Add unicode sequence parsing

scantist-ossops-m2 · Aug 22, 2015 · fa02bad · fa02bad
1 parent 0ae11a4
commit fa02bad
Show file tree

Hide file tree

Showing 3 changed files with 39 additions and 1 deletion.
diff --git a/src/context.cpp b/src/context.cpp
@@ -32,7 +32,6 @@
 #include "sass2scss.h"
 #include "prelexer.hpp"
 #include "emitter.hpp"
-#include "debugger.hpp"
 
 namespace Sass {
   using namespace Constants;

diff --git a/src/prelexer.cpp b/src/prelexer.cpp
@@ -78,6 +78,7 @@ namespace Sass {
     const char* identifier_alpha(const char* src)
     {
       return alternatives<
+               unicode_seq,
                alpha,
                unicode,
                exactly<'-'>,
@@ -90,6 +91,7 @@ namespace Sass {
     const char* identifier_alnum(const char* src)
     {
       return alternatives<
+               unicode_seq,
                alnum,
                unicode,
                exactly<'-'>,
@@ -153,6 +155,7 @@ namespace Sass {
               re_linebreak
             >,
             escape_seq,
+            unicode_seq,
             // skip interpolants
             interpolant,
             // skip non delimiters
@@ -176,6 +179,7 @@ namespace Sass {
               re_linebreak
             >,
             escape_seq,
+            unicode_seq,
             // skip interpolants
             interpolant,
             // skip non delimiters
@@ -832,6 +836,20 @@ namespace Sass {
       return (p == 0) ? t.end : 0;
     }
 
+    const char* unicode_seq(const char* src) {
+      return sequence <
+        alternatives <
+          exactly< 'U' >,
+          exactly< 'u' >
+        >,
+        exactly< '+' >,
+        padded_token <
+          6, xdigit,
+          exactly < '?' >
+        >
+      >(src);
+    }
+
     const char* static_component(const char* src) {
       return alternatives< identifier,
                            static_string,
@@ -926,5 +944,21 @@ namespace Sass {
       return sequence< number, optional_spaces, exactly<'/'>, optional_spaces, number >(src);
     }
 
+    template <size_t size, prelexer mx, prelexer pad>
+    const char* padded_token(const char* src)
+    {
+      size_t got = 0;
+      const char* pos = src;
+      while (got < size) {
+        if (!mx(pos)) break;
+        ++ pos; ++ got;
+      }
+      while (got < size) {
+        if (!pad(pos)) break;
+        ++ pos; ++ got;
+      }
+      return got ? pos : 0;
+    }
+
   }
 }
diff --git a/src/prelexer.hpp b/src/prelexer.hpp
@@ -227,6 +227,8 @@ namespace Sass {
     const char* kwd_charset_directive(const char* src);
     const char* kwd_extend(const char* src);
 
+    const char* unicode_seq(const char* src);
+
     const char* kwd_if_directive(const char* src);
     const char* kwd_else_directive(const char* src);
     const char* elseif_directive(const char* src);
@@ -388,6 +390,9 @@ namespace Sass {
       return counter;
     }
 
+    template <size_t size, prelexer mx, prelexer pad>
+    const char* padded_token(const char* src);
+
   }
 }