Merge pull request #59 from contour-terminal/fix/mktables-python-raw-str

mktables: Make use of Python raw strings for regex strings

Author: christianparpart

src/unicode/mktables.py

@@ -313,10 +313,10 @@ def __init__(self, _ucd_dir, _header_file, _impl_file):
         self.header = open(_header_file, 'w', encoding='utf-8', newline='\u000A')
         self.impl = open(_impl_file, 'w', encoding='utf-8', newline='\u000A')
 
-        self.singleValueRE = re.compile('([0-9A-F]+)\s*;\s*(\w+)\s*#\s*(.*)$')
-        self.rangeValueRE = re.compile('([0-9A-F]+)\.\.([0-9A-F]+)\s*;\s*(\w+)\s*#\s*(.*)$')
-        self.singleValueMultiRE = re.compile('([0-9A-F]+)\s*;\s*([\w\s]+)#\s*(.*)$')
-        self.rangeValueMultiRE = re.compile('([0-9A-F]+)\.\.([0-9A-F]+)\s*;\s*([\w\s]+)#\s*(.*)$')
+        self.singleValueRE = re.compile(r'([0-9A-F]+)\s*;\s*(\w+)\s*#\s*(.*)$')
+        self.rangeValueRE = re.compile(r'([0-9A-F]+)\.\.([0-9A-F]+)\s*;\s*(\w+)\s*#\s*(.*)$')
+        self.singleValueMultiRE = re.compile(r'([0-9A-F]+)\s*;\s*([\w\s]+)#\s*(.*)$')
+        self.rangeValueMultiRE = re.compile(r'([0-9A-F]+)\.\.([0-9A-F]+)\s*;\s*([\w\s]+)#\s*(.*)$')
 
         self.general_category_map = dict()
         self.general_category = list()
@@ -393,8 +393,8 @@ def load_property_value_aliases(self): # {{{
 
         with uopen(self.ucd_dir + '/' + PropertyValueAliases_fname) as f:
             # gc ; C   ; Other    # Cc | Cf | Cn | Co | Cs
-            headerRE = re.compile('^#\s*(\w+) \((\w+)\)$')
-            lineRE = re.compile('^(\w+)\s*;\s*([a-zA-Z0-9_\.]+)\s*;\s*([a-zA-Z0-9_]+).*$')
+            headerRE = re.compile(r'^#\s*(\w+) \((\w+)\)$')
+            lineRE = re.compile(r'^(\w+)\s*;\s*([a-zA-Z0-9_\.]+)\s*;\s*([a-zA-Z0-9_]+).*$')
             property_values = dict()
 
             while True:
@@ -514,7 +514,7 @@ def write_properties(self): # {{{
 
     def load_general_category(self): # {{{
         with uopen(self.ucd_dir + DerivedGeneralCategory_fname) as f:
-            headerRE = re.compile('^#\s*General_Category=(\w+)$')
+            headerRE = re.compile(r'^#\s*General_Category=(\w+)$')
             property_values = self.property_values['General_Category']
             cat_name = ''
             cats_grouped = dict()
@@ -646,7 +646,7 @@ def load_generic_properties(self, filename): # {{{
 
     def load_properties(self, filename: str, prop_key: str): # {{{
         with uopen(filename) as f:
-            headerRE = re.compile('^#\s*{}:\s*(\w+)$'.format(prop_key))
+            headerRE = re.compile(r'^#\s*{}:\s*(\w+)$'.format(prop_key))
 
             # collect
             props_name = ''
@@ -767,7 +767,7 @@ def write_scripts(self): # {{{
     def load_blocks(self): # {{{
         filename = self.ucd_dir + '/' + Blocks_fname
         with uopen(filename) as f:
-            line_regex = re.compile('^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+);\s*(.*)$')
+            line_regex = re.compile(r'^([0-9A-Fa-f]+)\.\.([0-9A-Fa-f]+);\s*(.*)$')
             blocks = list()
             while True:
                 line = f.readline()

src/unicode/scan.cpp

@@ -39,13 +39,13 @@ namespace unicode
 namespace
 {
 #if defined(__SSE2__)
-    int countTrailingZeroBits(unsigned int value) noexcept
+    [[maybe_unused]] int countTrailingZeroBits(unsigned int value) noexcept
     {
-#if defined(_WIN32)
+    #if defined(_WIN32)
         return _tzcnt_u32(value);
-#else
+    #else
         return __builtin_ctz(value);
-#endif
+    #endif
     }
 #endif
 
@@ -78,7 +78,7 @@ size_t detail::scan_for_text_ascii(string_view text, size_t maxColumnCount) noex
     auto input = text.data();
     auto const end = text.data() + min(text.size(), maxColumnCount);
 
-#if defined(__SSE2__) // TODO: support __aarch64__
+#if defined(__SSE2__)                                   // TODO: support __aarch64__
     __m128i const ControlCodeMax = _mm_set1_epi8(0x20); // 0..0x1F
     __m128i const Complex = _mm_set1_epi8(static_cast<char>(0x80));