Add cross-platform randomizer module.

The code uses two mechanisms for generating random numbers: srand()/rand(),
which is not thread-safe, and srandom()/random(), which is POSIX-specific.

Here I add a util/random.cc module that centralizes these calls, and unifies
some common usage patterns.  If the implementation is not good enough, we can
now change it in a single place.

To keep things simple, this uses the portable srand()/rand() but protects them
with a lock to avoid concurrency problems.

The hard part was to keep the regression tests passing: they rely on fixed
sequences of random numbers, so a small code change could break them very
thoroughly.  Util::rand(), for wide types like size_t, calls std::rand() not
once but twice.  This behaviour was generalized into utils::wide_rand() and
friends.

contrib/mira/Main.h

@@ -25,6 +25,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 #include "moses/Word.h"
 #include "moses/FF/FeatureFunction.h"
 #include "Decoder.h"
+#include "util/random.hh"
 
 typedef std::map<const Moses::FeatureFunction*, std::vector< float > > ProducerWeightMap;
 typedef std::pair<const Moses::FeatureFunction*, std::vector< float > > ProducerWeightPair;
@@ -37,8 +38,11 @@ template <class T> bool from_string(T& t, const std::string& s, std::ios_base& (
 
 struct RandomIndex {
   ptrdiff_t operator()(ptrdiff_t max) {
-    srand(time(0));  // Initialize random number generator with current time.
+    // TODO: Don't seed the randomizer here.  If this function gets called
-    return static_cast<ptrdiff_t> (rand() % max);
+    // multiple times in the same second, it will return the same value on
+    // each of those calls.
+    util::rand_init();
+    return util::rand_excl(max);
   }
 };
 

contrib/relent-filter/src/Main.cpp

@@ -42,6 +42,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 #include "RelativeEntropyCalc.h"
 #include "LexicalReordering.h"
 #include "LexicalReorderingState.h"
+#include "util/random.hh"
 
 #ifdef HAVE_PROTOBUF
 #include "hypergraph.pb.h"
@@ -205,7 +206,7 @@ int main(int argc, char** argv)
   
   
     //initialise random numbers
-    srand(time(NULL));
+    rand_init();
   
     // set up read/writing class
     IOWrapper* ioWrapper = GetIOWrapper(staticData);

mert/Data.cpp

@@ -287,7 +287,7 @@ void Data::createShards(size_t shard_count, float shard_size, const string& scor
     } else {
       //create shards by randomly sampling
       for (size_t i = 0; i < floor(shard_size+0.5); ++i) {
-        shard_contents.push_back(util::rand_int() % data_size);
+        shard_contents.push_back(util::rand_excl(data_size));
       }
     }
 

mert/Point.cpp

@@ -58,10 +58,8 @@ void Point::Randomize()
   UTIL_THROW_IF(m_min.size() != Point::m_dim, util::Exception, "Error");
   UTIL_THROW_IF(m_max.size() != Point::m_dim, util::Exception, "Error");
 
-  for (unsigned int i = 0; i < size(); i++) {
+  for (unsigned int i = 0; i < size(); i++)
-    const float scale = (m_max[i] - m_min[i]) / float(RAND_MAX);
+    operator[](i) = util::rand_incl(m_min[i], m_max[i]);
-    operator[](i) = m_min[i] + util::rand_int() * scale;
-  }
 }
 
 double Point::operator*(const FeatureStats& F) const

mert/TODO

@@ -5,11 +5,8 @@
 
 - check that --pairwise-ranked is compatible with all optimization metrics
 
-- Replace the standard rand() currently used in MERT and PRO with better
+- Use better random generators in util/random.cc, e.g. boost::mt19937.
-  random generators such as Boost's random generators (e.g., boost::mt19937).
+  - Support plugging of custom random generators.
-  - create a Random class to hide the details, i.e., how to generate
-    random numbers, which allows us to use custom random generators more
-    easily.
 
   Pros:
   - In MERT, you might want to use the random restarting technique to avoid

mert/evaluator.cpp

@@ -95,7 +95,7 @@ void EvaluatorUtil::evaluate(const string& candFile, int bootstrap, bool nbest_i
     for (int i = 0; i < bootstrap; ++i) {
       ScoreData scoredata(g_scorer);
       for (int j = 0; j < n; ++j) {
-        int randomIndex = util::rand_int() % n;
+        const int randomIndex = util::rand_excl(n);
         scoredata.add(entries[randomIndex], j);
       }
       g_scorer->setScoreData(&scoredata);
@@ -285,10 +285,10 @@ void InitSeed(const ProgramOption *opt)
 {
   if (opt->has_seed) {
     cerr << "Seeding random numbers with " << opt->seed << endl;
-    util::rand_int_init(opt->seed);
+    util::rand_init(opt->seed);
   } else {
     cerr << "Seeding random numbers with system clock " << endl;
-    util::rand_int_init();
+    util::rand_init();
   }
 }
 

mert/kbmira.cpp

@@ -40,6 +40,7 @@ de recherches du Canada
 #include <boost/scoped_ptr.hpp>
 
 #include "util/exception.hh"
+#include "util/random.hh"
 
 #include "BleuScorer.h"
 #include "HopeFearDecoder.h"
@@ -122,10 +123,10 @@ int main(int argc, char** argv)
 
   if (vm.count("random-seed")) {
     cerr << "Initialising random seed to " << seed << endl;
-    srand(seed);
+    util::rand_init(seed);
   } else {
     cerr << "Initialising random seed from system clock" << endl;
-    srand(time(NULL));
+    util::rand_init();
   }
 
   // Initialize weights

mert/mert.cpp

@@ -290,10 +290,10 @@ int main(int argc, char **argv)
 
   if (option.has_seed) {
     cerr << "Seeding random numbers with " << option.seed << endl;
-    util::rand_int_init(option.seed);
+    util::rand_init(option.seed);
   } else {
     cerr << "Seeding random numbers with system clock " << endl;
-    util::rand_int_init();
+    util::rand_init();
   }
 
   if (option.sparse_weights_file.size()) ++option.pdim;

mert/pro.cpp

@@ -43,6 +43,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 #include "ScoreDataIterator.h"
 #include "BleuScorer.h"
 #include "Util.h"
+#include "util/random.hh"
 
 using namespace std;
 using namespace MosesTuning;
@@ -141,10 +142,10 @@ int main(int argc, char** argv)
 
   if (vm.count("random-seed")) {
     cerr << "Initialising random seed to " << seed << endl;
-    srand(seed);
+    util::rand_init(seed);
   } else {
     cerr << "Initialising random seed from system clock" << endl;
-    srand(time(NULL));
+    util::rand_init();
   }
 
   if (scoreFiles.size() == 0 || featureFiles.size() == 0) {
@@ -211,11 +212,11 @@ int main(int argc, char** argv)
     vector<float> scores;
     size_t n_translations = hypotheses.size();
     for(size_t  i=0; i<n_candidates; i++) {
-      size_t rand1 = rand() % n_translations;
+      size_t rand1 = util::rand_excl(n_translations);
       pair<size_t,size_t> translation1 = hypotheses[rand1];
       float bleu1 = smoothedSentenceBleu(scoreDataIters[translation1.first]->operator[](translation1.second), bleuSmoothing, smoothBP);
 
-      size_t rand2 = rand() % n_translations;
+      size_t rand2 = util::rand_excl(n_translations);
       pair<size_t,size_t> translation2 = hypotheses[rand2];
       float bleu2 = smoothedSentenceBleu(scoreDataIters[translation2.first]->operator[](translation2.second), bleuSmoothing, smoothBP);
 

moses-cmd/MainVW.cpp

@@ -45,6 +45,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 #include "moses/FF/StatefulFeatureFunction.h"
 #include "moses/FF/StatelessFeatureFunction.h"
 #include "moses/TrainingTask.h"
+#include "util/random.hh"
 
 #ifdef HAVE_PROTOBUF
 #include "hypergraph.pb.h"
@@ -117,7 +118,7 @@ int main(int argc, char** argv)
 
 
     //initialise random numbers
-    srand(time(NULL));
+    util::rand_init();
 
     // set up read/writing class
     IFVERBOSE(1) {

moses/ExportInterface.cpp

@@ -27,6 +27,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 #include <sstream>
 #include <vector>
 
+#include "util/random.hh"
 #include "util/usage.hh"
 
 #ifdef WIN32
@@ -91,7 +92,7 @@ SimpleTranslationInterface::SimpleTranslationInterface(const string &mosesIni):
       exit(1);
     }
 
-    srand(time(NULL));
+    util::rand_init();
 
 }
 
@@ -185,7 +186,7 @@ batch_run()
   const StaticData& staticData = StaticData::Instance();
 
   //initialise random numbers
-  srand(time(NULL));
+  util::rand_init();
 
   IFVERBOSE(1) PrintUserTime("Created input-output object");
     

moses/Manager.cpp

@@ -54,6 +54,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 #endif
 
 #include "util/exception.hh"
+#include "util/random.hh"
 
 using namespace std;
 
@@ -418,7 +419,7 @@ void Manager::CalcLatticeSamples(size_t count, TrellisPathList &ret) const
       //cerr << endl;
 
       //draw the sample
-      float frandom = log((float)rand()/RAND_MAX);
+      const float frandom = log(util::rand_incl(0.0f, 1.0f));
       size_t position = 1;
       float sum = candidateScores[0];
       for (; position < candidateScores.size() && sum < frandom; ++position) {

moses/Parameter.cpp

@@ -31,6 +31,7 @@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 #include "InputFileStream.h"
 #include "StaticData.h"
 #include "util/exception.hh"
+#include "util/random.hh"
 #include <boost/program_options.hpp>
 
 
@@ -1392,7 +1393,7 @@ struct Credit {
     this->contact							= contact						;
     this->currentPursuits			= currentPursuits		;
     this->areaResponsibility	= areaResponsibility;
-    this->sortId							= rand() % 1000;
+    this->sortId							= util::rand_excl(1000);
   }
 
   bool operator<(const Credit &other) const {

moses/TranslationModel/DynSAInclude/hash.h

@@ -6,6 +6,7 @@
 #include "utils.h"
 #include "FileHandler.h"
 #include "util/exception.hh"
+#include "util/random.hh"
 
 using namespace Moses;
 typedef uint64_t P;   // largest input range is 2^64
@@ -162,7 +163,7 @@ void Hash_shiftAddXOR<T>::initSeeds()
 {
   v_ = new T[this->H_];
   for(count_t i=0; i < this->H_; i++)
-    v_[i] = Utils::rand<T>() + 1;
+    v_[i] = util::wide_rand<T>() + 1;
 }
 template <typename T>
 T Hash_shiftAddXOR<T>::hash(const char* s, count_t h)
@@ -187,9 +188,8 @@ void UnivHash_tableXOR<T>::initSeeds()
   // fill with random values
   for(count_t j=0; j < this->H_; j++) {
     table_[j] = new T[tblLen_];
-    for(count_t i=0; i < tblLen_; i++) {
+    for(count_t i=0; i < tblLen_; i++)
-      table_[j][i] = Utils::rand<T>(this->m_-1);
+      table_[j][i] = util::wide_rand_excl(this->m_-1);
-    }
   }
 }
 template <typename T>
@@ -218,7 +218,7 @@ void UnivHash_noPrimes<T>::initSeeds()
 {
   a_ = new P[this->H_];
   for(T i=0; i < this->H_; i++) {
-    a_[i] = Utils::rand<P>();
+    a_[i] = util::wide_rand<P>();
     if(a_[i] % 2 == 0) a_[i]++;  // a must be odd
   }
 }
@@ -284,8 +284,8 @@ void UnivHash_linear<T>::initSeeds()
     a_[i] = new T[MAX_NGRAM_ORDER];
     b_[i] = new T[MAX_NGRAM_ORDER];
     for(count_t j=0; j < MAX_NGRAM_ORDER; j++) {
-      a_[i][j] = 1 + Utils::rand<T>();
+      a_[i][j] = 1 + util::wide_rand<T>();
-      b_[i][j] = Utils::rand<T>();
+      b_[i][j] = util::wide_rand<T>();
     }
   }
 }

moses/TranslationModel/DynSAInclude/utils.h

@@ -62,22 +62,6 @@ public:
       str[i] = tolower(str[i]);
     }
   }
-  // TODO: interface with decent PRG
-  template<typename T>
-  static T rand(T mod_bnd = 0) {
-    T random = 0;
-    if(sizeof(T) <= 4) {
-      random = static_cast<T>(std::rand());
-    } else if(sizeof(T) == 8) {
-      random = static_cast<T>(std::rand());
-      random <<= 31;
-      random <<= 1;
-      random |= static_cast<T>(std::rand());
-    }
-    if(mod_bnd != 0)
-      return random % mod_bnd;
-    else return random;
-  }
 };
 
 #endif

moses/TranslationModel/DynSuffixArray.cpp

@@ -1,4 +1,6 @@
 #include "DynSuffixArray.h"
+#include "util/random.hh"
+
 #include <iostream>
 #include <boost/foreach.hpp>
 
@@ -315,33 +317,31 @@ int DynSuffixArray::Compare(int pos1, int pos2, int max)
   return 0;
 }
 
+namespace
+{
+/// Helper: swap two entries in an int array.
+inline void swap_ints(int array[], int one, int other)
+{
+  const int tmp = array[one];
+  array[one] = array[other];
+  array[other] = tmp;
+}
+}
+
 void DynSuffixArray::Qsort(int* array, int begin, int end)
 {
   if(end > begin) {
-    int index;
+    int index = util::rand_incl(begin, end);
     {
-      index = begin + (rand() % (end - begin + 1));
+      const int pivot = array[index];
-      int pivot = array[index];
+      swap_ints(array, index, end);
-      {
-        int tmp = array[index];
-        array[index] = array[end];
-        array[end] = tmp;
-      }
       for(int i=index=begin; i < end; ++i) {
         if (Compare(array[i], pivot, 20) <= 0) {
-          {
+          swap_ints(array, index, i);
-            int tmp = array[index];
-            array[index] = array[i];
-            array[i] = tmp;
           index++;
         }
       }
-      }
+      swap_ints(array, index, end);
-      {
-        int tmp = array[index];
-        array[index] = array[end];
-        array[end] = tmp;
-      }
     }
     Qsort(array, begin, index - 1);
     Qsort(array, index + 1,  end);

moses/TranslationModel/RuleTable/PhraseDictionaryFuzzyMatch.cpp

@@ -45,6 +45,7 @@
 #include "moses/TranslationModel/fuzzy-match/SentenceAlignment.h"
 #include "util/file.hh"
 #include "util/exception.hh"
+#include "util/random.hh"
 
 using namespace std;
 
@@ -62,8 +63,8 @@ char *mkdtemp(char *tempbuf)
     return NULL;
   }
 
-  srand((unsigned)time(0));
+  util::rand_init();
-  rand_value = (int)((rand() / ((double)RAND_MAX+1.0)) * 1e6);
+  rand_value = rand_excl(1e6);
   tempbase = strrchr(tempbuf, '/');
   tempbase = tempbase ? tempbase+1 : tempbuf;
   strcpy(tempbasebuf, tempbase);

moses/TranslationModel/UG/generic/sampling/Sampling.h

@@ -2,6 +2,9 @@
 #define __sampling_h
 #include <boost/dynamic_bitset.hpp>
 #include <vector>
+
+#include "util/random.hh"
+
 // Utility functions for proper sub-sampling.
 // (c) 2007-2012 Ulrich Germann
 
@@ -9,12 +12,6 @@
 namespace Moses
 {
 using namespace std;
-inline
-size_t
-randInt(size_t N)
-{
-  return N*(rand()/(RAND_MAX+1.));
-}
 
 // select a random sample of size /s/ without restitution from the range of
 // integers [0,N);
@@ -35,15 +32,15 @@ randomSample(vector<idx_t>& v, size_t s, size_t N)
   if (s*10<N) {
     boost::dynamic_bitset<uint64_t> check(N,0);
     for (size_t i = 0; i < v.size(); i++) {
-      size_t x = randInt(N);
+      size_t x = util::rand_excl(N);
-      while (check[x]) x = randInt(N);
+      while (check[x]) x = util::rand_excl(N);
       check[x]=true;
       v[i] = x;
     }
   } else {
     size_t m=0;
     for (size_t t = 0; m <= s && t < N; t++)
-      if (s==N || randInt(N-t) < s-m) v[m++] = t;
+      if (s==N || util::rand_excl(N-t) < s-m) v[m++] = t;
   }
 }
 

moses/TranslationModel/UG/mm/ug_mmbitext.cc

@@ -345,7 +345,7 @@
 // 	{
 // 	  boost::lock_guard<boost::mutex> lock(stats->lock);
 // 	  if (stats->raw_cnt == ctr) ++stats->raw_cnt;
-// 	  size_t rnum = randInt(stats->raw_cnt - ctr++);
+// 	  size_t rnum = util::rand_excl(stats->raw_cnt - ctr++);
 // 	  // cout << stats->raw_cnt << " " << ctr-1 << " " 
 // 	  // << rnum << " " << max_samples - stats->good << endl;
 // 	  if (rnum < max_samples - stats->good)

moses/TranslationModel/UG/mm/ug_tsa_array_entry.h

@@ -69,7 +69,7 @@ namespace ugdiss
     //   while (chosen < samplesize && next < stop)
     // 	{
     // 	  root->readEntry(next,*this);
-    // 	  if (randInt(N - sampled++) < samplesize - chosen)
+    // 	  if (util::rand_excl(N - sampled++) < samplesize - chosen)
     // 	    {
     // 	      ++chosen;
     // 	      return true;

moses/TranslationModel/UG/mm/ug_tsa_tree_iterator.h

@@ -9,6 +9,7 @@
 #include <iostream>
 #include "util/exception.hh"
 #include "moses/Util.h"
+#include "util/random.hh"
 //#include <cassert>
 
 // #include "ug_bv_iter.h"
@@ -894,13 +895,6 @@ namespace ugdiss
     return bv;
   }
 
-  inline
-  size_t 
-  randInt(size_t N)
-  {
-    return size_t(N*(rand()/(RAND_MAX+1.)));
-  }
-
   /// randomly select up to N occurrences of the sequence
   template<typename Token>
   sptr<vector<typename ttrack::Position> >
@@ -922,8 +916,8 @@ namespace ugdiss
         root->readEntry(I.next,I);
         
         // t: expected number of remaining samples
-        double t = (stop - I.pos)/root->aveIndexEntrySize();
+        const double t = (stop - I.pos)/root->aveIndexEntrySize();
-        double r = t*rand()/(RAND_MAX+1.);
+        const double r = util::rand_excl(t);
         if (r < N-m)
           {
             ret->at(m).offset = I.offset;

util/random.cc

@@ -19,21 +19,25 @@ namespace
 boost::mutex rand_lock;
 } // namespace
 
-void rand_int_init(unsigned int seed)
+void rand_init(unsigned int seed)
 {
   boost::lock_guard<boost::mutex> lock(rand_lock);
   srand(seed);
 }
 
 
-void rand_int_init()
+void rand_init()
 {
-  rand_int_init(time(NULL));
+  rand_init(time(NULL));
 }
 
+namespace internal
+{
+// This is the one call to the actual randomizer.  All else is built on this.
 int rand_int()
 {
   boost::lock_guard<boost::mutex> lock(rand_lock);
-  return rand();
+  return std::rand();
 }
+} // namespace internal
 } // namespace util

util/random.hh

@@ -1,32 +1,229 @@
 #ifndef UTIL_RANDOM_H
 #define UTIL_RANDOM_H
 
+#include <cstdlib>
+#include <limits>
+
 namespace util
 {
+/** Thread-safe, cross-platform random number generator.
+ *
+ * This is not for proper security-grade randomness, but should be "good
+ * enough" for producing arbitrary values of various numeric types.
+ *
+ * Before starting, call rand_init() to seed the randomizer.  There is no need
+ * to do this more than once; in fact doing it more often is likely to make the
+ * randomizer less effective.  Once that is done, call the rand(), rand_excl(),
+ * and rand_incl() functions as needed to generate pseudo-random numbers.
+ *
+ * Probability distribution is roughly uniform, but for integral types is
+ * skewed slightly towards lower numbers depending on how close "top" comes to
+ * RAND_MAX.
+ *
+ * For floating-point types, resolution is limited; there will actually be
+ * only RAND_MAX different possible values.
+ */
 
 /** Initialize randomizer with a fixed seed.
  *
  * After this, unless the randomizer gets seeded again, consecutive calls to
- * rand_int() will return a sequence of pseudo-random numbers determined by
+ * the random functions will return a sequence of pseudo-random numbers
- * the seed.  Every time the randomizer is seeded with this same seed, it will
+ * determined by the seed.  Every time the randomizer is seeded with this same
- * again start returning the same sequence of numbers.
+ * seed, it will again start returning the same sequence of numbers.
  */
-void rand_int_init(unsigned int);
+void rand_init(unsigned int);
 
 /** Initialize randomizer based on current time.
  *
  * Call this to make the randomizer return hard-to-predict numbers.  It won't
  * produce high-grade randomness, but enough to make the program act
  * differently on different runs.
+ *
+ * The seed will be based on the current time in seconds.  So calling it twice
+ * within the same second will just reset the randomizer to where it was before.
+ * Don't do that.
  */
-void rand_int_init();
+void rand_init();
+
 
 /** Return a pseudorandom number between 0 and RAND_MAX inclusive.
  *
  * Initialize (seed) the randomizer before starting to call this.
  */
+template<typename T> inline T rand();
+
+
+/** Return a pseudorandom number in the half-open interval [bottom, top).
+ *
+ * Generates a value between "bottom" (inclusive) and "top" (exclusive),
+ * assuming that (top - bottom) <= RAND_MAX.
+ */
+template<typename T> inline T rand_excl(T bottom, T top);
+
+
+/** Return a pseudorandom number in the half-open interval [0, top).
+ *
+ * Generates a value between 0 (inclusive) and "top" (exclusive), assuming that
+ * bottom <= RAND_MAX.
+ */
+template<typename T> inline T rand_excl(T top);
+
+
+/** Return a pseudorandom number in the open interval [bottom, top].
+ *
+ * Generates a value between "bottom" and "top" inclusive, assuming that
+ * (top - bottom) < RAND_MAX.
+ */
+template<typename T> inline T rand_incl(T bottom, T top);
+
+
+/** Return a pseudorandom number in the open interval [0, top].
+ *
+ * Generates a value between 0 and "top" inclusive, assuming that
+ * bottom < RAND_MAX.
+ */
+template<typename T> inline T rand_incl(T top);
+
+
+/** Return a pseudorandom number which may be larger than RAND_MAX.
+ *
+ * The requested type must be integral, and its size must be an even multiple
+ * of the size of an int.  The return value will combine one or more random
+ * ints into a single value, which could get quite large.
+ *
+ * The result is nonnegative.  Because the constituent ints are also
+ * nonnegative, the most significant bit in each of the ints will be zero,
+ * so for a wider type, there will be "gaps" in the range of possible outputs.
+ */
+template<typename T> inline T wide_rand();
+
+/** Return a pseudorandom number in [0, top), not limited to RAND_MAX.
+ *
+ * Works like wide_rand(), but if the requested type is wider than an int, it
+ * accommodates larger top values than an int can represent.
+ */
+template<typename T> inline T wide_rand_excl(T top);
+
+/** Return a pseudorandom number in [bottom, top), not limited to RAND_MAX.
+ *
+ * Works like wide_rand(), but if the requested type is wider than an int, it
+ * accommodates larger value ranges than an int can represent.
+ */
+template<typename T> inline T wide_rand_excl(T bottom, T top);
+
+/** Return a pseudorandom number in [0, top], not limited to RAND_MAX.
+ *
+ * Works like wide_rand(), but if the requested type is wider than an int, it
+ * accommodates larger top values than an int can represent.
+ */
+template<typename T> inline T wide_rand_incl(T top);
+
+/** Return a pseudorandom number in [bottom, top], not limited to RAND_MAX.
+ *
+ * Works like wide_rand(), but if the requested type is wider than an int, it
+ * accommodates larger top values than an int can represent.
+ */
+template<typename T> inline T wide_rand_incl(T bottom, T top);
+
+
+/// Implementation detail.  For the random module's internal use only.
+namespace internal
+{
+/// The central call to the randomizer upon which this whole module is built.
 int rand_int();
 
+/// Helper template: customize random values to required ranges.
+template<typename T, bool is_integer_type> struct random_scaler;
+
+/// Specialized random_scaler for integral types.
+template<typename T> struct random_scaler<T, true>
+{
+  static T rnd_excl(T value, T range) { return value % range; }
+  static T rnd_incl(T value, T range) { return value % (range + 1); }
+};
+
+/// Specialized random_scaler for non-integral types.
+template<typename T> struct random_scaler<T, false>
+{
+  static T rnd_excl(T value, T range)
+  {
+    // Promote RAND_MAX to T before adding one to avoid overflow.
+    return range * value / (T(RAND_MAX) + 1);
+  }
+  static T rnd_incl(T value, T range) { return range * value / RAND_MAX; }
+};
+
+/// Helper for filling a wider variable with random ints.
+template<typename T, size_t remaining_ints> struct wide_random_collector
+{
+  static T generate()
+  {
+    T one_int = util::rand<T>() << (8 * sizeof(int));
+    return one_int | wide_random_collector<T, remaining_ints-1>::generate();
+  }
+};
+/// Specialized wide_random_collector for generating just a single int.
+template<typename T> struct wide_random_collector<T, 1>
+{
+  static T generate() { return util::rand<T>(); }
+};
+
+} // namespace internal
+
+
+template<typename T> inline T rand()
+{
+  return T(util::internal::rand_int());
+}
+
+template<typename T> inline T rand_excl(T top)
+{
+  typedef internal::random_scaler<T, std::numeric_limits<T>::is_integer> scaler;
+  return scaler::rnd_excl(util::rand<T>(), top);
+}
+
+template<typename T> inline T rand_excl(T bottom, T top)
+{
+  return bottom + rand_excl(top - bottom);
+}
+
+template<typename T> inline T rand_incl(T top)
+{
+  typedef internal::random_scaler<T, std::numeric_limits<T>::is_integer> scaler;
+  return scaler::rnd_incl(util::rand<T>(), top);
+}
+
+template<typename T> inline T rand_incl(T bottom, T top)
+{
+  return bottom + rand_incl(top - bottom);
+}
+
+template<typename T> inline T wide_rand()
+{
+  return internal::wide_random_collector<T, sizeof(T)/sizeof(int)>::generate();
+}
+
+template<typename T> inline T wide_rand_excl(T top)
+{
+  typedef internal::random_scaler<T, std::numeric_limits<T>::is_integer> scaler;
+  return scaler::rnd_excl(util::wide_rand<T>(), top);
+}
+
+template<typename T> inline T wide_rand_excl(T bottom, T top)
+{
+  return bottom + wide_rand_excl(top - bottom);
+}
+
+template<typename T> inline T wide_rand_incl(T top)
+{
+  typedef internal::random_scaler<T, std::numeric_limits<T>::is_integer> scaler;
+  return scaler::rnd_incl(util::wide_rand<T>(), top);
+}
+
+template<typename T> inline T wide_rand_incl(T bottom, T top)
+{
+  return bottom + wide_rand_incl(top - bottom);
+}
 } // namespace util
 
 #endif

util/random_test.cc

@@ -1,3 +1,5 @@
+#include <cstdlib>
+
 #include "util/random.hh"
 
 #define BOOST_TEST_MODULE RandomTest
@@ -8,32 +10,182 @@ namespace util
 namespace
 {
 
-BOOST_AUTO_TEST_CASE(returns_different_consecutive_numbers)
+BOOST_AUTO_TEST_CASE(rand_int_returns_positive_no_greater_than_RAND_MAX)
 {
-  rand_int_init(99);
+  rand_init();
-  const int first = rand_int(), second = rand_int(), third = rand_int();
+  for (int i=0; i<100; i++)
+  {
+    const int random_number = rand<int>();
+    BOOST_CHECK(random_number >= 0);
+    BOOST_CHECK(random_number <= RAND_MAX);
+  }
+}
+
+BOOST_AUTO_TEST_CASE(rand_int_returns_different_consecutive_numbers)
+{
+  rand_init(99);
+  const int first = rand<int>(), second = rand<int>(), third = rand<int>();
   // Sometimes you'll get the same number twice in a row, but generally the
   // randomizer returns different numbers.
   BOOST_CHECK(second != first || third != first);
 }
 
-BOOST_AUTO_TEST_CASE(returns_different_numbers_for_different_seeds)
+BOOST_AUTO_TEST_CASE(rand_int_returns_different_numbers_for_different_seeds)
 {
-  rand_int_init(1);
+  rand_init(1);
-  const int one1 = rand_int(), one2 = rand_int();
+  const int one1 = rand<int>(), one2 = rand<int>();
-  rand_int_init(2);
+  rand_init(2);
-  const int two1 = rand_int(), two2 = rand_int();
+  const int two1 = rand<int>(), two2 = rand<int>();
-  BOOST_CHECK(two1 != one1 || two2 != two1);
+  BOOST_CHECK(two1 != one1 || two2 != one2);
 }
 
-BOOST_AUTO_TEST_CASE(returns_same_sequence_for_same_seed)
+BOOST_AUTO_TEST_CASE(rand_int_returns_same_sequence_for_same_seed)
 {
-  rand_int_init(1);
+  rand_init(1);
-  const int first = rand_int();
+  const int first = rand<int>();
-  rand_int_init(1);
+  rand_init(1);
-  const int second = rand_int();
+  const int second = rand<int>();
   BOOST_CHECK_EQUAL(first, second);
 }
 
+BOOST_AUTO_TEST_CASE(rand_excl_int_returns_number_in_range)
+{
+  const int bottom = 10, top = 50;
+  for (int i=0; i<100; i++)
+  {
+    const int random_number = rand_excl(bottom, top);
+    BOOST_CHECK(random_number >= bottom);
+    BOOST_CHECK(random_number < top);
+  }
+}
+
+BOOST_AUTO_TEST_CASE(rand_excl_int_covers_full_range)
+{
+  // The spread of random numbers really goes all the way from 0 (inclusive)
+  // to "top" (exclusive).  It's not some smaller subset.
+  // This test will randomly fail sometimes, though very very rarely, when the
+  // random numbers don't actually have enough different values.
+  const int bottom = 1, top = 4;
+  int lowest = 99, highest = -1;
+  for (int i=0; i<100; i++)
+  {
+    const int random_number = rand_excl(bottom, top);
+    lowest = std::min(lowest, random_number);
+    highest = std::max(highest, random_number);
+  }
+
+  BOOST_CHECK_EQUAL(lowest, bottom);
+  BOOST_CHECK_EQUAL(highest, top - 1);
+}
+
+BOOST_AUTO_TEST_CASE(rand_incl_int_returns_number_in_range)
+{
+  const int bottom = 10, top = 50;
+  for (int i=0; i<100; i++)
+  {
+    const int random_number = rand_incl(bottom, top);
+    BOOST_CHECK(random_number >= 0);
+    BOOST_CHECK(random_number <= top);
+  }
+}
+
+BOOST_AUTO_TEST_CASE(rand_incl_int_covers_full_range)
+{
+  // The spread of random numbers really goes all the way from 0 to "top"
+  // inclusive.  It's not some smaller subset.
+  // This test will randomly fail sometimes, though very very rarely, when the
+  // random numbers don't actually have enough different values.
+  const int bottom = 1, top = 4;
+  int lowest = 99, highest = -1;
+  for (int i=0; i<100; i++)
+  {
+    const int random_number = rand_incl(bottom, top);
+    lowest = std::min(lowest, random_number);
+    highest = std::max(highest, random_number);
+  }
+
+  BOOST_CHECK_EQUAL(lowest, bottom);
+  BOOST_CHECK_EQUAL(highest, top);
+}
+
+BOOST_AUTO_TEST_CASE(rand_excl_float_returns_float_in_range)
+{
+  const float bottom = 5, top = 10;
+  for (int i=0; i<100; i++)
+  {
+    const float random_number = rand_excl(bottom, top);
+    BOOST_CHECK(random_number >= bottom);
+    BOOST_CHECK(random_number < top);
+  }
+}
+
+BOOST_AUTO_TEST_CASE(rand_excl_float_returns_different_values)
+{
+  const float bottom = 5, top = 10;
+  float lowest = 99, highest = -1;
+  for (int i=0; i<10; i++)
+  {
+    const float random_number = rand_excl(bottom, top);
+    lowest = std::min(lowest, random_number);
+    highest = std::max(highest, random_number);
+  }
+  BOOST_CHECK(lowest < highest);
+}
+
+BOOST_AUTO_TEST_CASE(rand_float_incl_returns_float_in_range)
+{
+  const float bottom = 5, top = 10;
+  for (int i=0; i<1000; i++)
+  {
+    const float random_number = rand_excl(bottom, top);
+    BOOST_CHECK(random_number >= bottom);
+    BOOST_CHECK(random_number <= top);
+  }
+}
+
+BOOST_AUTO_TEST_CASE(rand_float_incl_returns_different_values)
+{
+  const float bottom = 0, top = 10;
+  float lowest = 99, highest = -1;
+  for (int i=0; i<10; i++)
+  {
+    const float random_number = rand_excl(bottom, top);
+    lowest = std::min(lowest, random_number);
+    highest = std::max(highest, random_number);
+  }
+  BOOST_CHECK(lowest < highest);
+}
+
+BOOST_AUTO_TEST_CASE(wide_rand_int_returns_different_numbers_in_range)
+{
+  for (int i=0; i<100; i++)
+  {
+    const int random_number = wide_rand<int>();
+    BOOST_CHECK(random_number >= 0);
+    BOOST_CHECK(random_number <= RAND_MAX);
+  }
+}
+
+BOOST_AUTO_TEST_CASE(wide_rand_long_long_returns_big_numbers)
+{
+  long long one = wide_rand<long long>(), two = wide_rand<long long>();
+  // This test will fail sometimes because of unlucky random numbers, but only
+  // very very rarely.
+  BOOST_CHECK(one > RAND_MAX || two > RAND_MAX);
+}
+
+BOOST_AUTO_TEST_CASE(wide_rand_excl_supports_larger_range)
+{
+  const long long top = 1000 * (long long)RAND_MAX;
+  long long
+    one = wide_rand_excl<long long>(top),
+    two = wide_rand_excl<long long>(top);
+  BOOST_CHECK(one < top);
+  BOOST_CHECK(two < top);
+  // This test will fail sometimes because of unlucky random numbers, but only
+  // very very rarely.
+  BOOST_CHECK(one > RAND_MAX || two > RAND_MAX);
+}
+
 } // namespace
 } // namespace util