Merge branch 'master' of ssh://cmph.git.sourceforge.net/gitroot/cmph/cmph

cxxmph/Makefile.am

@@ -1,9 +1,9 @@
 TESTS = $(check_PROGRAMS)
-check_PROGRAMS = seeded_hash_test mph_bits_test hollow_iterator_test mph_map_test mph_index_test trigraph_test map_tester_test
+check_PROGRAMS = seeded_hash_test mph_bits_test hollow_iterator_test mph_map_test mph_index_test trigraph_test map_tester_test string_util_test
 noinst_PROGRAMS = bm_index bm_map
 bin_PROGRAMS = cxxmph
 lib_LTLIBRARIES = libcxxmph.la
-libcxxmph_la_SOURCES = MurmurHash3.h MurmurHash3.cpp trigragh.h trigraph.cc mph_bits.h mph_bits.cc mph_index.h mph_index.cc seeded_hash.h stringpiece.h benchmark.h benchmark.cc
+libcxxmph_la_SOURCES = MurmurHash3.h MurmurHash3.cpp trigragh.h trigraph.cc mph_bits.h mph_bits.cc mph_index.h mph_index.cc seeded_hash.h stringpiece.h benchmark.h benchmark.cc string_util.cc
 libcxxmph_la_LDFLAGS = -version-info 0:0:0
 cxxmph_includedir = $(includedir)/cxxmph/
 cxxmph_include_HEADERS = mph_map.h mph_index.h MurmurHash3.h trigraph.h seeded_hash.h stringpiece.h hollow_iterator.h
@@ -34,5 +34,8 @@ seeded_hash_test_LDADD   = libcxxmph.la
 mph_bits_test_SOURCES = mph_bits_test.cc
 mph_bits_test_LDADD   = libcxxmph.la
 
+string_util_test_SOURCES = string_util_test.cc
+string_util_test_LDADD   = libcxxmph.la
+
 map_tester_test_SOURCES = map_tester.cc map_tester_test.cc
 

cxxmph/bm_index.cc

@@ -44,10 +44,12 @@ class BM_MPHIndexSearch : public SearchUrlsBenchmark {
   BM_MPHIndexSearch(const std::string& urls_file, int nsearches)
       : SearchUrlsBenchmark(urls_file, nsearches, 0) { }
   virtual void Run() {
+    uint64_t sum = 0;
     for (auto it = random_.begin(); it != random_.end(); ++it) {
       auto idx = index_.index(*it);
       // Collision check to be fair with STL
       if (strcmp(urls_[idx].c_str(), it->data()) != 0) idx = -1;
+      sum += idx;
     }
   }
  protected:
@@ -65,10 +67,12 @@ class BM_CmphIndexSearch : public SearchUrlsBenchmark {
       : SearchUrlsBenchmark(urls_file, nsearches, 0) { }
   ~BM_CmphIndexSearch() { if (index_) cmph_destroy(index_); }
   virtual void Run() {
+    uint64_t sum = 0;
     for (auto it = random_.begin(); it != random_.end(); ++it) {
       auto idx = cmph_search(index_, it->data(), it->length());
       // Collision check to be fair with STL
       if (strcmp(urls_[idx].c_str(), it->data()) != 0) idx = -1;
+      sum += idx;
     }
   }
  protected:
@@ -114,8 +118,10 @@ class BM_STLIndexSearch : public SearchUrlsBenchmark {
   BM_STLIndexSearch(const std::string& urls_file, int nsearches)
       : SearchUrlsBenchmark(urls_file, nsearches, 0) { }
   virtual void Run() {
+    uint64_t sum = 0;
     for (auto it = random_.begin(); it != random_.end(); ++it) {
       auto idx = index_.find(*it);
+      sum += idx->second;
     }
   }
  protected:

cxxmph/bm_map.cc

@@ -1,5 +1,5 @@
 #include <string>
-#include <tr1/unordered_map>
+#include <unordered_map>
 
 #include "bm_common.h"
 #include "mph_map.h"

cxxmph/map_tester.cc

@@ -1,8 +1,4 @@
 #include "map_tester.h"
 
 namespace cxxxmph {
-
-MapTester::MapTester() {}
-MapTester::~MapTester() {}
-
 }

cxxmph/map_tester.h

@@ -7,15 +7,18 @@
 #include <vector>
 #include <unordered_map>
 
+#include "string_util.h"
+
 namespace cxxmph {
 
 using namespace std;
 
-// template <template <class K, class V> class unordered_map> >
+template <template<typename...> class map_type>
 class MapTester {
  public:
-  MapTester();
+  MapTester() {}
-  ~MapTester();
+  ~MapTester() {}
+
   bool Run(string* errors) const {
     string e;
     if (!small_insert()) e += "small insert failed\n"; 
@@ -27,26 +30,26 @@ class MapTester {
     return !e.empty();
   }
   static bool small_insert() {
-    unordered_map<int64_t, int64_t> m;
+    map_type<int64_t, int64_t> m;
     // Start counting from 1 to not touch default constructed value bugs
     for (int i = 1; i < 12; ++i) m.insert(make_pair(i, i));
     return m.size() == 11;
   }
   static bool large_insert() {
-    unordered_map<int64_t, int64_t> m;
+    map_type<int64_t, int64_t> m;
     // Start counting from 1 to not touch default constructed value bugs
     for (int i = 1; i < 12 * 256 * 256; ++i) m.insert(make_pair(i, i));
     return m.size() == 12 * 256 * 256 - 1;
   }
   static bool small_search() {
-    unordered_map<int64_t, int64_t> m;
+    map_type<int64_t, int64_t> m;
     // Start counting from 1 to not touch default constructed value bugs
     for (int i = 1; i < 12; ++i) m.insert(make_pair(i, i));
     for (int i = 1; i < 12; ++i) if (m.find(i) == m.end()) return false; 
     return true;
   }
   static bool default_search() {
-    unordered_map<int64_t, int64_t> m;
+    map_type<int64_t, int64_t> m;
     if (m.find(0) != m.end()) return false;
     for (int i = 1; i < 256; ++i) m.insert(make_pair(i, i));
     if (m.find(0) != m.end()) return false;
@@ -56,7 +59,7 @@ class MapTester {
   }
   static bool large_search() {
     int nkeys = 10 * 1000;
-    unordered_map<int64_t, int64_t> m;
+    map_type<int64_t, int64_t> m;
     for (int i = 0; i < nkeys; ++i) m.insert(make_pair(i, i));
     for (int i = 0; i < nkeys; ++i) if (m.find(i) == m.end()) return false; 
     return true;
@@ -66,10 +69,10 @@ class MapTester {
     vector<string> keys;
     for (int i = 0; i < nkeys; ++i) {
       char buf[128];
-      snprintf(buf, sizeof(buf), "%lu", i);
+      cxxmph::format("%v", i);
       keys.push_back(buf);
     }
-    unordered_map<string, int64_t> m;
+    map_type<string, int64_t> m;
     for (int i = 0; i < nkeys; ++i) m.insert(make_pair(keys[i], i));
     for (int i = 0; i < nkeys; ++i) {
       auto it = m.find(keys[i]);

cxxmph/string_util.cc

@@ -0,0 +1,23 @@
+#include "string_util.h"
+
+#include <cassert>
+#include <cstdint>
+#include <iostream>
+#include <string>
+
+using namespace std;
+
+namespace cxxmph {
+
+bool stream_printf(
+    const std::string& format_string, uint32_t offset, std::ostream* out) {
+  if (offset == format_string.length()) return true;
+  assert(offset < format_string.length());
+  cerr << "length:" << format_string.length() << endl;
+  cerr << "offset:" << offset << endl;
+  auto txt = format_string.substr(offset, format_string.length() - offset);
+  *out << txt;
+  return true;
+}
+
+}  // namespace cxxmph

cxxmph/string_util.h

@@ -0,0 +1,137 @@
+#ifndef __CXXMPH_STRING_UTIL_H__
+#define __CXXMPH_STRING_UTIL_H__
+
+// Helper functions for string formatting and terminal output. Should be used
+// only for debugging and tests, since performance was not a concern.
+// Implemented using variadic templates because it is cool.
+//
+// Adds the extra format %v to the printf formatting language. Uses the method
+// cxxmph::tostr to implement custom printers and fallback to operator
+// ostream::operator<< otherwise.
+
+#include <cstdint>
+#include <cstdio>
+#include <cstring>
+#include <iostream>
+#include <string>
+#include <sstream>
+#include <utility>
+#include <vector>
+
+#define CXXMPH_DEBUGLN(fmt) variadic_print(__FILE__, __LINE__, &std::cerr, fmt)
+#define CXXMPH_INFOLN(fmt) variadic_print(__FILE__, __LINE__, &std::cout, fmt)
+
+namespace cxxmph {
+
+using std::pair;
+using std::string;
+using std::ostream;
+using std::vector;
+
+template <class T> void tostr(ostream *out, const T& v) {
+  *out << v;
+}
+inline void tostr(std::ostream* out, uint8_t v) {
+  *out << static_cast<uint32_t>(v);
+}
+template <class V>
+inline void tostr(ostream* out, const vector<V>& v) {
+  *out << "[";
+  for (uint32_t i = 0; i < v.size(); ++i) {
+    tostr(out, v[1]);
+    if (i != v.size() - 1)*out << " ";
+  }
+  *out << "]";
+}
+template <class F, class S>
+inline void tostr(ostream* out, const pair<F, S>& v) {
+  *out << "(";
+  tostr(out, v.first);
+  *out << ",";
+  tostr(out, v.second);
+  *out << ")";
+}
+
+template <typename... Args, class T >
+std::string infoln(const std::string& format_string, Args&&... args) {
+  return stream_printf(format_string, 0, &std::cout, std::forward<Args>(args)...);
+}
+
+bool stream_printf(
+    const std::string& format_string, uint32_t offset, std::ostream* out);
+
+template <bool ispod> struct pod_snprintf {};
+template <> struct pod_snprintf<false> {
+  template <class T>
+  int operator()(char*, size_t, const char*, const T&) {
+     return -1;
+  }
+};
+template <> struct pod_snprintf<true> {
+  template <class T>
+  int operator()(char* str, size_t size, const char* format, const T& v) {
+    return snprintf(str, size, format, v);
+  }
+};
+
+template <typename T, typename... Args>
+bool stream_printf(const std::string& format_string, uint32_t offset,
+                   std::ostream* out, const T& value, Args&&... args) {
+  auto b = format_string.c_str() + offset;
+  auto txt = format_string.c_str() + offset;
+  for (; *txt; ++txt) {
+    if (*txt == '%') {  
+      if (*(txt + 1) != '%') break;
+      *out << "%";
+      return stream_printf(format_string, offset + 2, out, value,
+                           std::forward<Args>(args)...);
+    }
+  }
+  if (txt != b) {
+    *out << string(b, txt - b);
+  }
+  auto fmt = txt + 1;
+  while (*fmt && *fmt != '%') ++fmt;
+  if (strncmp(txt, "%v", 2) == 0) {
+     txt += 2;
+     tostr(out, value);
+     if (txt != fmt) *out << string(txt, fmt);
+  } else {
+    char buf[256];  // Is this enough?
+    auto n = pod_snprintf<std::is_pod<T>::value>()(
+        buf, 256, std::string(txt, fmt).c_str(), value);
+    if (n < 0) return false;
+    *out << buf;
+  }
+  return stream_printf(format_string, fmt - format_string.c_str(), out,
+                       std::forward<Args>(args)...);
+}
+
+template <typename... Args>
+std::string format(const std::string& format_string, Args&&... args) {
+  std::ostringstream out;
+  if (!stream_printf(format_string, 0, &out, std::forward<Args>(args)...)) {
+    return std::string();
+  };
+  return out.str();
+}
+
+struct variadic_print {
+  variadic_print(const std::string& file, uint32_t line, std::ostream* out,
+                 const std::string& format_line) 
+     : file_(file), line_(line), out_(out), format_line_(format_line) {}
+  template <typename... Args>
+  void operator()(Args&&... args) {
+    std::string fancy_format = "%s:%d: ";
+    fancy_format += format_line_;
+    stream_printf(fancy_format, 0, out_, std::forward<Args>(args)...);
+  }
+  const std::string& file_;
+  const uint32_t& line_;
+  std::ostream* out_;
+  const std::string& format_line_;
+};
+
+}  // namespace cxxmph
+
+#endif // __CXXMPH_STRING_UTIL_H__

src/bdz.c

@@ -47,7 +47,7 @@ typedef cmph_uint32 * bdz_queue_t;
 
 static void bdz_alloc_queue(bdz_queue_t * queuep, cmph_uint32 nedges)
 {
-	(*queuep)=malloc(nedges*sizeof(cmph_uint32));
+	(*queuep)=(cmph_uint32 *)malloc(nedges*sizeof(cmph_uint32));
 };
 static void bdz_free_queue(bdz_queue_t * queue)
 {
@@ -65,9 +65,9 @@ typedef struct
 
 static void bdz_alloc_graph3(bdz_graph3_t * graph3, cmph_uint32 nedges, cmph_uint32 nvertices)
 {
-	graph3->edges=malloc(nedges*sizeof(bdz_edge_t));
+	graph3->edges=(bdz_edge_t *)malloc(nedges*sizeof(bdz_edge_t));
-	graph3->first_edge=malloc(nvertices*sizeof(cmph_uint32));
+	graph3->first_edge=(cmph_uint32 *)malloc(nvertices*sizeof(cmph_uint32));
-	graph3->vert_degree=malloc((size_t)nvertices);
+	graph3->vert_degree=(cmph_uint8 *)malloc((size_t)nvertices);
 };
 static void bdz_init_graph3(bdz_graph3_t * graph3, cmph_uint32 nedges, cmph_uint32 nvertices)
 {
@@ -163,7 +163,7 @@ static int bdz_generate_queue(cmph_uint32 nedges, cmph_uint32 nvertices, bdz_que
 	cmph_uint32 queue_head=0,queue_tail=0;
 	cmph_uint32 curr_edge;
 	cmph_uint32 tmp_edge;
-	cmph_uint8 * marked_edge =malloc((size_t)(nedges >> 3) + 1);
+	cmph_uint8 * marked_edge = (cmph_uint8 *)malloc((size_t)(nedges >> 3) + 1);
 	memset(marked_edge, 0, (size_t)(nedges >> 3) + 1);
 
 	for(i=0;i<nedges;i++){
@@ -424,7 +424,7 @@ static void assigning(bdz_config_data_t *bdz, bdz_graph3_t* graph3, bdz_queue_t
 	cmph_uint32 nedges=graph3->nedges;
 	cmph_uint32 curr_edge;
 	cmph_uint32 v0,v1,v2;
-	cmph_uint8 * marked_vertices =malloc((size_t)(bdz->n >> 3) + 1);
+	cmph_uint8 * marked_vertices = (cmph_uint8 *)malloc((size_t)(bdz->n >> 3) + 1);
         cmph_uint32 sizeg = (cmph_uint32)ceil(bdz->n/4.0);
 	bdz->g = (cmph_uint8 *)calloc((size_t)(sizeg), sizeof(cmph_uint8));
 	memset(marked_vertices, 0, (size_t)(bdz->n >> 3) + 1);
@@ -595,7 +595,7 @@ static inline cmph_uint32 rank(cmph_uint32 b, cmph_uint32 * ranktable, cmph_uint
 cmph_uint32 bdz_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
 	register cmph_uint32 vertex;
-	register bdz_data_t *bdz = mphf->data;
+	register bdz_data_t *bdz = (bdz_data_t *)mphf->data;
 	cmph_uint32 hl[3];
 	hash_vector(bdz->hl, key, keylen, hl);
 	hl[0] = hl[0] % bdz->r;
@@ -625,7 +625,7 @@ void bdz_destroy(cmph_t *mphf)
 void bdz_pack(cmph_t *mphf, void *packed_mphf)
 {
 	bdz_data_t *data = (bdz_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 
 	// packing hl type
 	CMPH_HASH hl_type = hash_get_type(data->hl);
@@ -681,7 +681,7 @@ cmph_uint32 bdz_search_packed(void *packed_mphf, const char *key, cmph_uint32 ke
 {
 
 	register cmph_uint32 vertex;
-	register CMPH_HASH hl_type  = *(cmph_uint32 *)packed_mphf;
+	register CMPH_HASH hl_type  = (CMPH_HASH)(*(cmph_uint32 *)packed_mphf);
 	register cmph_uint8 *hl_ptr = (cmph_uint8 *)(packed_mphf) + 4;
 
 	register cmph_uint32 *ranktable = (cmph_uint32*)(hl_ptr + hash_state_packed_size(hl_type));

src/bdz_ph.c

@@ -34,7 +34,7 @@ typedef cmph_uint32 * bdz_ph_queue_t;
 
 static void bdz_ph_alloc_queue(bdz_ph_queue_t * queuep, cmph_uint32 nedges)
 {
-	(*queuep)=malloc(nedges*sizeof(cmph_uint32));
+	(*queuep)=(cmph_uint32 *)malloc(nedges*sizeof(cmph_uint32));
 };
 static void bdz_ph_free_queue(bdz_ph_queue_t * queue)
 {
@@ -52,9 +52,9 @@ typedef struct
 
 static void bdz_ph_alloc_graph3(bdz_ph_graph3_t * graph3, cmph_uint32 nedges, cmph_uint32 nvertices)
 {
-	graph3->edges=malloc(nedges*sizeof(bdz_ph_edge_t));
+	graph3->edges=(bdz_ph_edge_t *)malloc(nedges*sizeof(bdz_ph_edge_t));
-	graph3->first_edge=malloc(nvertices*sizeof(cmph_uint32));
+	graph3->first_edge=(cmph_uint32 *)malloc(nvertices*sizeof(cmph_uint32));
-	graph3->vert_degree=malloc((size_t)nvertices);
+	graph3->vert_degree=(cmph_uint8 *)malloc((size_t)nvertices);
 };
 static void bdz_ph_init_graph3(bdz_ph_graph3_t * graph3, cmph_uint32 nedges, cmph_uint32 nvertices)
 {
@@ -148,7 +148,7 @@ static int bdz_ph_generate_queue(cmph_uint32 nedges, cmph_uint32 nvertices, bdz_
 	cmph_uint32 queue_head=0,queue_tail=0;
 	cmph_uint32 curr_edge;
 	cmph_uint32 tmp_edge;
-	cmph_uint8 * marked_edge =malloc((size_t)(nedges >> 3) + 1);
+	cmph_uint8 * marked_edge =(cmph_uint8 *)malloc((size_t)(nedges >> 3) + 1);
 	memset(marked_edge, 0, (size_t)(nedges >> 3) + 1);
 
 	for(i=0;i<nedges;i++){
@@ -381,7 +381,7 @@ static void assigning(bdz_ph_config_data_t *bdz_ph, bdz_ph_graph3_t* graph3, bdz
 	cmph_uint32 nedges=graph3->nedges;
 	cmph_uint32 curr_edge;
 	cmph_uint32 v0,v1,v2;
-	cmph_uint8 * marked_vertices =malloc((size_t)(bdz_ph->n >> 3) + 1);
+	cmph_uint8 * marked_vertices = (cmph_uint8 *)malloc((size_t)(bdz_ph->n >> 3) + 1);
 	cmph_uint32 sizeg = (cmph_uint32)ceil(bdz_ph->n/4.0);
 	bdz_ph->g = (cmph_uint8 *)calloc((size_t)sizeg, sizeof(cmph_uint8));
 	memset(marked_vertices, 0, (size_t)(bdz_ph->n >> 3) + 1);
@@ -507,7 +507,7 @@ void bdz_ph_load(FILE *f, cmph_t *mphf)
 
 cmph_uint32 bdz_ph_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	register bdz_ph_data_t *bdz_ph = mphf->data;
+	register bdz_ph_data_t *bdz_ph = (bdz_ph_data_t *)mphf->data;
 	cmph_uint32 hl[3];
 	register cmph_uint8 byte0, byte1, byte2;
 	register cmph_uint32 vertex;
@@ -547,7 +547,7 @@ void bdz_ph_destroy(cmph_t *mphf)
 void bdz_ph_pack(cmph_t *mphf, void *packed_mphf)
 {
 	bdz_ph_data_t *data = (bdz_ph_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 
 	// packing hl type
 	CMPH_HASH hl_type = hash_get_type(data->hl);
@@ -590,7 +590,7 @@ cmph_uint32 bdz_ph_packed_size(cmph_t *mphf)
 cmph_uint32 bdz_ph_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
 
-	register CMPH_HASH hl_type  = *(cmph_uint32 *)packed_mphf;
+	register CMPH_HASH hl_type  = (CMPH_HASH)*(cmph_uint32 *)packed_mphf;
 	register cmph_uint8 *hl_ptr = (cmph_uint8 *)(packed_mphf) + 4;
 
 	register cmph_uint8 * ptr = hl_ptr + hash_state_packed_size(hl_type);

src/bm_numbers.c

@@ -44,7 +44,7 @@ void bm_create(CMPH_ALGO algo, int iters) {
   cmph_config_t* config = NULL;
   cmph_t* mphf = NULL;
 
-  if (iters > g_numbers_len) {
+  if (iters > (int)g_numbers_len) {
     fprintf(stderr, "No input with proper size.");
     exit(-1);
   }

src/bmz.c

@@ -526,7 +526,7 @@ void bmz_load(FILE *f, cmph_t *mphf)
 
 cmph_uint32 bmz_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	bmz_data_t *bmz = mphf->data;
+	bmz_data_t *bmz = (bmz_data_t *)mphf->data;
 	cmph_uint32 h1 = hash(bmz->hashes[0], key, keylen) % bmz->n;
 	cmph_uint32 h2 = hash(bmz->hashes[1], key, keylen) % bmz->n;
 	DEBUGP("key: %.*s h1: %u h2: %u\n", keylen, key, h1, h2);
@@ -554,7 +554,7 @@ void bmz_pack(cmph_t *mphf, void *packed_mphf)
 {
 
 	bmz_data_t *data = (bmz_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 
 	// packing h1 type
 	CMPH_HASH h1_type = hash_get_type(data->hashes[0]);
@@ -606,12 +606,12 @@ cmph_uint32 bmz_packed_size(cmph_t *mphf)
  */
 cmph_uint32 bmz_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
-	register cmph_uint8 *h1_ptr = packed_mphf;
+	register cmph_uint8 *h1_ptr = (cmph_uint8 *)packed_mphf;
-	register CMPH_HASH h1_type  = *((cmph_uint32 *)h1_ptr);
+	register CMPH_HASH h1_type  = (CMPH_HASH)(*((cmph_uint32 *)h1_ptr));
 	h1_ptr += 4;
 
 	register cmph_uint8 *h2_ptr = h1_ptr + hash_state_packed_size(h1_type);
-	register CMPH_HASH h2_type  = *((cmph_uint32 *)h2_ptr);
+	register CMPH_HASH h2_type  = (CMPH_HASH)(*((cmph_uint32 *)h2_ptr));
 	h2_ptr += 4;
 
 	register cmph_uint32 *g_ptr = (cmph_uint32 *)(h2_ptr + hash_state_packed_size(h2_type));

src/bmz8.c

@@ -534,7 +534,7 @@ void bmz8_load(FILE *f, cmph_t *mphf)
 
 cmph_uint8 bmz8_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	bmz8_data_t *bmz8 = mphf->data;
+	bmz8_data_t *bmz8 = (bmz8_data_t *)mphf->data;
 	cmph_uint8 h1 = (cmph_uint8)(hash(bmz8->hashes[0], key, keylen) % bmz8->n);
 	cmph_uint8 h2 = (cmph_uint8)(hash(bmz8->hashes[1], key, keylen) % bmz8->n);
 	DEBUGP("key: %s h1: %u h2: %u\n", key, h1, h2);
@@ -561,7 +561,7 @@ void bmz8_destroy(cmph_t *mphf)
 void bmz8_pack(cmph_t *mphf, void *packed_mphf)
 {
 	bmz8_data_t *data = (bmz8_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 
 	// packing h1 type
 	CMPH_HASH h1_type = hash_get_type(data->hashes[0]);
@@ -612,12 +612,12 @@ cmph_uint32 bmz8_packed_size(cmph_t *mphf)
  */
 cmph_uint8 bmz8_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
-	register cmph_uint8 *h1_ptr = packed_mphf;
+	register cmph_uint8 *h1_ptr = (cmph_uint8 *)packed_mphf;
-	register CMPH_HASH h1_type  = *((cmph_uint32 *)h1_ptr);
+	register CMPH_HASH h1_type  = (CMPH_HASH)(*((cmph_uint32 *)h1_ptr));
 	h1_ptr += 4;
 
 	register cmph_uint8 *h2_ptr = h1_ptr + hash_state_packed_size(h1_type);
-	register CMPH_HASH h2_type  = *((cmph_uint32 *)h2_ptr);
+	register CMPH_HASH h2_type  = (CMPH_HASH)(*((cmph_uint32 *)h2_ptr));
 	h2_ptr += 4;
 
 	register cmph_uint8 *g_ptr = h2_ptr + hash_state_packed_size(h2_type);

src/brz.c

@@ -685,7 +685,7 @@ static cmph_uint32 brz_fch_search(brz_data_t *brz, const char *key, cmph_uint32
 
 cmph_uint32 brz_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	brz_data_t *brz = mphf->data;
+	brz_data_t *brz = (brz_data_t *)mphf->data;
 	cmph_uint32 fingerprint[3];
 	switch(brz->algo)
 	{
@@ -728,7 +728,7 @@ void brz_destroy(cmph_t *mphf)
 void brz_pack(cmph_t *mphf, void *packed_mphf)
 {
 	brz_data_t *data = (brz_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 	cmph_uint32 i,n;
 
 	// packing internal algo type
@@ -858,7 +858,7 @@ cmph_uint32 brz_packed_size(cmph_t *mphf)
 
 static cmph_uint32 brz_bmz8_search_packed(cmph_uint32 *packed_mphf, const char *key, cmph_uint32 keylen, cmph_uint32 * fingerprint)
 {
-	register CMPH_HASH h0_type = *packed_mphf++;
+	register CMPH_HASH h0_type = (CMPH_HASH)*packed_mphf++;
 	register cmph_uint32 *h0_ptr = packed_mphf;
 	packed_mphf = (cmph_uint32 *)(((cmph_uint8 *)packed_mphf) + hash_state_packed_size(h0_type));
 
@@ -867,9 +867,9 @@ static cmph_uint32 brz_bmz8_search_packed(cmph_uint32 *packed_mphf, const char *
 	register double c = (double)(*((cmph_uint64*)packed_mphf));
 	packed_mphf += 2;
 
-	register CMPH_HASH h1_type = *packed_mphf++;
+	register CMPH_HASH h1_type = (CMPH_HASH)*packed_mphf++;
 
-	register CMPH_HASH h2_type = *packed_mphf++;
+	register CMPH_HASH h2_type = (CMPH_HASH)*packed_mphf++;
 
 	register cmph_uint8 * size = (cmph_uint8 *) packed_mphf;
 	packed_mphf = (cmph_uint32 *)(size + k);
@@ -911,7 +911,7 @@ static cmph_uint32 brz_bmz8_search_packed(cmph_uint32 *packed_mphf, const char *
 
 static cmph_uint32 brz_fch_search_packed(cmph_uint32 *packed_mphf, const char *key, cmph_uint32 keylen, cmph_uint32 * fingerprint)
 {
-	register CMPH_HASH h0_type = *packed_mphf++;
+	register CMPH_HASH h0_type = (CMPH_HASH)*packed_mphf++;
 
 	register cmph_uint32 *h0_ptr = packed_mphf;
 	packed_mphf = (cmph_uint32 *)(((cmph_uint8 *)packed_mphf) + hash_state_packed_size(h0_type));
@@ -921,9 +921,9 @@ static cmph_uint32 brz_fch_search_packed(cmph_uint32 *packed_mphf, const char *k
 	register double c = (double)(*((cmph_uint64*)packed_mphf));
 	packed_mphf += 2;
 
-	register CMPH_HASH h1_type = *packed_mphf++;
+	register CMPH_HASH h1_type = (CMPH_HASH)*packed_mphf++;
 
-	register CMPH_HASH h2_type = *packed_mphf++;
+	register CMPH_HASH h2_type = (CMPH_HASH)*packed_mphf++;
 
 	register cmph_uint8 * size = (cmph_uint8 *) packed_mphf;
 	packed_mphf = (cmph_uint32 *)(size + k);
@@ -972,7 +972,7 @@ static cmph_uint32 brz_fch_search_packed(cmph_uint32 *packed_mphf, const char *k
 cmph_uint32 brz_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
 	register cmph_uint32 *ptr = (cmph_uint32 *)packed_mphf;
-	register CMPH_ALGO algo = *ptr++;
+	register CMPH_ALGO algo = (CMPH_ALGO)*ptr++;
 	cmph_uint32 fingerprint[3];
 	switch(algo)
 	{

src/chd.c

@@ -105,7 +105,7 @@ cmph_t *chd_new(cmph_config_t *mph, double c)
 	DEBUGP("packed_chd_phf_size = %u\n", packed_chd_phf_size);
 
 	/* Make sure that we have enough space to pack the mphf. */
-	packed_chd_phf = calloc((size_t)packed_chd_phf_size,(size_t)1);
+	packed_chd_phf = (cmph_uint8 *)calloc((size_t)packed_chd_phf_size,(size_t)1);
 
 	/* Pack the mphf. */
 	cmph_pack(chd_phf, packed_chd_phf);
@@ -229,14 +229,14 @@ static inline cmph_uint32 _chd_search(void * packed_chd_phf, void * packed_cr, c
 
 cmph_uint32 chd_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	register chd_data_t * chd = mphf->data;
+	register chd_data_t * chd = (chd_data_t *)mphf->data;
 	return _chd_search(chd->packed_chd_phf, chd->packed_cr, key, keylen);
 }
 
 void chd_pack(cmph_t *mphf, void *packed_mphf)
 {
 	chd_data_t *data = (chd_data_t *)mphf->data;
-	cmph_uint32 * ptr = packed_mphf;
+	cmph_uint32 * ptr = (cmph_uint32 *)packed_mphf;
 	cmph_uint8 * ptr8;
 
 	// packing packed_cr_size and packed_cr
@@ -263,7 +263,7 @@ cmph_uint32 chd_packed_size(cmph_t *mphf)
 cmph_uint32 chd_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
 
-	register cmph_uint32 * ptr = packed_mphf;
+	register cmph_uint32 * ptr = (cmph_uint32 *)packed_mphf;
 	register cmph_uint32 packed_cr_size = *ptr++;
 	register cmph_uint8 * packed_chd_phf = ((cmph_uint8 *) ptr) + packed_cr_size + sizeof(cmph_uint32);
 	return _chd_search(packed_chd_phf, ptr, key, keylen);

src/chd_ph.c

@@ -221,7 +221,7 @@ cmph_uint8 chd_ph_mapping(cmph_config_t *mph, chd_ph_bucket_t * buckets, chd_ph_
 	char * key = NULL;
 	cmph_uint32 keylen = 0;
 	chd_ph_map_item_t * map_item;
-	chd_ph_map_item_t * map_items = malloc(chd_ph->m*sizeof(chd_ph_map_item_t));
+	chd_ph_map_item_t * map_items = (chd_ph_map_item_t *)malloc(chd_ph->m*sizeof(chd_ph_map_item_t));
 	register cmph_uint32 mapping_iterations = 1000;
 	*max_bucket_size = 0;
 	while(1)
@@ -317,7 +317,7 @@ chd_ph_sorted_list_t * chd_ph_ordering(chd_ph_bucket_t ** _buckets, chd_ph_item_
 	};
 	sorted_lists[i-1].size = 0;
 	// Store the buckets in a new array which is sorted by bucket sizes
-	output_buckets = calloc(nbuckets, sizeof(chd_ph_bucket_t)); // everything is initialized with zero
+	output_buckets = (chd_ph_bucket_t *)calloc(nbuckets, sizeof(chd_ph_bucket_t)); // everything is initialized with zero
 //  	non_empty_buckets = nbuckets;
 
 	for(i = 0; i < nbuckets; i++)
@@ -901,7 +901,7 @@ void chd_ph_destroy(cmph_t *mphf)
 
 cmph_uint32 chd_ph_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	register chd_ph_data_t * chd_ph = mphf->data;
+	register chd_ph_data_t * chd_ph = (chd_ph_data_t *)mphf->data;
 	cmph_uint32 hl[3];
 	register cmph_uint32 disp,position;
 	register cmph_uint32 probe0_num,probe1_num;
@@ -921,7 +921,7 @@ cmph_uint32 chd_ph_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 void chd_ph_pack(cmph_t *mphf, void *packed_mphf)
 {
 	chd_ph_data_t *data = (chd_ph_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 
 	// packing hl type
 	CMPH_HASH hl_type = hash_get_type(data->hl);
@@ -959,7 +959,7 @@ cmph_uint32 chd_ph_packed_size(cmph_t *mphf)
 
 cmph_uint32 chd_ph_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
-	register CMPH_HASH hl_type  = *(cmph_uint32 *)packed_mphf;
+	register CMPH_HASH hl_type  = (CMPH_HASH)*(cmph_uint32 *)packed_mphf;
 	register cmph_uint8 *hl_ptr = (cmph_uint8 *)(packed_mphf) + 4;
 
 	register cmph_uint32 * ptr = (cmph_uint32 *)(hl_ptr + hash_state_packed_size(hl_type));

src/chm.c

@@ -276,7 +276,7 @@ void chm_load(FILE *f, cmph_t *mphf)
 
 cmph_uint32 chm_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	chm_data_t *chm = mphf->data;
+	chm_data_t *chm = (chm_data_t *)mphf->data;
 	cmph_uint32 h1 = hash(chm->hashes[0], key, keylen) % chm->n;
 	cmph_uint32 h2 = hash(chm->hashes[1], key, keylen) % chm->n;
 	DEBUGP("key: %s h1: %u h2: %u\n", key, h1, h2);
@@ -303,7 +303,7 @@ void chm_destroy(cmph_t *mphf)
 void chm_pack(cmph_t *mphf, void *packed_mphf)
 {
 	chm_data_t *data = (chm_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 
 	// packing h1 type
 	CMPH_HASH h1_type = hash_get_type(data->hashes[0]);
@@ -359,12 +359,12 @@ cmph_uint32 chm_packed_size(cmph_t *mphf)
  */
 cmph_uint32 chm_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
-	register cmph_uint8 *h1_ptr = packed_mphf;
+	register cmph_uint8 *h1_ptr = (cmph_uint8 *)packed_mphf;
-	register CMPH_HASH h1_type  = *((cmph_uint32 *)h1_ptr);
+	register CMPH_HASH h1_type  = (CMPH_HASH)(*((cmph_uint32 *)h1_ptr));
 	h1_ptr += 4;
 
 	register cmph_uint8 *h2_ptr = h1_ptr + hash_state_packed_size(h1_type);
-	register CMPH_HASH h2_type  = *((cmph_uint32 *)h2_ptr);
+	register CMPH_HASH h2_type  = (CMPH_HASH)(*((cmph_uint32 *)h2_ptr));
 	h2_ptr += 4;
 
 	register cmph_uint32 *g_ptr = (cmph_uint32 *)(h2_ptr + hash_state_packed_size(h2_type));

src/cmph_benchmark.c

@@ -70,7 +70,7 @@ void bm_register(const char* name, void (*func)(int), int iters) {
   benchmark.func = func;
   benchmark.iters = iters;
   assert(!find_benchmark(name));
-  global_benchmarks = realloc(
+  global_benchmarks = (benchmark_t *)realloc(
       global_benchmarks, (length + 2)*sizeof(benchmark_t));
   global_benchmarks[length] = benchmark;
   memset(&benchmark, 0, sizeof(benchmark_t));  // pivot

src/cmph_structs.c

@@ -49,7 +49,7 @@ cmph_t *__cmph_load(FILE *f)
 	{
 		if (strcmp(algo_name, cmph_names[i]) == 0)
 		{
-			algo = i;
+			algo = (CMPH_ALGO)(i);
 		}
 	}
 	if (algo == CMPH_COUNT)

src/fch.c

@@ -398,7 +398,7 @@ void fch_load(FILE *f, cmph_t *mphf)
 
 cmph_uint32 fch_search(cmph_t *mphf, const char *key, cmph_uint32 keylen)
 {
-	fch_data_t *fch = mphf->data;
+	fch_data_t *fch = (fch_data_t *)mphf->data;
 	cmph_uint32 h1 = hash(fch->h1, key, keylen) % fch->m;
 	cmph_uint32 h2 = hash(fch->h2, key, keylen) % fch->m;
 	h1 = mixh10h11h12 (fch->b, fch->p1, fch->p2, h1);
@@ -423,7 +423,7 @@ void fch_destroy(cmph_t *mphf)
 void fch_pack(cmph_t *mphf, void *packed_mphf)
 {
 	fch_data_t *data = (fch_data_t *)mphf->data;
-	cmph_uint8 * ptr = packed_mphf;
+	cmph_uint8 * ptr = (cmph_uint8 *)packed_mphf;
 
 	// packing h1 type
 	CMPH_HASH h1_type = hash_get_type(data->h1);
@@ -488,12 +488,12 @@ cmph_uint32 fch_packed_size(cmph_t *mphf)
  */
 cmph_uint32 fch_search_packed(void *packed_mphf, const char *key, cmph_uint32 keylen)
 {
-	register cmph_uint8 *h1_ptr = packed_mphf;
+	register cmph_uint8 *h1_ptr = (cmph_uint8 *)packed_mphf;
-	register CMPH_HASH h1_type  = *((cmph_uint32 *)h1_ptr);
+	register CMPH_HASH h1_type  = (CMPH_HASH)*((cmph_uint32 *)h1_ptr);
 	h1_ptr += 4;
 
 	register cmph_uint8 *h2_ptr = h1_ptr + hash_state_packed_size(h1_type);
-	register CMPH_HASH h2_type  = *((cmph_uint32 *)h2_ptr);
+	register CMPH_HASH h2_type  = (CMPH_HASH)*((cmph_uint32 *)h2_ptr);
 	h2_ptr += 4;
 
 	register cmph_uint32 *g_ptr = (cmph_uint32 *)(h2_ptr + hash_state_packed_size(h2_type));

src/fch_buckets.c

@@ -172,13 +172,13 @@ cmph_uint32 fch_buckets_get_nbuckets(fch_buckets_t * buckets)
 
 cmph_uint32 * fch_buckets_get_indexes_sorted_by_size(fch_buckets_t * buckets)
 {
-	cmph_uint32 i = 0;
+	cmph_int32 i = 0;
 	cmph_uint32 sum = 0, value;
 	cmph_uint32 *nbuckets_size = (cmph_uint32 *) calloc((size_t)buckets->max_size + 1, sizeof(cmph_uint32));
 	cmph_uint32 * sorted_indexes = (cmph_uint32 *) calloc((size_t)buckets->nbuckets, sizeof(cmph_uint32));
 
 	// collect how many buckets for each size.
-	for(i = 0; i < buckets->nbuckets; i++) nbuckets_size[fch_bucket_size(buckets->values + i)] ++;
+	for(i = 0; i < (int)buckets->nbuckets; i++) nbuckets_size[fch_bucket_size(buckets->values + i)] ++;
 
 	// calculating offset considering a decreasing order of buckets size.
 	value = nbuckets_size[buckets->max_size];
@@ -190,7 +190,7 @@ cmph_uint32 * fch_buckets_get_indexes_sorted_by_size(fch_buckets_t * buckets)
 		nbuckets_size[i] = sum;
 
 	}
-	for(i = 0; i < buckets->nbuckets; i++)
+	for(i = 0; i < (int)buckets->nbuckets; i++)
 	{
 		sorted_indexes[nbuckets_size[fch_bucket_size(buckets->values + i)]] = (cmph_uint32)i;
 		nbuckets_size[fch_bucket_size(buckets->values + i)] ++;

src/hash.c

@@ -101,7 +101,7 @@ hash_state_t *hash_state_load(const char *buf, cmph_uint32 buflen)
 	{
 		if (strcmp(buf, cmph_hash_names[i]) == 0)
 		{
-			hashfunc = i;
+			hashfunc = (CMPH_HASH)(i);
 			break;
 		}
 	}

src/main.c

@@ -177,7 +177,7 @@ int main(int argc, char **argv)
 				{
 					if (strcmp(cmph_names[i], optarg) == 0)
 					{
-						mph_algo = i;
+						mph_algo = (CMPH_ALGO)i;
 						valid = 1;
 						break;
 					}
@@ -197,7 +197,7 @@ int main(int argc, char **argv)
 					if (strcmp(cmph_hash_names[i], optarg) == 0)
 					{
 						hashes = (CMPH_HASH *)realloc(hashes, sizeof(CMPH_HASH) * ( nhashes + 2 ));
-						hashes[nhashes] = i;
+						hashes[nhashes] = (CMPH_HASH)i;
 						hashes[nhashes + 1] = CMPH_HASH_COUNT;
 						++nhashes;
 						valid = 1;

tests/packed_mphf_tests.c

@@ -147,7 +147,7 @@ int main(int argc, char **argv)
 	fprintf(stderr, "packed_size = %u\n", packed_size);
 	
 	/* Make sure that we have enough space to pack the mphf. */
-	cmph_uint8 * packed_mphf = calloc((size_t)packed_size,(size_t)1);
+	cmph_uint8 * packed_mphf = (cmph_uint8 *)calloc((size_t)packed_size,(size_t)1);
 
 	/* Pack the mphf. */
 	cmph_pack(mphf, packed_mphf);