Better handling of parenthesis inside string literals

When scanning for the matching parenthesis, string literals
were not taken into account, which would inevitably lead to
errors if one wanted to do a selection on a string containing
a parenthesis.

Now it uses a state-machine to skip the ones that are located
inside literals.

CHANGELOG

@@ -2,8 +2,8 @@
 - Improve error reporting
 - Release is now signed with PGP
 - Doesn't crash on network errors
-- Fixed optimization introduced in 2.2
+- Fixed optimization introduced in 2.2 that did not hold in all cases
-  that did not hold in all cases.
+- Better handling of parenthesis inside string literals
 
 2.3
 - Very small release. The windows setup now installs the C++ library

relational/parser.py

@@ -291,7 +291,21 @@ def _find_matching_parenthesis(expression, start=0, openpar=u'(', closepar=u')')
     close parenthesis to the 1st open parenthesis found
     starting from start (0 by default)'''
     par_count = 0  # Count of parenthesis
+
+    string = False
+    escape = False
+
     for i in range(start, len(expression)):
+        if expression[i] == '\'' and not escape:
+            string = not string
+        if expression[i] == '\\' and not escape:
+            escape = True
+        else:
+            escape = False
+        if string:
+            continue
+
+
         if expression[i] == openpar:
             par_count += 1
         elif expression[i] == closepar:
@@ -299,6 +313,30 @@ def _find_matching_parenthesis(expression, start=0, openpar=u'(', closepar=u')')
             if par_count == 0:
                 return i  # Closing parenthesis of the parameter
 
+def _find_token(haystack, needle):
+    '''
+    Like the string function find, but
+    ignores tokens that are within a string
+    literal.
+    '''
+    r = -1
+    string = False
+    escape = False
+
+    for i in range(len(haystack)):
+        if haystack[i] == '\'' and not escape:
+            string = not string
+        if haystack[i] == '\\' and not escape:
+            escape = True
+        else:
+            escape = False
+        if string:
+            continue
+
+        if haystack[i:].startswith(needle):
+            return i
+    return r
+
 
 def tokenize(expression):
     '''This function converts a relational expression into a list where
@@ -331,7 +369,7 @@ def tokenize(expression):
                 par = expression.find(
                     '(', _find_matching_parenthesis(expression))
             else:  # Expression without parenthesis, so adding what's between start and parenthesis as whole
-                par = expression.find('(')
+                par = _find_token(expression, '(')
 
             items.append(expression[:par].strip())
                          # Inserting parameter of the operator

test/par1.query

@@ -0,0 +1 @@
+σ name=='(' (people)

test/par1.result

@@ -0,0 +1 @@
+id,name,chief,age

test/par2.query

@@ -0,0 +1 @@
+σ (name=='(') (people)

test/par2.result

@@ -0,0 +1 @@
+id,name,chief,age

test/par3.query

@@ -0,0 +1 @@
+σ (name==')') (people)

test/par3.result

@@ -0,0 +1 @@
+id,name,chief,age

test/par4.query

@@ -0,0 +1 @@
+σ name==')' (people)

test/par4.result

@@ -0,0 +1 @@
+id,name,chief,age