core: Add C++ classes for read-only memory-mapped files and UNIX file descriptors; Remove Zstandard's dependency on Boost.

components/core/CMakeLists.txt

@@ -322,6 +322,8 @@ set(SOURCE_FILES_unitTest
         src/clp/ffi/search/Subquery.hpp
         src/clp/ffi/search/WildcardToken.cpp
         src/clp/ffi/search/WildcardToken.hpp
+        src/clp/FileDescriptor.cpp
+        src/clp/FileDescriptor.hpp
         src/clp/FileReader.cpp
         src/clp/FileReader.hpp
         src/clp/FileWriter.cpp
@@ -359,6 +361,8 @@ set(SOURCE_FILES_unitTest
         src/clp/LogTypeDictionaryWriter.cpp
         src/clp/LogTypeDictionaryWriter.hpp
         src/clp/math_utils.hpp
+        src/clp/MemoryMappedFileView.cpp
+        src/clp/MemoryMappedFileView.hpp
         src/clp/MessageParser.cpp
         src/clp/MessageParser.hpp
         src/clp/MySQLDB.cpp
@@ -454,6 +458,7 @@ set(SOURCE_FILES_unitTest
         tests/test-kql.cpp
         tests/test-main.cpp
         tests/test-math_utils.cpp
+        tests/test-MemoryMappedFile.cpp
         tests/test-NetworkReader.cpp
         tests/test-ParserWithUserSchema.cpp
         tests/test-query_methods.cpp

components/core/src/clp/FileDescriptor.cpp

@@ -0,0 +1,55 @@
+#include "FileDescriptor.hpp"
+
+#include <fcntl.h>
+#include <sys/stat.h>
+#include <unistd.h>
+
+#include <cerrno>
+#include <cstddef>
+#include <string_view>
+
+#include "ErrorCode.hpp"
+#include "type_utils.hpp"
+
+namespace clp {
+FileDescriptor::FileDescriptor(
+        std::string_view path,
+        Mode mode,
+        CloseFailureCallback close_failure_callback
+)
+        : m_fd{open(path.data(), enum_to_underlying_type(mode))},
+          m_mode{mode},
+          m_close_failure_callback{close_failure_callback} {
+    if (-1 == m_fd) {
+        throw OperationFailed(
+                ErrorCode_errno,
+                __FILE__,
+                __LINE__,
+                "Failed to open file descriptor in path: " + std::string{path}
+        );
+    }
+}
+
+FileDescriptor::~FileDescriptor() {
+    if (-1 == m_fd) {
+        return;
+    }
+    if (0 != close(m_fd) && nullptr != m_close_failure_callback) {
+        m_close_failure_callback(errno);
+    }
+}
+
+auto FileDescriptor::get_size() const -> size_t {
+    struct stat stat_result {};
+
+    if (0 != fstat(m_fd, &stat_result)) {
+        throw OperationFailed(
+                ErrorCode_errno,
+                __FILE__,
+                __LINE__,
+                "Failed to stat file using file descriptor."
+        );
+    }
+    return static_cast<size_t>(stat_result.st_size);
+}
+}  // namespace clp

components/core/src/clp/FileDescriptor.hpp

@@ -0,0 +1,98 @@
+#ifndef CLP_FILEDESCRIPTOR_HPP
+#define CLP_FILEDESCRIPTOR_HPP
+
+#include <fcntl.h>
+
+#include <cstddef>
+#include <string>
+#include <string_view>
+#include <utility>
+
+#include "ErrorCode.hpp"
+#include "TraceableException.hpp"
+
+namespace clp {
+/**
+ * A class to wrap C style file descriptor.
+ */
+class FileDescriptor {
+public:
+    // Types
+    /**
+     * `close` is called in the destructor to close the file descriptor. However, `close` may return
+     * an error indicated by `errno`. This type alias defines a callback to handle the `close`
+     * failure in the destructor.
+     * The signature of the callback: void close_failure_callback(int errno)
+     */
+    using CloseFailureCallback = void (*)(int);
+
+    class OperationFailed : public TraceableException {
+    public:
+        OperationFailed(
+                ErrorCode error_code,
+                char const* const filename,
+                int line_number,
+                std::string msg
+        )
+                : TraceableException{error_code, filename, line_number},
+                  m_msg{std::move(msg)} {}
+
+        [[nodiscard]] auto what() const noexcept -> char const* override { return m_msg.c_str(); }
+
+    private:
+        std::string m_msg;
+    };
+
+    /**
+     * A C++ wrapper for Unix oflag that describes the operation mode.
+     */
+    // NOLINTNEXTLINE(performance-enum-size)
+    enum class Mode : int {
+        ReadOnly = O_RDONLY,
+        CreateForWrite = O_WRONLY | O_CREAT | O_TRUNC,
+    };
+
+    // Constructors
+    /**
+     * @param path
+     * @param mode
+     * @param close_failure_callback
+     */
+    FileDescriptor(
+            std::string_view path,
+            Mode mode,
+            CloseFailureCallback close_failure_callback = nullptr
+    );
+
+    // Destructor
+    ~FileDescriptor();
+
+    // Disable copy/move constructors/assignment operators
+    FileDescriptor(FileDescriptor const&) = delete;
+    FileDescriptor(FileDescriptor&&) = delete;
+    auto operator=(FileDescriptor const&) -> FileDescriptor& = delete;
+    auto operator=(FileDescriptor&&) -> FileDescriptor& = delete;
+
+    /**
+     * @return The raw fd.
+     */
+    [[nodiscard]] auto get_raw_fd() const -> int { return m_fd; }
+
+    /**
+     * @return The size of the file.
+     */
+    [[nodiscard]] auto get_size() const -> size_t;
+
+    /**
+     * @return The operation mode.
+     */
+    [[nodiscard]] auto get_mode() const -> Mode { return m_mode; }
+
+private:
+    int m_fd{-1};
+    Mode m_mode;
+    CloseFailureCallback m_close_failure_callback{nullptr};
+};
+}  // namespace clp
+
+#endif

components/core/src/clp/MemoryMappedFileView.cpp

@@ -0,0 +1,41 @@
+#include "MemoryMappedFileView.hpp"
+
+#include <sys/mman.h>
+
+#include <string>
+#include <string_view>
+
+#include "ErrorCode.hpp"
+#include "FileDescriptor.hpp"
+
+namespace clp {
+MemoryMappedFileView::MemoryMappedFileView(std::string_view path) {
+    FileDescriptor const fd{path, FileDescriptor::Mode::ReadOnly};
+    auto const file_size{fd.get_size()};
+    if (0 == file_size) {
+        // If the file is empty, we should call `mmap` since a length greater than 0 is expected.
+        // The default construct should already set buffer length to 0.
+        return;
+    }
+    auto* mmap_ptr{mmap(nullptr, file_size, PROT_READ, MAP_PRIVATE, fd.get_raw_fd(), 0)};
+    if (MAP_FAILED == mmap_ptr) {
+        throw OperationFailed(
+                ErrorCode_errno,
+                __FILE__,
+                __LINE__,
+                "`mmap` failed to map the file in path: " + std::string{path}
+        );
+    }
+    m_data = mmap_ptr;
+    m_buf_size = file_size;
+}
+
+MemoryMappedFileView::~MemoryMappedFileView() {
+    if (0 == m_buf_size) {
+        // If the mapped region has a length of 0, we should not call `munmap` since 0 is an invalid
+        // length to unmap a region.
+        return;
+    }
+    munmap(m_data, m_buf_size);
+}
+}  // namespace clp

components/core/src/clp/MemoryMappedFileView.hpp

@@ -0,0 +1,66 @@
+#ifndef CLP_MEMORYMAPPEDFILEVIEW_HPP
+#define CLP_MEMORYMAPPEDFILEVIEW_HPP
+
+#include <cstddef>
+#include <span>
+#include <string>
+#include <string_view>
+#include <utility>
+
+#include "ErrorCode.hpp"
+#include "TraceableException.hpp"
+
+namespace clp {
+/**
+ * A class for mapping a read-only file into memory. It maintains the memory buffer created by the
+ * underlying `mmap` system call and provides methods to get a view of the memory buffer.
+ */
+class MemoryMappedFileView {
+public:
+    // Types
+    class OperationFailed : public TraceableException {
+    public:
+        OperationFailed(
+                ErrorCode error_code,
+                char const* const filename,
+                int line_number,
+                std::string msg
+        )
+                : TraceableException{error_code, filename, line_number},
+                  m_msg{std::move(msg)} {}
+
+        [[nodiscard]] auto what() const noexcept -> char const* override { return m_msg.c_str(); }
+
+    private:
+        std::string m_msg;
+    };
+
+    // Constructors
+    /**
+     * @param path The path of the file to map.
+     */
+    MemoryMappedFileView(std::string_view path);
+
+    // Destructor
+    ~MemoryMappedFileView();
+
+    // Disable copy/move constructors/assignment operators
+    MemoryMappedFileView(MemoryMappedFileView const&) = delete;
+    MemoryMappedFileView(MemoryMappedFileView&&) = delete;
+    auto operator=(MemoryMappedFileView const&) -> MemoryMappedFileView& = delete;
+    auto operator=(MemoryMappedFileView&&) -> MemoryMappedFileView& = delete;
+
+    /**
+     * @return A view to the mapped file in memory.
+     */
+    [[nodiscard]] auto get_view() const -> std::span<char> {
+        return std::span<char>{static_cast<char*>(m_data), m_buf_size};
+    }
+
+private:
+    void* m_data{nullptr};
+    size_t m_buf_size{0};
+};
+}  // namespace clp
+
+#endif

components/core/src/clp/clg/CMakeLists.txt

@@ -18,6 +18,8 @@ set(
         ../ffi/ir_stream/decoding_methods.cpp
         ../ffi/ir_stream/decoding_methods.hpp
         ../ffi/ir_stream/decoding_methods.inc
+        ../FileDescriptor.cpp
+        ../FileDescriptor.hpp
         ../FileReader.cpp
         ../FileReader.hpp
         ../FileWriter.cpp
@@ -41,6 +43,8 @@ set(
         ../LogTypeDictionaryEntry.cpp
         ../LogTypeDictionaryEntry.hpp
         ../LogTypeDictionaryReader.hpp
+        ../MemoryMappedFileView.cpp
+        ../MemoryMappedFileView.hpp
         ../MySQLDB.cpp
         ../MySQLDB.hpp
         ../MySQLParamBindings.cpp

components/core/src/clp/clo/CMakeLists.txt

@@ -20,6 +20,8 @@ set(
         ../ffi/ir_stream/decoding_methods.cpp
         ../ffi/ir_stream/decoding_methods.hpp
         ../ffi/ir_stream/decoding_methods.inc
+        ../FileDescriptor.cpp
+        ../FileDescriptor.hpp
         ../FileReader.cpp
         ../FileReader.hpp
         ../FileWriter.cpp
@@ -36,6 +38,8 @@ set(
         ../LogTypeDictionaryEntry.cpp
         ../LogTypeDictionaryEntry.hpp
         ../LogTypeDictionaryReader.hpp
+        ../MemoryMappedFileView.cpp
+        ../MemoryMappedFileView.hpp
         ../networking/socket_utils.cpp
         ../networking/socket_utils.hpp
         ../networking/SocketOperationFailed.hpp

components/core/src/clp/clp/CMakeLists.txt

@@ -25,6 +25,8 @@ set(
         ../ffi/ir_stream/decoding_methods.inc
         ../ffi/ir_stream/encoding_methods.cpp
         ../ffi/ir_stream/encoding_methods.hpp
+        ../FileDescriptor.cpp
+        ../FileDescriptor.hpp
         ../FileReader.cpp
         ../FileReader.hpp
         ../FileWriter.cpp
@@ -60,6 +62,8 @@ set(
         ../LogTypeDictionaryWriter.cpp
         ../LogTypeDictionaryWriter.hpp
         ../math_utils.hpp
+        ../MemoryMappedFileView.cpp
+        ../MemoryMappedFileView.hpp
         ../MessageParser.cpp
         ../MessageParser.hpp
         ../MySQLDB.cpp

components/core/src/clp/make_dictionaries_readable/CMakeLists.txt

@@ -4,6 +4,8 @@ set(
         ../dictionary_utils.hpp
         ../DictionaryEntry.hpp
         ../DictionaryReader.hpp
+        ../FileDescriptor.cpp
+        ../FileDescriptor.hpp
         ../FileReader.cpp
         ../FileReader.hpp
         ../FileWriter.cpp
@@ -13,6 +15,8 @@ set(
         ../LogTypeDictionaryEntry.cpp
         ../LogTypeDictionaryEntry.hpp
         ../LogTypeDictionaryReader.hpp
+        ../MemoryMappedFileView.cpp
+        ../MemoryMappedFileView.hpp
         ../ParsedMessage.cpp
         ../ParsedMessage.hpp
         ../ReaderInterface.cpp