER),\
     @echo compiling single-threaded dynamic library $(LIBVER))
-	$(CC) $(FLAGS) $^ $(SONAME_FLAGS) -o $@
+	$(CC) $(FLAGS) $^ $(LDFLAGS) $(SONAME_FLAGS) -o $@
 	@echo creating versioned links
-	$(LN) -sf $@ libzstd.$(SHARED_EXT_MAJOR)
-	$(LN) -sf $@ libzstd.$(SHARED_EXT)
+	ln -sf $@ libzstd.$(SHARED_EXT_MAJOR)
+	ln -sf $@ libzstd.$(SHARED_EXT)
 
 $(LIBZSTD): $(ZSTD_DYNLIB)
-	$(CP) $< $@
+	cp -f $< $@
 
 endif  # ifndef BUILD_DIR
 endif  # if windows
@@ -255,7 +255,7 @@ libzstd-nomt: $(ZSTD_NOMT_FILES)
         echo "Error: Found zstdmt in list."; \
         exit 1; \
     fi
-	$(CC) $(FLAGS) $^ $(SONAME_FLAGS) -o $@
+	$(CC) $(FLAGS) $^ $(LDFLAGS) $(SONAME_FLAGS) -o $@
 
 .PHONY: clean
 clean:
@@ -268,7 +268,7 @@ clean:
 #-----------------------------------------------------------------------------
 # make install is validated only for below listed environments
 #-----------------------------------------------------------------------------
-ifneq (,$(filter $(INSTALL_OS_LIST),$(UNAME)))
+ifneq (,$(filter Linux Darwin GNU/kFreeBSD GNU OpenBSD FreeBSD NetBSD DragonFly SunOS Haiku AIX MSYS_NT% CYGWIN_NT%,$(UNAME)))
 
 lib: libzstd.pc
 
@@ -363,8 +363,8 @@ install-shared:
 	[ -e $(DESTDIR)$(LIBDIR) ] || $(INSTALL) -d -m 755 $(DESTDIR)$(LIBDIR)/
 	@echo Installing shared library
 	$(INSTALL_PROGRAM) $(LIBZSTD) $(DESTDIR)$(LIBDIR)
-	$(LN) -sf $(LIBZSTD) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT_MAJOR)
-	$(LN) -sf $(LIBZSTD) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT)
+	ln -sf $(LIBZSTD) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT_MAJOR)
+	ln -sf $(LIBZSTD) $(DESTDIR)$(LIBDIR)/libzstd.$(SHARED_EXT)
 
 .PHONY: install-includes
 install-includes:

data/{vendor → ext/zstd_native}/zstd/lib/README.md

@@ -7,29 +7,15 @@ in order to make it easier to select or exclude features.
 
 #### Building
 
-A `Makefile` script is provided, supporting [Makefile conventions](https://www.gnu.org/prep/standards/html_node/Makefile-Conventions.html#Makefile-Conventions),
+`Makefile` script is provided, supporting [Makefile conventions](https://www.gnu.org/prep/standards/html_node/Makefile-Conventions.html#Makefile-Conventions),
 including commands variables, staged install, directory variables and standard targets.
 - `make` : generates both static and dynamic libraries
-- `make install` : install libraries, headers and pkg-config in local system directories
+- `make install` : install libraries and headers in target system directories
 
-`libzstd` default scope is extensive, including compression, decompression, dictionary builder,
-and support for decoding legacy formats >= v0.5.0 by default.
+`libzstd` default scope is pretty large, including compression, decompression, dictionary builder,
+and support for decoding legacy formats >= v0.5.0.
 The scope can be reduced on demand (see paragraph _modular build_).
 
-#### Multiarch Support
-
-For multiarch systems (like Debian/Ubuntu), libraries should be installed to architecture-specific directories.
-When creating packages for such systems, use the `LIBDIR` variable to specify the correct multiarch path:
-
-```bash
-# For x86_64 systems on Ubuntu/Debian:
-make install PREFIX=/usr LIBDIR=/usr/lib/x86_64-linux-gnu
-
-# For ARM64 systems on Ubuntu/Debian:
-make install PREFIX=/usr LIBDIR=/usr/lib/aarch64-linux-gnu
-```
-
-This will not only install the files in the correct directories, but also generate the correct paths for `pkg-config`.
 
 #### Multithreading support
 
@@ -207,11 +193,6 @@ The file structure is designed to make this selection manually achievable for an
   and assembly decoding loops. You may want to use this macro if these loops are
   slower on your platform.
 
-- The macro `ZDICT_QSORT` can enforce selection of a specific sorting variant,
-  which is useful when autodetection fails, for example with older versions of `musl`.
-  For this scenario, it can be set as `ZDICT_QSORT=ZDICT_QSORT_C90`.
-  Other selectable suffixes are `_GNU`, `_APPLE`, `_MSVC` and `_C11`.
-
 #### Windows : using MinGW+MSYS to create DLL
 
 DLL can be created using MinGW+MSYS with the `make libzstd` command.

data/{vendor → ext/zstd_native}/zstd/lib/common/bits.h

@@ -21,7 +21,7 @@ MEM_STATIC unsigned ZSTD_countTrailingZeros32_fallback(U32 val)
                                                 30, 22, 20, 15, 25, 17, 4, 8,
                                                 31, 27, 13, 23, 21, 19, 16, 7,
                                                 26, 12, 18, 6, 11, 5, 10, 9};
-        return DeBruijnBytePos[((U32) ((val & (0-val)) * 0x077CB531U)) >> 27];
+        return DeBruijnBytePos[((U32) ((val & -(S32) val) * 0x077CB531U)) >> 27];
     }
 }
 

data/{vendor → ext/zstd_native}/zstd/lib/common/compiler.h

@@ -218,18 +218,6 @@
 #  if defined(__ARM_NEON) || defined(_M_ARM64)
 #    define ZSTD_ARCH_ARM_NEON
 #  endif
-#  if defined(__ARM_FEATURE_SVE)
-#    define ZSTD_ARCH_ARM_SVE
-#  endif
-#  if defined(__ARM_FEATURE_SVE2)
-#    define ZSTD_ARCH_ARM_SVE2
-#  endif
-#  if defined(__riscv) && defined(__riscv_vector)
-#    if ((defined(__GNUC__) && !defined(__clang__) && __GNUC__ >= 14) || \
-        (defined(__clang__) && __clang_major__ >= 19))
-        #define ZSTD_ARCH_RISCV_RVV
-#  endif
-#endif
 #
 #  if defined(ZSTD_ARCH_X86_AVX2)
 #    include <immintrin.h>
@@ -239,12 +227,6 @@
 #  elif defined(ZSTD_ARCH_ARM_NEON)
 #    include <arm_neon.h>
 #  endif
-#  if defined(ZSTD_ARCH_ARM_SVE) || defined(ZSTD_ARCH_ARM_SVE2)
-#    include <arm_sve.h>
-#  endif
-#  if defined(ZSTD_ARCH_RISCV_RVV)
-#    include <riscv_vector.h>
-#  endif
 #endif
 
 /* C-language Attributes are added in C23. */
@@ -375,9 +357,9 @@ ptrdiff_t ZSTD_wrappedPtrDiff(unsigned char const* lhs, unsigned char const* rhs
  */
 MEM_STATIC
 ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
-const void* ZSTD_wrappedPtrAdd(const void* ptr, ptrdiff_t add)
+unsigned char const* ZSTD_wrappedPtrAdd(unsigned char const* ptr, ptrdiff_t add)
 {
-    return (const char*)ptr + add;
+    return ptr + add;
 }
 
 /**
@@ -388,9 +370,9 @@ const void* ZSTD_wrappedPtrAdd(const void* ptr, ptrdiff_t add)
  */
 MEM_STATIC
 ZSTD_ALLOW_POINTER_OVERFLOW_ATTR
-const void* ZSTD_wrappedPtrSub(const void* ptr, ptrdiff_t sub)
+unsigned char const* ZSTD_wrappedPtrSub(unsigned char const* ptr, ptrdiff_t sub)
 {
-    return (const char*)ptr - sub;
+    return ptr - sub;
 }
 
 /**
@@ -400,9 +382,9 @@ const void* ZSTD_wrappedPtrSub(const void* ptr, ptrdiff_t sub)
  * @returns `ptr + add` except it defines `NULL + 0 == NULL`.
  */
 MEM_STATIC
-void* ZSTD_maybeNullPtrAdd(void* ptr, ptrdiff_t add)
+unsigned char* ZSTD_maybeNullPtrAdd(unsigned char* ptr, ptrdiff_t add)
 {
-    return add > 0 ? (char*)ptr + add : ptr;
+    return add > 0 ? ptr + add : ptr;
 }
 
 /* Issue #3240 reports an ASAN failure on an llvm-mingw build. Out of an

data/{vendor → ext/zstd_native}/zstd/lib/common/portability_macros.h

@@ -168,23 +168,4 @@
 # define ZSTD_CET_ENDBRANCH
 #endif
 
-/**
- * ZSTD_IS_DETERMINISTIC_BUILD must be set to 0 if any compilation macro is
- * active that impacts the compressed output.
- *
- * NOTE: ZSTD_MULTITHREAD is allowed to be set or unset.
- */
-#if defined(ZSTD_CLEVEL_DEFAULT) \
-    || defined(ZSTD_EXCLUDE_DFAST_BLOCK_COMPRESSOR) \
-    || defined(ZSTD_EXCLUDE_GREEDY_BLOCK_COMPRESSOR) \
-    || defined(ZSTD_EXCLUDE_LAZY_BLOCK_COMPRESSOR) \
-    || defined(ZSTD_EXCLUDE_LAZY2_BLOCK_COMPRESSOR) \
-    || defined(ZSTD_EXCLUDE_BTLAZY2_BLOCK_COMPRESSOR) \
-    || defined(ZSTD_EXCLUDE_BTOPT_BLOCK_COMPRESSOR) \
-    || defined(ZSTD_EXCLUDE_BTULTRA_BLOCK_COMPRESSOR)
-# define ZSTD_IS_DETERMINISTIC_BUILD 0
-#else
-# define ZSTD_IS_DETERMINISTIC_BUILD 1
-#endif
-
 #endif /* ZSTD_PORTABILITY_MACROS_H */

data/{vendor → ext/zstd_native}/zstd/lib/common/threading.c

@@ -143,14 +143,7 @@ int ZSTD_pthread_mutex_init(ZSTD_pthread_mutex_t* mutex, pthread_mutexattr_t con
     *mutex = (pthread_mutex_t*)ZSTD_malloc(sizeof(pthread_mutex_t));
     if (!*mutex)
         return 1;
-    {
-        int const ret = pthread_mutex_init(*mutex, attr);
-        if (ret != 0) {
-            ZSTD_free(*mutex);
-            *mutex = NULL;
-        }
-        return ret;
-    }
+    return pthread_mutex_init(*mutex, attr);
 }
 
 int ZSTD_pthread_mutex_destroy(ZSTD_pthread_mutex_t* mutex)
@@ -171,14 +164,7 @@ int ZSTD_pthread_cond_init(ZSTD_pthread_cond_t* cond, pthread_condattr_t const*
     *cond = (pthread_cond_t*)ZSTD_malloc(sizeof(pthread_cond_t));
     if (!*cond)
         return 1;
-    {
-        int const ret = pthread_cond_init(*cond, attr);
-        if (ret != 0) {
-            ZSTD_free(*cond);
-            *cond = NULL;
-        }
-        return ret;
-    }
+    return pthread_cond_init(*cond, attr);
 }
 
 int ZSTD_pthread_cond_destroy(ZSTD_pthread_cond_t* cond)

data/{vendor → ext/zstd_native}/zstd/lib/common/zstd_common.c

@@ -46,12 +46,3 @@ ZSTD_ErrorCode ZSTD_getErrorCode(size_t code) { return ERR_getErrorCode(code); }
 /*! ZSTD_getErrorString() :
  *  provides error code string from enum */
 const char* ZSTD_getErrorString(ZSTD_ErrorCode code) { return ERR_getErrorString(code); }
-
-int ZSTD_isDeterministicBuild(void)
-{
-#if ZSTD_IS_DETERMINISTIC_BUILD
-    return 1;
-#else
-    return 0;
-#endif
-}

data/{vendor → ext/zstd_native}/zstd/lib/common/zstd_internal.h

@@ -168,7 +168,7 @@ static UNUSED_ATTR const U32 OF_defaultNormLog = OF_DEFAULTNORMLOG;
 *  Shared functions to include for inlining
 *********************************************/
 static void ZSTD_copy8(void* dst, const void* src) {
-#if defined(ZSTD_ARCH_ARM_NEON) && !defined(__aarch64__)
+#if defined(ZSTD_ARCH_ARM_NEON)
     vst1_u8((uint8_t*)dst, vld1_u8((const uint8_t*)src));
 #else
     ZSTD_memcpy(dst, src, 8);
@@ -185,8 +185,6 @@ static void ZSTD_copy16(void* dst, const void* src) {
     vst1q_u8((uint8_t*)dst, vld1q_u8((const uint8_t*)src));
 #elif defined(ZSTD_ARCH_X86_SSE2)
     _mm_storeu_si128((__m128i*)dst, _mm_loadu_si128((const __m128i*)src));
-#elif defined(ZSTD_ARCH_RISCV_RVV)
-    __riscv_vse8_v_u8m1((uint8_t*)dst, __riscv_vle8_v_u8m1((const uint8_t*)src, 16), 16);
 #elif defined(__clang__)
     ZSTD_memmove(dst, src, 16);
 #else
@@ -215,7 +213,7 @@ typedef enum {
  *           The src buffer must be before the dst buffer.
  */
 MEM_STATIC FORCE_INLINE_ATTR
-void ZSTD_wildcopy(void* dst, const void* src, size_t length, ZSTD_overlap_e const ovtype)
+void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e const ovtype)
 {
     ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src;
     const BYTE* ip = (const BYTE*)src;

data/ext/zstd_native/zstd/lib/compress/hist.c

@@ -0,0 +1,191 @@
+/* ******************************************************************
+ * hist : Histogram functions
+ * part of Finite State Entropy project
+ * Copyright (c) Meta Platforms, Inc. and affiliates.
+ *
+ *  You can contact the author at :
+ *  - FSE source repository : https://github.com/Cyan4973/FiniteStateEntropy
+ *  - Public forum : https://groups.google.com/forum/#!forum/lz4c
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+****************************************************************** */
+
+/* --- dependencies --- */
+#include "../common/mem.h"             /* U32, BYTE, etc. */
+#include "../common/debug.h"           /* assert, DEBUGLOG */
+#include "../common/error_private.h"   /* ERROR */
+#include "hist.h"
+
+
+/* --- Error management --- */
+unsigned HIST_isError(size_t code) { return ERR_isError(code); }
+
+/*-**************************************************************
+ *  Histogram functions
+ ****************************************************************/
+void HIST_add(unsigned* count, const void* src, size_t srcSize)
+{
+    const BYTE* ip = (const BYTE*)src;
+    const BYTE* const end = ip + srcSize;
+
+    while (ip<end) {
+        count[*ip++]++;
+    }
+}
+
+unsigned HIST_count_simple(unsigned* count, unsigned* maxSymbolValuePtr,
+                           const void* src, size_t srcSize)
+{
+    const BYTE* ip = (const BYTE*)src;
+    const BYTE* const end = ip + srcSize;
+    unsigned maxSymbolValue = *maxSymbolValuePtr;
+    unsigned largestCount=0;
+
+    ZSTD_memset(count, 0, (maxSymbolValue+1) * sizeof(*count));
+    if (srcSize==0) { *maxSymbolValuePtr = 0; return 0; }
+
+    while (ip<end) {
+        assert(*ip <= maxSymbolValue);
+        count[*ip++]++;
+    }
+
+    while (!count[maxSymbolValue]) maxSymbolValue--;
+    *maxSymbolValuePtr = maxSymbolValue;
+
+    {   U32 s;
+        for (s=0; s<=maxSymbolValue; s++)
+            if (count[s] > largestCount) largestCount = count[s];
+    }
+
+    return largestCount;
+}
+
+typedef enum { trustInput, checkMaxSymbolValue } HIST_checkInput_e;
+
+/* HIST_count_parallel_wksp() :
+ * store histogram into 4 intermediate tables, recombined at the end.
+ * this design makes better use of OoO cpus,
+ * and is noticeably faster when some values are heavily repeated.
+ * But it needs some additional workspace for intermediate tables.
+ * `workSpace` must be a U32 table of size >= HIST_WKSP_SIZE_U32.
+ * @return : largest histogram frequency,
+ *           or an error code (notably when histogram's alphabet is larger than *maxSymbolValuePtr) */
+static size_t HIST_count_parallel_wksp(
+                                unsigned* count, unsigned* maxSymbolValuePtr,
+                                const void* source, size_t sourceSize,
+                                HIST_checkInput_e check,
+                                U32* const workSpace)
+{
+    const BYTE* ip = (const BYTE*)source;
+    const BYTE* const iend = ip+sourceSize;
+    size_t const countSize = (*maxSymbolValuePtr + 1) * sizeof(*count);
+    unsigned max=0;
+    U32* const Counting1 = workSpace;
+    U32* const Counting2 = Counting1 + 256;
+    U32* const Counting3 = Counting2 + 256;
+    U32* const Counting4 = Counting3 + 256;
+
+    /* safety checks */
+    assert(*maxSymbolValuePtr <= 255);
+    if (!sourceSize) {
+        ZSTD_memset(count, 0, countSize);
+        *maxSymbolValuePtr = 0;
+        return 0;
+    }
+    ZSTD_memset(workSpace, 0, 4*256*sizeof(unsigned));
+
+    /* by stripes of 16 bytes */
+    {   U32 cached = MEM_read32(ip); ip += 4;
+        while (ip < iend-15) {
+            U32 c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+            c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+            c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+            c = cached; cached = MEM_read32(ip); ip += 4;
+            Counting1[(BYTE) c     ]++;
+            Counting2[(BYTE)(c>>8) ]++;
+            Counting3[(BYTE)(c>>16)]++;
+            Counting4[       c>>24 ]++;
+        }
+        ip-=4;
+    }
+
+    /* finish last symbols */
+    while (ip<iend) Counting1[*ip++]++;
+
+    {   U32 s;
+        for (s=0; s<256; s++) {
+            Counting1[s] += Counting2[s] + Counting3[s] + Counting4[s];
+            if (Counting1[s] > max) max = Counting1[s];
+    }   }
+
+    {   unsigned maxSymbolValue = 255;
+        while (!Counting1[maxSymbolValue]) maxSymbolValue--;
+        if (check && maxSymbolValue > *maxSymbolValuePtr) return ERROR(maxSymbolValue_tooSmall);
+        *maxSymbolValuePtr = maxSymbolValue;
+        ZSTD_memmove(count, Counting1, countSize);   /* in case count & Counting1 are overlapping */
+    }
+    return (size_t)max;
+}
+
+/* HIST_countFast_wksp() :
+ * Same as HIST_countFast(), but using an externally provided scratch buffer.
+ * `workSpace` is a writable buffer which must be 4-bytes aligned,
+ * `workSpaceSize` must be >= HIST_WKSP_SIZE
+ */
+size_t HIST_countFast_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
+                          const void* source, size_t sourceSize,
+                          void* workSpace, size_t workSpaceSize)
+{
+    if (sourceSize < 1500) /* heuristic threshold */
+        return HIST_count_simple(count, maxSymbolValuePtr, source, sourceSize);
+    if ((size_t)workSpace & 3) return ERROR(GENERIC);  /* must be aligned on 4-bytes boundaries */
+    if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);
+    return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, trustInput, (U32*)workSpace);
+}
+
+/* HIST_count_wksp() :
+ * Same as HIST_count(), but using an externally provided scratch buffer.
+ * `workSpace` size must be table of >= HIST_WKSP_SIZE_U32 unsigned */
+size_t HIST_count_wksp(unsigned* count, unsigned* maxSymbolValuePtr,
+                       const void* source, size_t sourceSize,
+                       void* workSpace, size_t workSpaceSize)
+{
+    if ((size_t)workSpace & 3) return ERROR(GENERIC);  /* must be aligned on 4-bytes boundaries */
+    if (workSpaceSize < HIST_WKSP_SIZE) return ERROR(workSpace_tooSmall);
+    if (*maxSymbolValuePtr < 255)
+        return HIST_count_parallel_wksp(count, maxSymbolValuePtr, source, sourceSize, checkMaxSymbolValue, (U32*)workSpace);
+    *maxSymbolValuePtr = 255;
+    return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, workSpace, workSpaceSize);
+}
+
+#ifndef ZSTD_NO_UNUSED_FUNCTIONS
+/* fast variant (unsafe : won't check if src contains values beyond count[] limit) */
+size_t HIST_countFast(unsigned* count, unsigned* maxSymbolValuePtr,
+                     const void* source, size_t sourceSize)
+{
+    unsigned tmpCounters[HIST_WKSP_SIZE_U32];
+    return HIST_countFast_wksp(count, maxSymbolValuePtr, source, sourceSize, tmpCounters, sizeof(tmpCounters));
+}
+
+size_t HIST_count(unsigned* count, unsigned* maxSymbolValuePtr,
+                 const void* src, size_t srcSize)
+{
+    unsigned tmpCounters[HIST_WKSP_SIZE_U32];
+    return HIST_count_wksp(count, maxSymbolValuePtr, src, srcSize, tmpCounters, sizeof(tmpCounters));
+}
+#endif

data/{vendor → ext/zstd_native}/zstd/lib/compress/hist.h

@@ -35,11 +35,7 @@ unsigned HIST_isError(size_t code);  /**< tells if a return value is an error co
 
 /* --- advanced histogram functions --- */
 
-#if defined(__ARM_FEATURE_SVE2)
-#define HIST_WKSP_SIZE_U32 0
-#else
 #define HIST_WKSP_SIZE_U32 1024
-#endif
 #define HIST_WKSP_SIZE    (HIST_WKSP_SIZE_U32 * sizeof(unsigned))
 /** HIST_count_wksp() :
  *  Same as HIST_count(), but using an externally provided scratch buffer.

data/{vendor → ext/zstd_native}/zstd/lib/compress/huf_compress.c

@@ -425,7 +425,6 @@ static U32 HUF_setMaxHeight(nodeElt* huffNode, U32 lastNonNull, U32 targetNbBits
                  * gain back half the rank.
                  */
                 U32 nBitsToDecrease = ZSTD_highbit32((U32)totalCost) + 1;
-                assert(nBitsToDecrease <= HUF_TABLELOG_MAX+1);
                 for ( ; nBitsToDecrease > 1; nBitsToDecrease--) {
                     U32 const highPos = rankLast[nBitsToDecrease];
                     U32 const lowPos = rankLast[nBitsToDecrease-1];