summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--builder.py53
-rw-r--r--test/unittest_builder.py40
2 files changed, 89 insertions, 4 deletions
diff --git a/builder.py b/builder.py
index f744d607..1b850d8d 100644
--- a/builder.py
+++ b/builder.py
@@ -47,6 +47,45 @@ from _ast import PyCF_ONLY_AST
def parse(string):
return compile(string, "<string>", 'exec', PyCF_ONLY_AST)
+if sys.version_info >= (3, 0):
+ from tokenize import detect_encoding
+
+ def open_source_file(filename):
+ byte_stream = open(filename, 'bU')
+ encoding = detect_encoding(byte_stream.readline)[0]
+ stream = open(filename, 'U', encoding=encoding)
+ try:
+ data = stream.read()
+ except UnicodeError, uex: # wrong encodingg
+ # detect_encoding returns utf-8 if no encoding specified
+ msg = 'Wrong (%s) or no encoding specified' % encoding
+ raise ASTNGBuildingException(msg)
+ return stream, encoding, data
+
+else:
+ import re
+
+ _ENCODING_RGX = re.compile("[^#]*#*.*coding[:=]\s*([^\s]+)")
+
+ def _guess_encoding(string):
+ """get encoding from a python file as string or return None if not found
+ """
+ # check for UTF-8 byte-order mark
+ if string.startswith('\xef\xbb\xbf'):
+ return 'UTF-8'
+ for line in string.split('\n', 2)[:2]:
+ # check for encoding declaration
+ match = _ENCODING_RGX.match(line)
+ if match is not None:
+ return match.group(1)
+
+ def open_source_file(filename):
+ """get data for parsing a file"""
+ stream = open(filename, 'U')
+ data = stream.read()
+ encoding = _guess_encoding(data)
+ return stream, encoding, data
+
# ast NG builder ##############################################################
MANAGER = ASTNGManager()
@@ -101,10 +140,14 @@ class ASTNGBuilder:
path is expected to be a python source file
"""
try:
- data = open(path, 'U').read()
- except IOError, ex:
- msg = 'Unable to load file %r (%s)' % (path, ex)
+ stream, encoding, data = open_source_file(path)
+ except IOError, exc:
+ msg = 'Unable to load file %r (%s)' % (path, exc)
raise ASTNGBuildingException(msg)
+ except SyntaxError, exc: # py3k encoding specification error
+ raise ASTNGBuildingException(exc)
+ except LookupError, exc: # unknown encoding
+ raise ASTNGBuildingException(exc)
# get module name if necessary, *before modifying sys.path*
if modname is None:
try:
@@ -113,10 +156,12 @@ class ASTNGBuilder:
modname = splitext(basename(path))[0]
# build astng representation
try:
- sys.path.insert(0, dirname(path))
+ sys.path.insert(0, dirname(path)) # XXX (syt) iirk
node = self.string_build(data, modname, path)
finally:
sys.path.pop(0)
+ node.file_encoding = encoding
+ node.file_stream = stream
return node
def string_build(self, data, modname='', path=None):
diff --git a/test/unittest_builder.py b/test/unittest_builder.py
index 89fba46b..2f086266 100644
--- a/test/unittest_builder.py
+++ b/test/unittest_builder.py
@@ -43,8 +43,10 @@ from logilab.astng.nodes import Module
from logilab.astng.bases import YES, BUILTINS_NAME
from logilab.astng.as_string import as_string
from logilab.astng.manager import ASTNGManager
+
MANAGER = ASTNGManager()
+
from unittest_inference import get_name_node
import data
@@ -682,5 +684,43 @@ def func():
self.assertEqual(chain.value, 'None')
+guess_encoding = builder._guess_encoding
+
+class TestGuessEncoding(TestCase):
+
+ def testEmacs(self):
+ e = guess_encoding('# -*- coding: UTF-8 -*-')
+ self.failUnlessEqual(e, 'UTF-8')
+ e = guess_encoding('# -*- coding:UTF-8 -*-')
+ self.failUnlessEqual(e, 'UTF-8')
+ e = guess_encoding('''
+ ### -*- coding: ISO-8859-1 -*-
+ ''')
+ self.failUnlessEqual(e, 'ISO-8859-1')
+ e = guess_encoding('''
+
+ ### -*- coding: ISO-8859-1 -*-
+ ''')
+ self.failUnlessEqual(e, None)
+
+ def testVim(self):
+ e = guess_encoding('# vim:fileencoding=UTF-8')
+ self.failUnlessEqual(e, 'UTF-8')
+ e = guess_encoding('''
+ ### vim:fileencoding=ISO-8859-1
+ ''')
+ self.failUnlessEqual(e, 'ISO-8859-1')
+ e = guess_encoding('''
+
+ ### vim:fileencoding= ISO-8859-1
+ ''')
+ self.failUnlessEqual(e, None)
+
+ def testUTF8(self):
+ e = guess_encoding('\xef\xbb\xbf any UTF-8 data')
+ self.failUnlessEqual(e, 'UTF-8')
+ e = guess_encoding(' any UTF-8 data \xef\xbb\xbf')
+ self.failUnlessEqual(e, None)
+
if __name__ == '__main__':
unittest_main()