Move lexer helper functions to QUtil

3 files changed, 80 insertions, 60 deletions

diff --git a/include/qpdf/QUtil.hh b/include/qpdf/QUtil.hh
index f92cc7fc..eba4ef61 100644
--- a/include/qpdf/QUtil.hh
+++ b/include/qpdf/QUtil.hh
@@ -157,6 +157,21 @@ namespace QUtil
     // exception will be thrown.
     QPDF_DLL
     RandomDataProvider* getRandomDataProvider();
+
+    // These routines help the tokenizer recognize certain character
+    // classes without using ctype, which we avoid because of locale
+    // considerations.
+    QPDF_DLL
+    bool is_hex_digit(char);
+
+    QPDF_DLL
+    bool is_space(char);
+
+    QPDF_DLL
+    bool is_digit(char);
+
+    QPDF_DLL
+    bool is_number(char const*);
 };
 
 #endif // __QUTIL_HH__
diff --git a/libqpdf/QPDFTokenizer.cc b/libqpdf/QPDFTokenizer.cc
index 90961006..40f04f34 100644
--- a/libqpdf/QPDFTokenizer.cc
+++ b/libqpdf/QPDFTokenizer.cc
@@ -6,66 +6,11 @@
 
 #include <qpdf/QTC.hh>
 #include <qpdf/QPDFExc.hh>
+#include <qpdf/QUtil.hh>
 
 #include <stdexcept>
 #include <string.h>
 
-// See note above about ctype.
-static bool is_hex_digit(char ch)
-{
-    return (strchr("0123456789abcdefABCDEF", ch) != 0);
-}
-static bool is_space(char ch)
-{
-    return (strchr(" \f\n\r\t\v", ch) != 0);
-}
-static bool is_digit(char ch)
-{
-    return ((ch >= '0') && (ch <= '9'));
-}
-static bool
-is_number(std::string const& str)
-{
-    // ^[\+\-]?(\.\d+|\d+(\.\d+)?)$
-    char const* p = str.c_str();
-    if (! *p)
-    {
-        return false;
-    }
-    if ((*p == '-') || (*p == '+'))
-    {
-        ++p;
-    }
-    bool found_dot = false;
-    bool found_digit = false;
-    for (; *p; ++p)
-    {
-        if (*p == '.')
-        {
-            if (found_dot)
-            {
-                // only one dot
-                return false;
-            }
-            if (! *(p+1))
-            {
-                // dot can't be last
-                return false;
-            }
-            found_dot = true;
-        }
-        else if (is_digit(*p))
-        {
-            found_digit = true;
-        }
-        else
-        {
-            return false;
-        }
-    }
-    return found_digit;
-}
-
 QPDFTokenizer::QPDFTokenizer() :
     pound_special_in_name(true),
     allow_eof(false)
@@ -117,7 +62,7 @@ QPDFTokenizer::resolveLiteral()
             if ((*p == '#') && this->pound_special_in_name)
             {
                 if (p[1] && p[2] &&
-                    is_hex_digit(p[1]) && is_hex_digit(p[2]))
+                    QUtil::is_hex_digit(p[1]) && QUtil::is_hex_digit(p[2]))
                 {
                     char num[3];
                     num[0] = p[1];
@@ -153,7 +98,7 @@ QPDFTokenizer::resolveLiteral()
         }
         val = nval;
     }
-    else if (is_number(val))
+    else if (QUtil::is_number(val.c_str()))
     {
         if (val.find('.') != std::string::npos)
         {
@@ -447,7 +392,7 @@ QPDFTokenizer::presentCharacter(char ch)
 	    }
 	    val = nval;
 	}
-	else if (is_hex_digit(ch))
+	else if (QUtil::is_hex_digit(ch))
 	{
 	    val += ch;
 	}
@@ -554,7 +499,7 @@ QPDFTokenizer::readToken(PointerHolder<InputSource> input,
 	}
 	else
 	{
-	    if (is_space(static_cast<unsigned char>(ch)) &&
+	    if (QUtil::is_space(static_cast<unsigned char>(ch)) &&
 		(input->getLastOffset() == offset))
 	    {
 		++offset;
diff --git a/libqpdf/QUtil.cc b/libqpdf/QUtil.cc
index 37cfda9e..86de07f2 100644
--- a/libqpdf/QUtil.cc
+++ b/libqpdf/QUtil.cc
@@ -456,3 +456,63 @@ QUtil::srandom(unsigned int seed)
     srand(seed);
 #endif
 }
+
+bool
+QUtil::is_hex_digit(char ch)
+{
+    return (strchr("0123456789abcdefABCDEF", ch) != 0);
+}
+
+bool
+QUtil::is_space(char ch)
+{
+    return (strchr(" \f\n\r\t\v", ch) != 0);
+}
+
+bool
+QUtil::is_digit(char ch)
+{
+    return ((ch >= '0') && (ch <= '9'));
+}
+
+bool
+QUtil::is_number(char const* p)
+{
+    // ^[\+\-]?(\.\d+|\d+(\.\d+)?)$
+    if (! *p)
+    {
+        return false;
+    }
+    if ((*p == '-') || (*p == '+'))
+    {
+        ++p;
+    }
+    bool found_dot = false;
+    bool found_digit = false;
+    for (; *p; ++p)
+    {
+        if (*p == '.')
+        {
+            if (found_dot)
+            {
+                // only one dot
+                return false;
+            }
+            if (! *(p+1))
+            {
+                // dot can't be last
+                return false;
+            }
+            found_dot = true;
+        }
+        else if (QUtil::is_digit(*p))
+        {
+            found_digit = true;
+        }
+        else
+        {
+            return false;
+        }
+    }
+    return found_digit;
+}