Support reading and writing of RDA (a.k.a. RData) files.

Internally, RDA is pretty much just a wrapper around a RDS containing a
pairlist, so it's easy enough to do. Also did some refactoring to improve
sharing of code between the RDA and RDS readers/writers.

Author: LTLA

docs/Doxyfile

@@ -796,8 +796,12 @@ INPUT                  = ../include/rds2cpp/StringEncoding.hpp \
                          ../include/rds2cpp/Environment.hpp \
                          ../include/rds2cpp/Symbol.hpp \
                          ../include/rds2cpp/RdsFile.hpp \
+                         ../include/rds2cpp/Version.hpp \
+                         ../include/rds2cpp/RdaFile.hpp \
                          ../include/rds2cpp/parse_rds.hpp \
                          ../include/rds2cpp/write_rds.hpp \
+                         ../include/rds2cpp/parse_rda.hpp \
+                         ../include/rds2cpp/write_rda.hpp \
                          ../include/rds2cpp/rds2cpp.hpp \
                          ../README.md
 

include/rds2cpp/RdaFile.hpp

@@ -0,0 +1,72 @@
+#ifndef RDS2CPP_RDAFILE_HPP
+#define RDS2CPP_RDAFILE_HPP
+
+#include <array>
+#include <cstdint>
+
+#include "RObject.hpp"
+#include "Environment.hpp"
+#include "ExternalPointer.hpp"
+#include "Symbol.hpp"
+#include "Version.hpp"
+
+/**
+ * @file RdaFile.hpp
+ *
+ * @brief Information about an RDA file.
+ */
+
+namespace rds2cpp {
+
+/**
+ * @brief Contents of the parsed RDA file.
+ */
+struct RdaFile {
+    /**
+     * Version of the R serialization format.
+     */
+    std::int32_t format_version = 3;
+
+    /**
+     * R version used to write the file. 
+     */
+    Version writer_version;
+
+    /**
+     * Minimum R version required to read the file.
+     */
+    Version reader_version;
+
+    /**
+     * String encoding used to write the file, for conversion of unflagged strings.
+     */
+    StringEncoding encoding = StringEncoding::UTF8;
+
+    /**
+     * The unserialized pairlist containing all saved objects.
+     * This pairlist will be tagged with the object names. 
+     */
+    PairList contents;
+
+    /**
+     * All environments inside the file.
+     * This can be referenced by position using `EnvironmentIndex::index`.
+     */
+    std::vector<Environment> environments;
+
+    /**
+     * All symbols inside the file.
+     * This can be referenced by position using `SymbolIndex::index`.
+     */
+    std::vector<Symbol> symbols;
+
+    /**
+     * All external pointers inside the file.
+     * This can be referenced by position using `ExternalPointerIndex::index`.
+     */
+    std::vector<ExternalPointer> external_pointers;
+};
+
+}
+
+#endif

include/rds2cpp/RdsFile.hpp

@@ -8,6 +8,7 @@
 #include "Environment.hpp"
 #include "ExternalPointer.hpp"
 #include "Symbol.hpp"
+#include "Version.hpp"
 
 /**
  * @file RdsFile.hpp
@@ -17,37 +18,17 @@
 
 namespace rds2cpp {
 
-/**
- * @brief Major-minor-patch version.
- */
-struct Version {
-    /**
-     * Major version.
-     */
-    std::int16_t major = 3;
-
-    /**
-     * Minor version.
-     */
-    std::uint8_t minor = 5;
-
-    /**
-     * Patch version.
-     */
-    std::uint8_t patch = 0;
-};
-
 /**
  * @brief Contents of the parsed RDS file.
  */
 struct RdsFile {
     /**
-     * Version of the RDS format.
+     * Version of the R serialization format.
      */
     std::int32_t format_version = 3;
 
     /**
-     * R version used to write the file as major-minor-patch integers.
+     * R version used to write the file. 
      */
     Version writer_version;
 
@@ -68,19 +49,19 @@ struct RdsFile {
 
     /**
      * All environments inside the file.
-     * This can be referenced by the `index` in `EnvironmentIndex`.
+     * This can be referenced by position using `EnvironmentIndex::index`.
      */
     std::vector<Environment> environments;
 
     /**
      * All symbols inside the file.
-     * This can be referenced by the `index` in `SymbolIndex`.
+     * This can be referenced by position using `SymbolIndex::index`.
      */
     std::vector<Symbol> symbols;
 
     /**
      * All external pointers inside the file.
-     * This can be referenced by the `index` in `ExternalPointerIndex`.
+     * This can be referenced by position using `ExternalPointerIndex::index`.
      */
     std::vector<ExternalPointer> external_pointers;
 };

include/rds2cpp/Version.hpp

@@ -0,0 +1,69 @@
+#ifndef RDS2CPP_VERSION_HPP
+#define RDS2CPP_VERSION_HPP
+
+#include <cstdint>
+
+#include "sanisizer/sanisizer.hpp"
+
+#include "utils_parse.hpp"
+#include "utils_write.hpp"
+
+/**
+ * @file Version.hpp
+ *
+ * @brief R version information.
+ */
+
+namespace rds2cpp {
+
+/**
+ * @brief Major-minor-patch version number, usually of R itself.
+ */
+struct Version {
+    /**
+     * Major version.
+     */
+    std::int16_t major = 3;
+
+    /**
+     * Minor version.
+     */
+    std::uint8_t minor = 5;
+
+    /**
+     * Patch version.
+     */
+    std::uint8_t patch = 0;
+};
+
+/**
+ * @cond
+ */
+template<typename Source_>
+Version parse_version(Source_& src) {
+    // Using int32_t for the versions as the R_Version macro operates with signed integers AFIACT.
+    auto writer_version = quick_integer<std::int32_t>(src);
+    Version output;
+    output.major = (writer_version >> 16);
+    output.minor = (writer_version >> 8) & 255;
+    output.patch = writer_version & 255;
+    return output;
+}
+
+template<class BufferedWriter_>
+void write_version(const Version& version, BufferedWriter_& bufwriter) {
+    // Mimic the behavior of the R_Version macro.
+    inject_integer<std::int32_t, std::int32_t>(
+        sanisizer::product_unsafe<std::int32_t>(version.major, 65536) +
+        sanisizer::product_unsafe<std::int32_t>(version.minor, 256) +
+        static_cast<std::int32_t>(version.patch),
+        bufwriter
+    );
+}
+/**
+ * @endcond
+ */
+
+}
+
+#endif