Fixed compilation error and detected iterator problem.

cxxmph/Makefile.am

@@ -1,4 +1,5 @@
-bin_PROGRAMS = cmph_hash_map_test mphtable_test trigraph_test
+noinst_PROGRAMS = cmph_hash_map_test mphtable_test trigraph_test
+bin_PROGRAMS = cxxmph
 lib_LTLIBRARIES = libcxxmph.la
 include_HEADERS = cmph_hash_map.h mphtable.h MurmurHash2.h trigraph.h cmph_hash_function.h stringpiece.h
 
@@ -13,3 +14,6 @@ mphtable_test_SOURCES = mphtable_test.cc
 
 trigraph_test_LDADD   = libcxxmph.la
 trigraph_test_SOURCES = trigraph_test.cc
+
+cxxmph_LDADD   = libcxxmph.la
+cxxmph_SOURCES = cxxmph.cc

cxxmph/cmph_hash_map.h

@@ -5,27 +5,14 @@
 #include "MurmurHash2.h"
 #include "mphtable.h"
 
-namespace __gnu_cxx {
-template <> struct hash<std::string> {
-  std::size_t operator()(std::string const& s) const {
-    return MurmurHash2(s.c_str(), s.length(), 1 /* seed */);
-  }
-};
-template <> struct hash<long long int> {
-  std::size_t operator()(const long long int& s) const {
-    return MurmurHash2(reinterpret_cast<const char*>(&s), sizeof(long long int), 1 /* seed */);
-  }
-};
-} // namespace __gnu_cxx
-
 namespace cxxmph {
 
 // Save on repetitive typing.
-#define CMPH_TMPL_SPEC template <class Key, class Data, class HashFcn, class EqualKey, class Alloc> 
+#define CMPH_TMPL_SPEC template <class Key, class Data, class HashFcn, class EqualKey, class Alloc>
 #define CMPH_CLASS_SPEC cmph_hash_map<Key, Data, HashFcn, EqualKey, Alloc>
 #define CMPH_METHOD_DECL(r, m) CMPH_TMPL_SPEC typename CMPH_CLASS_SPEC::r CMPH_CLASS_SPEC::m
 
-template <class Key, class Data, class HashFcn = __gnu_cxx::hash<Key>, class EqualKey = std::equal_to<Key>, class Alloc = std::allocator<Data> > 
+template <class Key, class Data, class HashFcn = __gnu_cxx::hash<Key>, class EqualKey = std::equal_to<Key>, class Alloc = std::allocator<Data> >
 class cmph_hash_map {
  public:
   typedef Key key_type;
@@ -132,10 +119,10 @@ CMPH_METHOD_DECL(const_iterator, begin)() const { return values_.begin(); }
 CMPH_METHOD_DECL(const_iterator, end)() const { return values_.end(); }
 CMPH_METHOD_DECL(bool_type, empty)() const { return values_.empty(); }
 
-CMPH_METHOD_DECL(void_type, clear)() { 
+CMPH_METHOD_DECL(void_type, clear)() {
   values_.clear();
   slack_.clear();
-  table_.clear(); 
+  table_.clear();
 }
 
 CMPH_METHOD_DECL(void_type, erase)(iterator pos) {
@@ -163,6 +150,8 @@ CMPH_METHOD_DECL(const_iterator, find)(const key_type& k) const {
 CMPH_METHOD_DECL(iterator, find)(const key_type& k) {
   if (!slack_.empty()) {
     typename slack_type::const_iterator it = slack_.find(k);
+    // TODO(davi) this is broken, it->second should be an integer
+    // otherwise I cannot access values_ iterators.
     if (it != slack_.end()) return values_.begin() + it->second;
   }
   if (table_.size() == 0) return end();
@@ -172,7 +161,6 @@ CMPH_METHOD_DECL(iterator, find)(const key_type& k) {
   }
   return end();
 }
-  
 
 CMPH_METHOD_DECL(data_type&, operator[])(const key_type& k) {
   return insert(std::make_pair(k, data_type())).first->second;

cxxmph/cmph_hash_map_test.cc

@@ -1,5 +1,6 @@
 #include "cmph_hash_map.h"
 
+#include <cstdio>
 #include <cstdlib>
 #include <iostream>
 #include <string>
@@ -28,4 +29,13 @@ int main(int argc, char** argv) {
        std::cerr << "Search " << i - 1 << " gives " << h.find(buf)->second << std::endl;
     }
   }
+  for (int j = 0; j < 100; ++j) {
+    for (int i = 1000; i > 0; --i) {
+       char buf[10];    
+       snprintf(buf, 10, "%d", i*100 - 1);
+       h.find(buf);
+       std::cerr << "Search " << i*100 - 1 << " gives " << h.find(buf)->second << std::endl;
+    }
+  }
+
 }

cxxmph/cxxmph.cc

@@ -0,0 +1,68 @@
+// Copyright 2010 Google Inc. All Rights Reserved.
+// Author: davi@google.com (Davi Reis)
+
+#include <getopt.h>
+
+#include <fstream>
+#include <iostream>
+#include <string>
+#include <vector>
+
+#include "cmph_hash_map.h"
+#include "config.h"
+
+using std::cerr;
+using std::cout;
+using std::endl;
+using std::getline;
+using std::ifstream;
+using std::string;
+using std::vector;
+
+using cxxmph::cmph_hash_map;
+
+void usage(const char* prg) {
+  cerr << "usage: " << prg << "[-v] [-h] [-V]" << endl;
+}
+void usage_long(const char* prg) {
+  usage(prg);
+  cerr << "   -h\t print this help message" << endl;
+  cerr << "   -V\t print version number and exit" << endl;
+  cerr << "   -v\t increase verbosity (may be used multiple times)" << endl;
+}
+
+int main(int argc, char** argv) {
+
+  int verbosity = 0;
+  while (1) {
+    char ch = (char)getopt(argc, argv, "hv");
+    if (ch == -1) break;
+    switch (ch) {
+      case 'h':
+        usage_long(argv[0]);
+        return 0;
+      case 'V':
+        std::cout << VERSION << std::endl;
+        return 0;
+      case 'v':
+        ++verbosity;
+        break;
+    }
+  }
+  if (optind != argc - 1) {
+    usage(argv[0]);
+    return 1;
+  }
+  vector<string> keys;
+  ifstream f(argv[optind]);
+  string buffer;
+  while (!getline(f, buffer).eof()) keys.push_back(buffer);
+  cmph_hash_map<const char*, string> table;
+  for (int i = 0; i < keys.size(); ++i) table[keys[i].c_str()] = keys[i];
+  cmph_hash_map<const char*, string>::const_iterator it = table.begin();
+  cmph_hash_map<const char*, string>::const_iterator end = table.end();
+  for (; it != end; ++it) {
+    cout << (it - table.begin()) << ": " << it->first
+         <<" -> " << it->second << endl;
+  }
+}