string: add tokeniser

Makefile.am

@@ -399,6 +399,7 @@ TEST_BIN = 			\
     test/sha1			\
     test/sha2			\
     test/signal			\
+    test/string			\
     test/stringid		\
     test/strongdef		\
     test/tuple			\

string.cpp

@@ -11,15 +11,16 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  *
- * Copyright 2011 Danny Robson <danny@nerdcruft.net>
+ * Copyright 2011-2016 Danny Robson <danny@nerdcruft.net>
  */
 
 #include "./string.hpp"
 
-#include "./cast.hpp"
-
 #include <cstring>
 
+using util::tokeniser;
+
+///////////////////////////////////////////////////////////////////////////////
 // TODO: Horribly inefficient, but God help you if you're relying on this
 // being efficient in the first place.
 bool
@@ -27,3 +28,89 @@ strbegins (const char *restrict str,
            const char *restrict prefix) {
     return 0 == strncmp (prefix, str, strlen (prefix));
 }
+
+
+///////////////////////////////////////////////////////////////////////////////
+tokeniser::tokeniser (const string_type &_value,
+                      value_type _separator):
+    m_value (_value),
+    m_separator (_separator)
+{ }
+
+
+//-----------------------------------------------------------------------------
+tokeniser::iterator
+tokeniser::cbegin (void) const
+{
+    return iterator (
+        {m_value.cbegin (), m_value.cend ()},
+        m_separator
+    );
+}
+
+
+//-----------------------------------------------------------------------------
+tokeniser::iterator
+tokeniser::cend (void) const
+{
+    return iterator (
+        {m_value.cend (), m_value.cend ()},
+        m_separator
+    );
+}
+
+
+///////////////////////////////////////////////////////////////////////////////
+tokeniser::iterator::iterator (range_type _range, value_type _separator):
+    m_separator (_separator),
+    m_range (_range.cbegin (),
+             std::find (_range.cbegin (),
+                        _range.cend (),
+                        _separator)),
+    m_end (_range.cend ())
+{ }
+
+
+//-----------------------------------------------------------------------------
+tokeniser::iterator
+tokeniser::iterator::operator++ (int)
+{
+    iterator res(*this);
+    ++*this;
+    return res;
+}
+
+
+//-----------------------------------------------------------------------------
+tokeniser::iterator&
+tokeniser::iterator::operator++ (void)
+{
+    auto newend = m_range.cend ();
+    if (newend != m_end) {
+        CHECK_EQ (*m_range.cend (), m_separator);
+        newend++;
+    }
+
+    m_range = {
+        newend,
+        std::find (newend, m_end, m_separator)
+    };
+
+    return *this;
+}
+
+
+//-----------------------------------------------------------------------------
+tokeniser::iterator::range_type
+tokeniser::iterator::operator* (void) const
+{
+    return m_range;
+}
+
+
+//-----------------------------------------------------------------------------
+bool
+tokeniser::iterator::operator== (const iterator &rhs) const
+{
+    return m_range == rhs.m_range && m_separator == rhs.m_separator;
+}

string.hpp

@@ -11,15 +11,59 @@
  * See the License for the specific language governing permissions and
  * limitations under the License.
  *
- * Copyright 2011 Danny Robson <danny@nerdcruft.net>
+ * Copyright 2011-2016 Danny Robson <danny@nerdcruft.net>
  */
 
 #ifndef __UTIL_STRING_HPP
 #define __UTIL_STRING_HPP
 
+#include "./view.hpp"
+
+
 bool
 strbegins(const char *restrict str,
           const char *restrict prefix);
 
+
+namespace util {
+    struct tokeniser {
+    public:
+        using string_type = std::string;
+        using value_type  = string_type::value_type;
+
+        tokeniser (const std::string &value, char separator);
+        tokeniser (std::string &&value, char separator) = delete;
+
+        struct iterator {
+        public:
+            using range_type = util::view<string_type::const_iterator>;
+
+            iterator (range_type range, char separator);
+
+            iterator  operator++ (int);
+            iterator& operator++ (void);
+
+            range_type operator* (void) const;
+
+            bool operator== (const iterator&) const;
+
+        private:
+            const value_type  m_separator;
+            range_type m_range;
+            string_type::const_iterator m_end;
+        };
+
+        iterator cbegin (void) const;
+        iterator begin  (void) const;
+
+        iterator cend (void) const;
+        iterator end  (void) const;
+
+    private:
+        const string_type &m_value;
+        const value_type m_separator;
+    };
+}
+
 #endif // __UTIL_STRING_HPP
 

test/string.cpp

@@ -0,0 +1,35 @@
+#include <cruft/util/tap.hpp>
+#include <cruft/util/string.hpp>
+#include <cruft/util/types.hpp>
+
+int
+main (int, char**)
+{
+    util::TAP::logger tap;
+    
+    const char csv[] = "\0,a,123,,this is a test,";
+    const std::string values[] = {
+        { "\0", 1 },
+        { "a" },
+        { "123" },
+        { "" },
+        { "this is a test" },
+        { "" }
+    };
+
+    std::string str (std::cbegin (csv), std::cbegin (csv) + elems (csv));
+    auto tok = util::tokeniser (str, ',');
+    auto t_cursor = tok.cbegin ();
+    auto v_cursor = std::cbegin (values);
+
+    tap.expect_eq (*t_cursor++, *v_cursor++, "tokeniser, single letter");
+    tap.expect_eq (*t_cursor++, *v_cursor++, "tokeniser, three digits");
+    tap.expect_eq (*t_cursor++, *v_cursor++, "tokeniser, embedded null");
+    tap.expect_eq (*t_cursor++, *v_cursor++, "tokeniser, empty string");
+    tap.expect_eq (*t_cursor++, *v_cursor++, "tokeniser, string with spaces");
+    tap.expect_eq (*t_cursor++, *v_cursor++, "tokeniser, trailing empty");
+
+    tap.expect_eq (t_cursor, tok.cend (), "tokeniser iterator at end");
+
+    return tap.status ();
+}