Squashed 'src/snark/' content from commit 9ada3f8

git-subtree-dir: src/snark
git-subtree-split: 9ada3f84ab484c57b2247c2f41091fd6a0916573

src/common/serialization.hpp

@@ -0,0 +1,104 @@
+/** @file
+ *****************************************************************************
+
+ Declaration of serialization routines and constants.
+
+ *****************************************************************************
+ * @author     This file is part of libsnark, developed by SCIPR Lab
+ *             and contributors (see AUTHORS).
+ * @copyright  MIT license (see LICENSE file)
+ *****************************************************************************/
+
+#ifndef SERIALIZATION_HPP_
+#define SERIALIZATION_HPP_
+
+#include <istream>
+#include <map>
+#include <ostream>
+#include <set>
+#include <vector>
+
+namespace libsnark {
+
+/*
+ * @todo
+ * The serialization is fragile. Shoud be rewritten using a standard, portable-format
+ * library like boost::serialize.
+ *
+ * However, for now the following conventions are used within the code.
+ *
+ * All algebraic objects support either binary or decimal output using
+ * the standard C++ stream operators (operator<<, operator>>).
+ *
+ * The binary mode is activated by defining a BINARY_OUTPUT
+ * preprocessor macro (e.g. g++ -DBINARY_OUTPUT ...).
+ *
+ * Binary output assumes that the stream is to be binary read at its
+ * current position so any white space should be consumed beforehand.
+ *
+ * Consecutive algebraic objects are separated by OUTPUT_NEWLINE and
+ * within themselves (e.g. X and Y coordinates for field elements) with
+ * OUTPUT_SEPARATOR (as defined below).
+ *
+ * Therefore to dump two integers, two Fp elements and another integer
+ * one would:
+ *
+ * out << 3 << "\n";
+ * out << 4 << "\n";
+ * out << FieldT(56) << OUTPUT_NEWLINE;
+ * out << FieldT(78) << OUTPUT_NEWLINE;
+ * out << 9 << "\n";
+ *
+ * Then reading back it its reader's responsibility (!) to consume "\n"
+ * after 4, but Fp::operator<< will correctly consume OUTPUT_NEWLINE.
+ *
+ * The reader should also consume "\n" after 9, so that another field
+ * element can be properly chained. This is especially important for
+ * binary output.
+ *
+ * The binary serialization of algebraic objects is currently *not*
+ * portable between machines of different word sizes.
+ */
+
+#ifdef BINARY_OUTPUT
+#define OUTPUT_NEWLINE ""
+#define OUTPUT_SEPARATOR ""
+#else
+#define OUTPUT_NEWLINE "\n"
+#define OUTPUT_SEPARATOR " "
+#endif
+
+inline void consume_newline(std::istream &in);
+inline void consume_OUTPUT_NEWLINE(std::istream &in);
+inline void consume_OUTPUT_SEPARATOR(std::istream &in);
+
+inline void output_bool(std::ostream &out, const bool b);
+
+inline void output_bool_vector(std::ostream &out, const std::vector<bool> &v);
+
+template<typename T>
+T reserialize(const T &obj);
+
+template<typename T>
+std::ostream& operator<<(std::ostream& out, const std::vector<T> &v);
+
+template<typename T>
+std::istream& operator>>(std::ostream& out, std::vector<T> &v);
+
+template<typename T1, typename T2>
+std::ostream& operator<<(std::ostream& out, const std::map<T1, T2> &m);
+
+template<typename T1, typename T2>
+std::istream& operator>>(std::istream& in, std::map<T1, T2> &m);
+
+template<typename T>
+std::ostream& operator<<(std::ostream& out, const std::set<T> &s);
+
+template<typename T>
+std::istream& operator>>(std::istream& in, std::set<T> &s);
+
+} // libsnark
+
+#include "common/serialization.tcc"
+
+#endif // SERIALIZATION_HPP_