Documented varint

src/common/base58.cpp

@@ -99,7 +99,7 @@ namespace tools
         static decoded_block_sizes instance;
 
       private:
-        std::vector<int> m_data;
+        std::vector<int> m_data;p
       };
 
       decoded_block_sizes decoded_block_sizes::instance;

src/common/util.cpp

@@ -47,7 +47,7 @@ using namespace epee;
 
 namespace tools
 {
-  std::function<void(void)> signal_handler::m_handler;
+  std::function<void(void)> signal_handler::m_handler; 
 
 #ifdef WIN32
   std::string get_windows_version_display_string()

src/common/util.h

@@ -136,8 +136,8 @@ namespace tools
     /*! \breif calles m_handler */
     static void handle_signal()
     {
-      /* static std::mutex m_mutex; */
+      static std::mutex m_mutex;
-      /* std::unique_lock<std::mutex> lock(m_mutex); */
+      std::unique_lock<std::mutex> lock(m_mutex);
       m_handler();
     }
 

src/common/varint.h

@@ -35,54 +35,94 @@
 #include <utility>
 #include <sstream>
 #include <string>
+/*! \file varint.h
+ * \breif provides the implementation of varint's
+ * 
+ * The representation of varints is rather odd. The first bit of each
+ * octet is significant, it represents wheter there is another part
+ * waiting to be read. For example 0x8002 would return 0x200, even
+ * though 0x02 does not have its msb set. The actual way they are read
+ * is as follows: Strip the msb of each byte, then from left to right,
+ * read in what remains, placing it in reverse, into the buffer. Thus,
+ * the following bit stream: 0xff02 would return 0x027f. 0xff turns
+ * into 0x7f, is placed on the beggining of the buffer, then 0x02 is
+ * unchanged, since its msb is not set, and placed at the end of the
+ * buffer.
+ */
 
 namespace tools {
 
-    template<typename OutputIt, typename T>
+  /*! \breif Error codes for varint
-    typename std::enable_if<std::is_integral<T>::value && std::is_unsigned<T>::value, void>::type
+   */
-    write_varint(OutputIt &&dest, T i) {
+  enum {
-        while (i >= 0x80) {
+    /* \breif Represents the overflow error */
-            *dest++ = (static_cast<char>(i) & 0x7f) | 0x80;
+    EVARINT_OVERFLOW = -1,
-            i >>= 7;
+    /* \breif Represents a non conical represnetation */
-        }
+    EVARINT_REPRESENT = -2,
-        *dest++ = static_cast<char>(i);
+  };
+
+  /*! \breif writes a varint to a stream.
+   */
+  template<typename OutputIt, typename T>
+  /* Requires T to be both an integral type and unsigned, should be a compile error if it is not */
+  typename std::enable_if<std::is_integral<T>::value && std::is_unsigned<T>::value, void>::type 
+  write_varint(OutputIt &&dest, T i) {
+    /* Make sure that there is one after this */
+    while (i >= 0x80) {
+      *dest = (static_cast<char>(i) & 0x7f) | 0x80; 
+      ++dest;
+      i >>= 7;			/* I should be in multiples of 7, this should just get the next part */
     }
+    /* writes the last one to dest */
+    *dest = static_cast<char>(i);
+    dest++;			/* Seems kinda pointless... */
+  }
 
-    template<typename t_type>
+  /*! \breif Returns the string that represents the varint
-    std::string get_varint_data(const t_type& v)
+   */
+  template<typename T>
+    std::string get_varint_data(const T& v)
     {
       std::stringstream ss;
       write_varint(std::ostreambuf_iterator<char>(ss), v);
       return ss.str();
     }
-
+  /*! \breif reads in the varint as pointer to by InputIt into i
-    template<int bits, typename InputIt, typename T>
+   */ 
+  template<int bits, typename InputIt, typename T>
     typename std::enable_if<std::is_integral<T>::value && std::is_unsigned<T>::value && 0 <= bits && bits <= std::numeric_limits<T>::digits, int>::type
-    read_varint(InputIt &&first, InputIt &&last, T &i) {
+    read_varint(InputIt &&first, InputIt &&last, T &write) {
-        int read = 0;
+    int read = 0;
-        i = 0;
+    write = 0;
-        for (int shift = 0;; shift += 7) {
+    for (int shift = 0;; shift += 7) {
-            if (first == last) {
+      if (first == last) {
-                return read; // End of input.
+	return read; 
-            }
+      }
-            unsigned char byte = *first++;
+      unsigned char byte = *first;
-            ++read;
+      ++first;
-            if (shift + 7 >= bits && byte >= 1 << (bits - shift)) {
+      ++read;
-                return -1; // Overflow.
+      if (shift + 7 >= bits && byte >= 1 << (bits - shift)) {
-            }
+	return EVARINT_OVERFLOW;
-            if (byte == 0 && shift != 0) {
+      }
-                return -2; // Non-canonical representation.
+      if (byte == 0 && shift != 0) {
-            }
+	return EVARINT_REPRESENT;
-            i |= static_cast<T>(byte & 0x7f) << shift;
+      }
-            if ((byte & 0x80) == 0) {
+
-                break;
+      write |= static_cast<T>(byte & 0x7f) << shift; /* Does the actualy placing into write, stripping the first bit */
-            }
+
-        }
+      /* If there is no next */
-        return read;
+      if ((byte & 0x80) == 0) {
+	break;
+      }
     }
+    return read;
+  }
 
-    template<typename InputIt, typename T>
+  /*! \breif Wrapper around the other read_varint,
+   *  Sets template parameters for you.
+   */
+  template<typename InputIt, typename T>
     int read_varint(InputIt &&first, InputIt &&last, T &i) {
-        return read_varint<std::numeric_limits<T>::digits, InputIt, T>(std::move(first), std::move(last), i);
+    return read_varint<std::numeric_limits<T>::digits, InputIt, T>(std::move(first), std::move(last), i);
-    }
+  }
 }