- unicode literals need to just be handled differently if they have utf-8

encoded in them vs. unicode escaping.  not worth figuring out how to combine
these right now

4 files changed, 15 insertions, 5 deletions

diff --git a/lib/sqlalchemy/dialects/sqlite/base.py b/lib/sqlalchemy/dialects/sqlite/base.py
index d742a71b5..1ca8f4e64 100644
--- a/lib/sqlalchemy/dialects/sqlite/base.py
+++ b/lib/sqlalchemy/dialects/sqlite/base.py
@@ -812,7 +812,7 @@ class SQLiteDialect(default.DefaultDialect):
             coltype = sqltypes.NullType()
 
         if default is not None:
-            default = str(default)
+            default = util.text_type(default)
 
         return {
             'name': name,
diff --git a/lib/sqlalchemy/sql/compiler.py b/lib/sqlalchemy/sql/compiler.py
index d51dd625a..c3aea159a 100644
--- a/lib/sqlalchemy/sql/compiler.py
+++ b/lib/sqlalchemy/sql/compiler.py
@@ -1868,11 +1868,11 @@ class DDLCompiler(engine.Compiled):
                     first_pk = True
             except exc.CompileError as ce:
                 util.raise_from_cause(
-                    exc.CompileError(util.u("(in table '%s', column '%s'): %s" % (
+                    exc.CompileError(util.u("(in table '%s', column '%s'): %s") % (
                                                 table.description,
                                                 column.name,
                                                 ce.args[0]
-                                            ))))
+                                            )))
 
         const = self.create_table_constraints(table)
         if const:
@@ -2344,7 +2344,7 @@ class IdentifierPreparer(object):
         lc_value = value.lower()
         return (lc_value in self.reserved_words
                 or value[0] in self.illegal_initial_characters
-                or not self.legal_characters.match(str(value))
+                or not self.legal_characters.match(util.text_type(value))
                 or (lc_value != value))
 
     def quote_schema(self, schema, force):
diff --git a/lib/sqlalchemy/util/__init__.py b/lib/sqlalchemy/util/__init__.py
index 72baa9f36..b1f1a2db6 100644
--- a/lib/sqlalchemy/util/__init__.py
+++ b/lib/sqlalchemy/util/__init__.py
@@ -8,7 +8,7 @@ from .compat import callable, cmp, reduce,  \
     threading, py3k, py2k, jython, pypy, cpython, win32, \
     pickle, dottedgetter, parse_qsl, namedtuple, next, WeakSet, reraise, \
     raise_from_cause, text_type, string_types, int_types, binary_type, \
-    quote_plus, with_metaclass, print_, itertools_filterfalse, u, b,\
+    quote_plus, with_metaclass, print_, itertools_filterfalse, u, ue, b,\
     unquote_plus
 
 from ._collections import KeyedTuple, ImmutableContainer, immutabledict, \
diff --git a/lib/sqlalchemy/util/compat.py b/lib/sqlalchemy/util/compat.py
index bc7a0fe21..94b35f019 100644
--- a/lib/sqlalchemy/util/compat.py
+++ b/lib/sqlalchemy/util/compat.py
@@ -46,6 +46,9 @@ if py3k:
     def u(s):
         return s
 
+    def ue(s):
+        return s
+
     def b(s):
         return s.encode("latin-1")
 
@@ -79,6 +82,13 @@ else:
         return (ord(byte) for byte in buf)
 
     def u(s):
+        # this differs from what six does, which doesn't support non-ASCII
+        # strings - we only use u() with
+        # literal source strings, and all our source files with non-ascii
+        # in them (all are tests) are utf-8 encoded.
+        return unicode(s, "utf-8")
+
+    def ue(s):
         return unicode(s, "unicode_escape")
 
     def b(s):