Use encoding (default is utf-8) and unicode

[sqlconvert.git] / mysql2sql / process_tokens.py

diff --git a/mysql2sql/process_tokens.py b/mysql2sql/process_tokens.py
index 37752e7c146a8d76ad63f9b7ccbbe395066eec4e..1e74ac9b69a5a416bb6cf8b0418219095b73b357 100644 (file)
--- a/mysql2sql/process_tokens.py
+++ b/mysql2sql/process_tokens.py
@@ -1,6 +1,6 @@
 
 from sqlparse import parse
-from sqlparse.sql import Statement
+from sqlparse.compat import PY3
 from sqlparse.tokens import Name, Error, Punctuation, Comment, Newline, \
     Whitespace
 
@@ -26,34 +26,49 @@ def find_error(token_list):
     return False
 
 
+if PY3:
+    xrange = range
+
+
 class StatementGrouper(object):
-    def __init__(self):
-        self.statements = []
-        self.tokens = []
+    """Collect lines and reparse until the last statement is complete"""
+
+    def __init__(self, encoding=None):
         self.lines = []
+        self.statements = []
+        self.encoding = encoding
 
     def process_line(self, line):
-        lines = self.lines
-        lines.append(line)
-        tokens = parse('\n'.join(lines))[0]
-        self.process_tokens(tokens)
-        self.lines = []
+        self.lines.append(line)
+        self.process_lines()
 
-    def process_tokens(self, tokens):
-        for token in tokens:
-            self.tokens.append(token)
-            if (token.ttype == Punctuation) and (token.value == ';'):
-                self.statements.append(Statement(self.tokens))
-                self.tokens = []
+    def process_lines(self):
+        statements = parse(''.join(self.lines), encoding=self.encoding)
+        last_stmt = statements[-1]
+        for i in xrange(len(last_stmt.tokens) - 1, 0, -1):
+            token = last_stmt.tokens[i]
+            if token.ttype in (Comment.Single, Comment.Multiline,
+                               Newline, Whitespace):
+                continue
+            if token.ttype is Punctuation and token.value == ';':
+                break  # The last statement is complete
+            # The last statement is still incomplete - wait for the next line
+            return
+        self.lines = []
+        self.statements = statements
 
     def get_statements(self):
-        for statement in self.statements:
-            yield statement
+        for stmt in self.statements:
+            yield stmt
         self.statements = []
 
     def close(self):
-        for token in self.tokens:
+        if not self.lines:
+            return
+        tokens = parse(''.join(self.lines), encoding=self.encoding)
+        for token in tokens:
             if (token.ttype not in (Comment.Single, Comment.Multiline,
                                     Newline, Whitespace)):
-                raise ValueError("Incomplete SQL statement: %s" % self.tokens)
-        return self.tokens
+                raise ValueError("Incomplete SQL statement: %s" %
+                                 tokens)
+        return tokens