Add ucase(), lcase(), casecmp(), alpha(), digit(), alnum()

10 files changed, 288 insertions, 13 deletions

diff --git a/butl/buildfile b/butl/buildfile
index d2d16ac..7e65b6c 100644
--- a/butl/buildfile
+++ b/butl/buildfile
@@ -20,7 +20,7 @@ lib{butl}:                         \
 {hxx     txx    }{ string-table  } \
 {hxx         cxx}{ timestamp     } \
 {hxx         cxx}{ triplet       } \
-{hxx            }{ utility       } \
+{hxx ixx        }{ utility       } \
 {hxx            }{ vector-view   } \
 {hxx            }{ version       } \
 {hxx         cxx}{ win32-utility }
diff --git a/butl/path b/butl/path
index 7322a15..2040f5f 100644
--- a/butl/path
+++ b/butl/path
@@ -14,6 +14,8 @@
 
 #include <butl/export>
 
+#include <butl/utility>
+
 namespace butl
 {
   // Wish list/ideas for improvements.
@@ -195,7 +197,7 @@ namespace butl
       for (size_type i (0), n (ln < rn ? ln : rn); i != n; ++i)
       {
 #ifdef _WIN32
-        C lc (tolower (l[i])), rc (tolower (r[i]));
+        C lc (lcase (l[i])), rc (lcase (r[i]));
 #else
         C lc (l[i]), rc (r[i]);
 #endif
diff --git a/butl/path.ixx b/butl/path.ixx
index 26a9b3d..d1fa34c 100644
--- a/butl/path.ixx
+++ b/butl/path.ixx
@@ -3,7 +3,6 @@
 // license   : MIT; see accompanying LICENSE file
 
 #ifdef _WIN32
-#  include <cctype>  // tolower(), toupper()
 #  include <cwctype> // towlower(), towupper()
 #endif
 
@@ -14,7 +13,7 @@ namespace butl
   inline char path_traits<char>::
   tolower (char c)
   {
-    return std::tolower (c);
+    return lcase (c);
   }
 
   template <>
@@ -28,7 +27,7 @@ namespace butl
   inline char path_traits<char>::
   toupper (char c)
   {
-    return std::toupper (c);
+    return ucase (c);
   }
 
   template <>
diff --git a/butl/process.cxx b/butl/process.cxx
index 6b41165..75e8718 100644
--- a/butl/process.cxx
+++ b/butl/process.cxx
@@ -12,7 +12,6 @@
 
 #  include <io.h>        // _open_osfhandle(), _get_osfhandle(), _close()
 #  include <fcntl.h>     // _O_TEXT
-#  include <string.h>    // _stricmp() @@ CASE
 #  include <stdlib.h>    // _MAX_PATH, getenv()
 #  include <sys/types.h> // stat
 #  include <sys/stat.h>  // stat(), S_IS*
@@ -25,6 +24,7 @@
 
 #include <cassert>
 
+#include <butl/utility>  // casecmp()
 #include <butl/fdstream> // fdnull(), fdclose()
 
 using namespace std;
@@ -288,7 +288,7 @@ namespace butl
       // support those, it will have to be handled differently.
       //
       const char* e (r.extension ());
-      if (e == nullptr || _stricmp (e, "exe") != 0) // @@ CASE
+      if (e == nullptr || casecmp (e, "exe") != 0)
         r += ".exe";
 
       // Only check that the file exists since the executable mode is set
diff --git a/butl/sha256.cxx b/butl/sha256.cxx
index 7c0b21e..fb1fcc8 100644
--- a/butl/sha256.cxx
+++ b/butl/sha256.cxx
@@ -9,9 +9,11 @@
 #include <stdint.h>
 #include <stddef.h> // size_t
 
-#include <cctype>    // isxdigit(), toupper(), tolower()
+#include <cctype>    // isxdigit()
 #include <stdexcept> // invalid_argument
 
+#include <butl/utility> // ucase(), lcase()
+
 using SHA256_CTX = butl::sha256::context;
 
 extern "C"
@@ -101,7 +103,7 @@ namespace butl
       if (i > 0 && i % 2 == 0)
         f += ":";
 
-      f += toupper (c);
+      f += ucase (c);
     }
 
     return f;
@@ -131,7 +133,7 @@ namespace butl
         if (!isxdigit (c))
           bad ();
 
-        s += tolower (c);
+        s += lcase (c);
       }
     }
 
diff --git a/butl/utility b/butl/utility
index 101d503..757edc2 100644
--- a/butl/utility
+++ b/butl/utility
@@ -5,14 +5,73 @@
 #ifndef BUTL_UTILITY
 #define BUTL_UTILITY
 
-#include <cstddef> // std::size_t
+#include <string>
+#include <cstddef> // size_t
 #include <utility> // forward()
-#include <cstring> // strcmp
+#include <cstring> // strcmp(), strlen()
 
 #include <butl/export>
 
 namespace butl
 {
+  // Convert ASCII character/string case. If there is no upper/lower case
+  // counterpart, leave the character unchanged. The POSIX locale (also known
+  // as C locale) must be the current application locale. Otherwise the
+  // behavior is undefined.
+  //
+  // Note that the POSIX locale specifies behaviour on data consisting
+  // entirely of characters from the portable character set (subset of ASCII
+  // including 103 non-negative characters and English alphabet letters in
+  // particular) and the control character set (more about them at
+  // http://pubs.opengroup.org/onlinepubs/009696899/basedefs/xbd_chap06.html).
+  //
+  // Also note that according to the POSIX locale definition the case
+  // conversion can be applied only to [A-Z] and [a-z] character ranges being
+  // translated to each other (more about that at
+  // http://pubs.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap07.html#tag_07_02)
+  //
+  char ucase (char);
+  std::string ucase (const char*, std::size_t = std::string::npos);
+  std::string ucase (const std::string&);
+  std::string& ucase (std::string&);
+  LIBBUTL_EXPORT void ucase (char*, std::size_t);
+
+  char lcase (char);
+  std::string lcase (const char*, std::size_t = std::string::npos);
+  std::string lcase (const std::string&);
+  std::string& lcase (std::string&);
+  LIBBUTL_EXPORT void lcase (char*, std::size_t);
+
+  // Compare ASCII characters/strings ignoring case. Behave as if characters
+  // had been converted to the lower case and then byte-compared. Return a
+  // negative, zero or positive value if the left hand side is less, equal or
+  // greater than the right hand side, respectivelly. The POSIX locale (also
+  // known as C locale) must be the current application locale. Otherwise the
+  // behavior is undefined.
+  //
+  // The optional size argument specifies the maximum number of characters
+  // to compare.
+  //
+  int casecmp (char, char);
+
+  int casecmp (const std::string&, const std::string&,
+               std::size_t = std::string::npos);
+
+  int casecmp (const std::string&, const char*,
+               std::size_t = std::string::npos);
+
+  LIBBUTL_EXPORT int casecmp (const char*, const char*,
+                              std::size_t = std::string::npos);
+
+  bool
+  alpha (char);
+
+  bool
+  digit (char);
+
+  bool
+  alnum (char);
+
   // Key comparators (i.e., to be used in sets, maps, etc).
   //
   struct compare_c_string
@@ -67,4 +126,6 @@ namespace butl
   reverse_iterate (T&& x) {return reverse_range<T> (std::forward<T> (x));}
 }
 
+#include <butl/utility.ixx>
+
 #endif // BUTL_UTILITY
diff --git a/butl/utility.ixx b/butl/utility.ixx
new file mode 100644
index 0000000..b0d83c9
--- /dev/null
+++ b/butl/utility.ixx
@@ -0,0 +1,136 @@
+// file      : butl/utility.ixx -*- C++ -*-
+// copyright : Copyright (c) 2014-2016 Code Synthesis Ltd
+// license   : MIT; see accompanying LICENSE file
+
+#ifndef _WIN32
+#  include <strings.h> // strcasecmp(), strncasecmp()
+#else
+#  include <string.h> // _stricmp(), _strnicmp()
+#endif
+
+#include <cctype> // toupper(), tolower(), isalpha(), isdigit(), isalnum()
+
+namespace butl
+{
+  inline char
+  ucase (char c)
+  {
+    return std::toupper (c);
+  }
+
+  inline std::string
+  ucase (const char* s, std::size_t n)
+  {
+    std::string r (s, n == std::string::npos ? std::strlen (s) : n);
+    return ucase (r);
+  }
+
+  inline std::string
+  ucase (const std::string& s)
+  {
+    return ucase (s.c_str (), s.size ());
+  }
+
+  inline std::string&
+  ucase (std::string& s)
+  {
+    if (size_t n = s.size ())
+    {
+      s.front () = s.front (); // Force copy in CoW.
+      ucase (const_cast<char*> (s.data ()), n);
+    }
+    return s;
+  }
+
+  inline void
+  ucase (char* s, std::size_t n)
+  {
+    for (const char* e (s + n); s != e; ++s)
+      *s = ucase (*s);
+  }
+
+  inline char
+  lcase (char c)
+  {
+    return std::tolower (c);
+  }
+
+  inline std::string
+  lcase (const char* s, std::size_t n)
+  {
+    std::string r (s, n == std::string::npos ? std::strlen (s) : n);
+    return lcase (r);
+  }
+
+  inline std::string
+  lcase (const std::string& s)
+  {
+    return lcase (s.c_str (), s.size ());
+  }
+
+  inline std::string&
+  lcase (std::string& s)
+  {
+    if (size_t n = s.size ())
+    {
+      s.front () = s.front (); // Force copy in CoW.
+      lcase (const_cast<char*> (s.data ()), n);
+    }
+    return s;
+  }
+
+  inline void
+  lcase (char* s, std::size_t n)
+  {
+    for (const char* e (s + n); s != e; ++s)
+      *s = lcase (*s);
+  }
+
+  inline int
+  casecmp (char l, char r)
+  {
+    l = lcase (l);
+    r = lcase (r);
+    return l < r ? -1 : (l > r ? 1 : 0);
+  }
+
+  inline int
+  casecmp (const std::string& l, const std::string& r, std::size_t n)
+  {
+    return casecmp (l.c_str (), r.c_str (), n);
+  }
+
+  inline int
+  casecmp (const std::string& l, const char* r, std::size_t n)
+  {
+    return casecmp (l.c_str (), r, n);
+  }
+
+  inline int
+  casecmp (const char* l, const char* r, std::size_t n)
+  {
+#ifndef _WIN32
+    return n == std::string::npos ? strcasecmp (l, r) : strncasecmp (l, r, n);
+#else
+    return n == std::string::npos ? _stricmp (l, r) : _strnicmp (l, r, n);
+#endif
+  }
+
+  inline bool
+  alpha (char c)
+  {
+    return std::isalpha (c);
+  }
+
+  inline bool
+  digit (char c)
+  {
+    return std::isdigit (c);
+  }
+
+  inline bool
+  alnum (char c)
+  {
+    return std::isalnum (c);
+  }
+}
diff --git a/tests/buildfile b/tests/buildfile
index 36e2ce6..fd2589a 100644
--- a/tests/buildfile
+++ b/tests/buildfile
@@ -3,7 +3,7 @@
 # license   : MIT; see accompanying LICENSE file
 
 d = base64/ cpfile/ dir-iterator/ fdstream/ link/ pager/ path/ prefix-map/ \
-    process/ sha256/ timestamp/ triplet/
+    process/ sha256/ strcase/ timestamp/ triplet/
 
 ./: $d
 include $d
diff --git a/tests/strcase/buildfile b/tests/strcase/buildfile
new file mode 100644
index 0000000..a18c730
--- /dev/null
+++ b/tests/strcase/buildfile
@@ -0,0 +1,7 @@
+# file      : tests/strcase/buildfile
+# copyright : Copyright (c) 2014-2016 Code Synthesis Ltd
+# license   : MIT; see accompanying LICENSE file
+
+exe{driver}: cxx{driver} ../../butl/lib{butl}
+
+include ../../butl/
diff --git a/tests/strcase/driver.cxx b/tests/strcase/driver.cxx
new file mode 100644
index 0000000..4741856
--- /dev/null
+++ b/tests/strcase/driver.cxx
@@ -0,0 +1,68 @@
+// file      : tests/strcase/driver.cxx -*- C++ -*-
+// copyright : Copyright (c) 2014-2016 Code Synthesis Ltd
+// license   : MIT; see accompanying LICENSE file
+
+#include <string>
+#include <cassert>
+
+#include <butl/utility>
+
+using namespace std;
+using namespace butl;
+
+int
+main ()
+{
+  const string upper ("+/0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ");
+  const string lower ("+/0123456789abcdefghijklmnopqrstuvwxyz");
+
+  assert (casecmp (upper, lower) == 0);
+  assert (casecmp (upper, lower, upper.size ()) == 0);
+  assert (casecmp (upper, lower, 100) == 0);
+  assert (casecmp ("a", "A1") < 0);
+  assert (casecmp ("A1", "a") > 0);
+  assert (casecmp ("a", "A1", 1) == 0);
+  assert (casecmp ("A1", "a", 1) == 0);
+  assert (casecmp ("a", "b", 0) == 0);
+
+  for (size_t i (0); i < upper.size (); ++i)
+  {
+    assert (casecmp (upper[i], lower[i]) == 0);
+
+    if (i > 0)
+    {
+      assert (casecmp (upper[i], lower[i - 1]) > 0);
+      assert (casecmp (lower[i - 1], upper[i]) < 0);
+    }
+  }
+
+  // As casecmp() compares strings as if they have been converted to the
+  // lower case the characters [\]^_` (located between 'Z' and 'a' in the ASCII
+  // table) evaluates as less than any alphabetic character.
+  //
+  string ascii_91_96 ("[\\]^_`");
+  for (const auto& c: ascii_91_96)
+  {
+    assert (casecmp (&c, "A", 1) < 0);
+    assert (casecmp (&c, "a", 1) < 0);
+  }
+
+  assert (ucase (lower) == upper);
+  assert (lcase (upper) == lower);
+
+  assert (ucase (lower.c_str (), 20) == string (upper, 0, 20));
+  assert (lcase (upper.c_str (), 20) == string (lower, 0, 20));
+
+  assert (ucase (lower.c_str (), 0) == string ());
+  assert (lcase (upper.c_str (), 0) == string ());
+
+  assert (ucase ("") == string ());
+  assert (lcase ("") == string ());
+
+  string s (upper);
+  assert (lcase (s) == lower);
+
+  s = lower;
+  ucase (const_cast<char*> (s.data ()), s.size ());
+  assert (s == upper);
+}