pybcsv 1.5.2__tar.gz → 1.5.4__tar.gz

{pybcsv-1.5.2 → pybcsv-1.5.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pybcsv
-Version: 1.5.2
+Version: 1.5.4
 Summary: High-performance Python bindings for the BCSV (Binary CSV) library with pandas integration
 Keywords: csv,binary,data-science,compression,performance,pandas,numpy,big-data,data-processing
 Author-Email: Tobias Weber <weber.tobias.md@gmail.com>

{pybcsv-1.5.2 → pybcsv-1.5.4}/include/bcsv/bitset.hpp

@@ -1928,6 +1928,12 @@ constexpr Bitset<N>& Bitset<N>::assignRange(const Bitset<M>& src,
     // Fast path: destination is word-aligned
     if (dst_shift == 0) {
         const size_t base = offset / WORD_BITS;
+        // GCC false-positive: bounds are verified by the asserts above, but
+        // the compiler cannot prove it for dynamic (N=npos) specialisations.
+#if defined(__GNUC__) && !defined(__clang__)
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Warray-bounds"
+#endif
         for (size_t i = 0; i < full_words; ++i) {
             word_t s = (i < src_wc) ? src_data[i] : word_t{0};
             dst_data[base + i] = s;
@@ -1937,6 +1943,9 @@ constexpr Bitset<N>& Bitset<N>::assignRange(const Bitset<M>& src,
             word_t& d = dst_data[base + full_words];
             d = (d & ~tail_mask) | (s & tail_mask);
         }
+#if defined(__GNUC__) && !defined(__clang__)
+#pragma GCC diagnostic pop
+#endif
         clearUnusedBits();
         return *this;
     }

{pybcsv-1.5.2 → pybcsv-1.5.4}/include/bcsv/definitions.h

@@ -65,21 +65,25 @@ namespace bcsv {
     // See VERSIONING.md for compatibility rules (A/B/C).
 
     // Constants for the binary file format
+    // uint32 constants store magic as little-endian integers; char arrays hold the on-disk byte order.
+    constexpr uint32_t BCSV_MAGIC = 0x56534342;         // LE bytes: "BCSV"
+    // NOTE: PCKT_MAGIC on-disk bytes read "PKCT" (K/C transposed vs the intended "PCKT").
+    // Kept as-is to avoid invalidating all existing files; correct in the next breaking change.
+    constexpr uint32_t PCKT_MAGIC = 0x54434B50;         // LE bytes: "PKCT"
+    constexpr uint32_t FOOTER_BIDX_MAGIC = 0x58444942;  // LE bytes: "BIDX"
+    constexpr uint32_t FOOTER_EIDX_MAGIC = 0x58444945;  // LE bytes: "EIDX"
     constexpr char MAGIC_BYTES_BCSV[4] = { 'B', 'C', 'S', 'V' };
-    constexpr char MAGIC_BYTES_PCKT[4] = { 'P', 'C', 'K', 'T' };
+    constexpr char MAGIC_BYTES_PCKT[4] = { 'P', 'K', 'C', 'T' };  // matches LE of PCKT_MAGIC
     constexpr char MAGIC_BYTES_FOOTER_BIDX[4] = { 'B', 'I', 'D', 'X' };
-    constexpr char MAGIC_BYTES_FOOTER_EIDX[4] = { 'E', 'I', 'D', 'X' }; 
-    constexpr uint32_t BCSV_MAGIC = 0x56534342;   // "BCSV" in little-endian
-    constexpr uint32_t PCKT_MAGIC = 0x54434B50;   // "PCKT" in little-endian
-    constexpr uint32_t FOOTER_BIDX_MAGIC = 0x58444942;   // "BIDX" in little-endian
-    constexpr uint32_t FOOTER_EIDX_MAGIC = 0x58444945;   // "EIDX" in little-endian
+    constexpr char MAGIC_BYTES_FOOTER_EIDX[4] = { 'E', 'I', 'D', 'X' };
     constexpr uint32_t PCKT_TERMINATOR = 0x3FFFFFFF ;    // Marker value to indicate end of packet data (no more rows to come)
     constexpr size_t MAX_COLUMN_COUNT  = std::numeric_limits<uint16_t>::max();  // Maximum number of columns
     constexpr size_t MAX_COLUMN_LENGTH = std::numeric_limits<uint16_t>::max();  // Maximum width of column content
     constexpr size_t MAX_STRING_LENGTH = std::numeric_limits<uint16_t>::max(); // Maximum length of string data (wire format uses uint16_t lengths)
     constexpr size_t MAX_ROW_LENGTH    = (1ULL << 24) - 2 ;                     // ~16 MB maximum row size, 4b BLE encoding (2 bits for length), reserve 0xFFFF for terminator.
-    constexpr size_t MIN_PACKET_SIZE   = 64 * 1024;                             // 64KB minimum packet size    
-    constexpr size_t MAX_PACKET_SIZE   = 1024 * 1024 * 1024;                    // 1GB maximum packet size
+    constexpr size_t MIN_PACKET_SIZE         = 64 * 1024;                       // 64KB minimum packet size
+    constexpr size_t DEFAULT_PACKET_SIZE_KB  = 8192;                            // 8MB default packet size (in KB, for Writer::open blockSizeKB param)
+    constexpr size_t MAX_PACKET_SIZE         = 1024 * 1024 * 1024;              // 1GB maximum packet size
     /**
      * @brief Feature flag bit positions (Reserved for future optional features)
      * 

{pybcsv-1.5.2 → pybcsv-1.5.4}/include/bcsv/packet_header.h

@@ -35,7 +35,7 @@ namespace bcsv {
      * ```
      * Offset | Size | Field                  | Description
      * -------|------|------------------------|----------------------------------------
-     *   0    |  4   | magic[4]               | "PCKT" (0x50 0x43 0x4B 0x54)
+     *   0    |  4   | magic[4]               | PCKT_MAGIC (0x50 0x4B 0x43 0x54 on disk)
      *   4    |  8   | firstRowIndex          | Absolute row index (0-based, file-wide)
      *  12    |  4   | headerChecksum         | xxHash32 of bytes 0-12 (magic + firstRowIndex)
      * 
@@ -46,7 +46,7 @@ namespace bcsv {
     #pragma pack(push, 1)
     struct PacketHeader {
         
-        uint32_t magic;           ///< Magic number: "PCKT" (0x50 0x43 0x4B 0x54)
+        uint32_t magic;           ///< Magic number: PCKT_MAGIC (0x50 0x4B 0x43 0x54 on disk)
         uint64_t first_row_index; ///< Absolute row index (0-based, file-wide)
         uint32_t checksum;        ///< xxHash32 of bytes 0-19 (magic + firstRowIndex + prevPayloadChecksum)
         
@@ -73,7 +73,7 @@ namespace bcsv {
         
         /**
          * @brief Validate magic number
-         * @return true if magic number is "PCKT"
+         * @return true if magic number matches PCKT_MAGIC
          */
         bool isValidMagic() const {
             return magic == PCKT_MAGIC;
@@ -168,9 +168,11 @@ namespace bcsv {
                     break;  // Not enough data for a complete header
                 }
                 
-                // Search for "PCKT" in the buffer using string_view
+                // Search for PCKT_MAGIC bytes in the buffer
+                static constexpr auto pckt_le = std::bit_cast<std::array<char, 4>>(PCKT_MAGIC);
+                const std::string_view magic_sv(pckt_le.data(), 4);
                 std::string_view view(buffer.data(), static_cast<size_t>(validBytes));
-                size_t pos = view.find("PCKT");
+                size_t pos = view.find(magic_sv);
                 
                 while (pos != std::string_view::npos) {
                     // Check if we have enough bytes for complete header
@@ -204,7 +206,7 @@ namespace bcsv {
                     }
                     
                     // Invalid header, search for next occurrence after this position
-                    pos = view.find("PCKT", pos + 1);
+                    pos = view.find(magic_sv, pos + 1);
                 }
                 
                 // If no match found and we read a full chunk, prepare for next iteration

{pybcsv-1.5.2 → pybcsv-1.5.4}/include/bcsv/writer.h

@@ -89,7 +89,7 @@ namespace bcsv {
         const FilePath&         filePath() const                { return file_path_; }
         const LayoutType&       layout() const                  { return row_.layout(); }
         bool                    isOpen() const                  { return stream_.is_open(); }
-        bool                    open(const FilePath& filepath, bool overwrite = false, size_t compressionLevel = 1, size_t blockSizeKB = 64, FileFlags flags = FileFlags::NONE);
+        bool                    open(const FilePath& filepath, bool overwrite = false, size_t compressionLevel = 1, size_t blockSizeKB = DEFAULT_PACKET_SIZE_KB, FileFlags flags = FileFlags::NONE);
         RowType&                row()                           { return row_; }
         const RowType&          row() const                     { return row_; }
         size_t                  rowCount() const                { return row_cnt_; }

pybcsv-1.5.4/pybcsv/_version.py

@@ -0,0 +1,24 @@
+# file generated by vcs-versioning
+# don't change, don't track in version control
+from __future__ import annotations
+
+__all__ = [
+    "__version__",
+    "__version_tuple__",
+    "version",
+    "version_tuple",
+    "__commit_id__",
+    "commit_id",
+]
+
+version: str
+__version__: str
+__version_tuple__: tuple[int | str, ...]
+version_tuple: tuple[int | str, ...]
+commit_id: str | None
+__commit_id__: str | None
+
+__version__ = version = '1.5.4'
+__version_tuple__ = version_tuple = (1, 5, 4)
+
+__commit_id__ = commit_id = 'gc04e73b58'

{pybcsv-1.5.2 → pybcsv-1.5.4}/pybcsv/bindings.cpp

@@ -278,7 +278,7 @@ namespace {
         PyWriter pw(layout, row_codec);
         pw.visit([&](auto& w) {
             nb::gil_scoped_release release;
-            if (!w.open(filename, true, compression_level, 64, flags))
+            if (!w.open(filename, true, compression_level, bcsv::DEFAULT_PACKET_SIZE_KB, flags))
                 throw std::runtime_error("Failed to open file for writing: " + filename);
             for (size_t r = 0; r < num_rows; ++r) {
                 auto& row = w.row();
@@ -580,7 +580,7 @@ NB_MODULE(_bcsv, m) {
             }
             return success;
         }, nb::arg("filename"), nb::arg("overwrite") = false, nb::arg("compression_level") = 1,
-           nb::arg("block_size_kb") = 64, nb::arg("flags") = DEFAULT_FILE_FLAGS)
+           nb::arg("block_size_kb") = bcsv::DEFAULT_PACKET_SIZE_KB, nb::arg("flags") = DEFAULT_FILE_FLAGS)
         .def("write_row", [](PyWriter& pw, const nb::list& values) {
             const auto& layout = pw.visit([](auto& w) -> const bcsv::Layout& { return w.layout(); });
             if (values.size() != layout.columnCount())

pybcsv-1.5.2/pybcsv/_version.py

@@ -1,34 +0,0 @@
-# file generated by setuptools-scm
-# don't change, don't track in version control
-
-__all__ = [
-    "__version__",
-    "__version_tuple__",
-    "version",
-    "version_tuple",
-    "__commit_id__",
-    "commit_id",
-]
-
-TYPE_CHECKING = False
-if TYPE_CHECKING:
-    from typing import Tuple
-    from typing import Union
-
-    VERSION_TUPLE = Tuple[Union[int, str], ...]
-    COMMIT_ID = Union[str, None]
-else:
-    VERSION_TUPLE = object
-    COMMIT_ID = object
-
-version: str
-__version__: str
-__version_tuple__: VERSION_TUPLE
-version_tuple: VERSION_TUPLE
-commit_id: COMMIT_ID
-__commit_id__: COMMIT_ID
-
-__version__ = version = '1.5.2'
-__version_tuple__ = version_tuple = (1, 5, 2)
-
-__commit_id__ = commit_id = 'ga084b947a'