# HG changeset patch
# User Paul Boddie <paul@boddie.org.uk>
# Date 1493935827 -7200
# Node ID 8dbedbb8ef8b60cb736e216f88acc364409380be
# Parent  3993165616f88b901f0897fe41fd0d7115d1454f
Moved patterns into the parser, introducing functionality for pattern re-use.

diff -r 3993165616f8 -r 8dbedbb8ef8b moinformat/__init__.py
--- a/moinformat/__init__.py	Thu May 04 22:39:00 2017 +0200
+++ b/moinformat/__init__.py	Fri May 05 00:10:27 2017 +0200
@@ -19,15 +19,13 @@
 this program.  If not, see <http://www.gnu.org/licenses/>.
 """
 
-from moinformat.parsing import ParserBase, TokenStream, new_block
+from moinformat.parsing import ParserBase, TokenStream, get_patterns, new_block
 from moinformat.serialisers import serialise
 from moinformat.tree import Break, DefItem, DefTerm, FontStyle, Heading, \
                             Larger, ListItem, Monospace, Region, Rule, Smaller, \
                             Subscript, Superscript, TableAttr, TableAttrs, \
                             TableCell, TableRow, Text, Underline
 
-import re
-
 # Regular expressions.
 
 syntax = {
@@ -103,23 +101,6 @@
     "attrvalue"     : r"""=(?P<x>['"])(.*?)(?P=x)""",
     }
 
-# Define pattern details.
-
-table_pattern_names = ["attrname", "colour", "colspan", "halign", "rowspan", "tableattrsend", "valign", "width"]
-
-inline_pattern_names = ["fontstyle", "larger", "monospace", "smaller", "sub", "super", "underline"]
-
-def inline_patterns_for(name):
-    names = inline_pattern_names[:]
-    names[names.index(name)] = "%send" % name
-    return names
-
-# Define patterns for the regular expressions.
-
-patterns = {}
-for name, value in syntax.items():
-    patterns[name] = re.compile(value, re.UNICODE | re.MULTILINE)
-
 
 
 class Parser(ParserBase):
@@ -139,13 +120,37 @@
 
         ParserBase.__init__(self, formats)
 
+    # Pattern details.
+
+    patterns = get_patterns(syntax)
+
+    table_pattern_names = [
+        "attrname", "colour", "colspan", "halign", "rowspan", "tableattrsend",
+        "valign", "width"
+        ]
+
+    inline_pattern_names = [
+        "fontstyle", "larger", "monospace", "smaller", "sub", "super", "underline",
+        ]
+
+    region_pattern_names = inline_pattern_names + [
+        "break", "heading", "defterm", "defterm_empty", "listitem",
+        "listitem_alpha", "listitem_dot", "listitem_num", "listitem_roman",
+        "regionstart", "regionend", "rule", "tablerow",
+        ]
+
+    def inline_patterns_for(self, name):
+        names = self.inline_pattern_names[:]
+        names[names.index(name)] = "%send" % name
+        return names
+
+    # Principal parser methods.
+
     def get_items(self, s):
 
         "Return a sequence of token items for 's'."
 
-        return TokenStream(s, patterns)
-
-    # Principal parser methods.
+        return TokenStream(s, self.patterns)
 
     def parse(self, s):
 
@@ -175,16 +180,7 @@
         "Parse the data provided by 'items' to populate a wiki 'region'."
 
         new_block(region)
-
-        self.parse_region_details(items, region, inline_pattern_names + [
-            "break", "heading",
-            "defterm", "defterm_empty",
-            "listitem", "listitem_alpha", "listitem_dot", "listitem_num",
-            "listitem_roman",
-            "regionstart", "regionend",
-            "rule",
-            "tablerow",
-            ])
+        self.parse_region_details(items, region, self.region_pattern_names)
 
     # Parser methods supporting different page features.
 
@@ -274,7 +270,7 @@
 
         span = FontStyle([], emphasis, strong)
         if not double:
-            self.parse_region_details(items, span, inline_pattern_names)
+            self.parse_region_details(items, span, self.inline_pattern_names)
         region.append_inline(span)
 
     def parse_halign(self, items, attrs):
@@ -293,7 +289,7 @@
         level = len(items.read_match(2))
         start_pad = items.read_match(3)
         heading = Heading([], level, start_extra, start_pad)
-        self.parse_region_details(items, heading, ["headingend"] + inline_pattern_names)
+        self.parse_region_details(items, heading, ["headingend"] + self.inline_pattern_names)
         region.add(heading)
         new_block(region)
 
@@ -354,7 +350,7 @@
         "Handle the start of table attributes within 'cell'."
 
         attrs = TableAttrs([])
-        self.parse_region_details(items, attrs, table_pattern_names)
+        self.parse_region_details(items, attrs, self.table_pattern_names)
 
         # Test the validity of the attributes.
 
@@ -444,7 +440,7 @@
         "Handle an inline region."
 
         span = cls([])
-        self.parse_region_details(items, span, inline_patterns_for(pattern_name))
+        self.parse_region_details(items, span, self.inline_patterns_for(pattern_name))
         region.append_inline(span)
 
     def parse_larger(self, items, region):
diff -r 3993165616f8 -r 8dbedbb8ef8b moinformat/parsing.py
--- a/moinformat/parsing.py	Thu May 04 22:39:00 2017 +0200
+++ b/moinformat/parsing.py	Fri May 05 00:10:27 2017 +0200
@@ -20,6 +20,27 @@
 """
 
 from moinformat.tree import Block, Region, Text
+import re
+
+# Pattern management.
+
+def get_patterns(syntax):
+
+    "Define patterns for the regular expressions in the 'syntax' mapping."
+
+    patterns = {}
+    for name, value in syntax.items():
+        patterns[name] = re.compile(value, re.UNICODE | re.MULTILINE)
+    return patterns
+
+def combine_patterns(patterns, syntax):
+
+    "Combine 'patterns' with those defined by the given 'syntax' mapping."
+
+    p = {}
+    p.update(patterns)
+    p.update(get_patterns(syntax))
+    return p
 
 # Tokenising functions.