Returned old folding algorithm, as the current implementation fails in some cases (https://github.com/collective/icalendar/issues/72)

2012-10-25 19:16:33 +04:00 · 2012-10-25 19:16:33 +04:00 · 7f4120da4c
commit 7f4120da4c
--- a/src/icalendar/parser.py
+++ b/src/icalendar/parser.py
@ -16,7 +16,7 @@ from icalendar.caselessdict import CaselessDict
 SequenceTypes = [TupleType, ListType]


-def foldline(text, lenght=75, newline='\r\n'):
+def foldline(text, length=75, newline='\r\n'):
    """Make a string folded per RFC5545 (each line must be less than 75 octets)

    >>> from icalendar.parser import foldline
@ -29,20 +29,52 @@ def foldline(text, lenght=75, newline='\r\n'):
    ... # doctest: +NORMALIZE_WHITESPACE
    'Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vestibulum \\r\\n
    convallis imperdiet dui posuere.'
-
-#    >>> uuu = u'alfdkadäääüüaskd'
-#    >>> foldline(uuu, length=3)
-#    u'alf\ndka\nd\xe4\xe4\n\xe4\xfc\xfc\nask\nd'
-
    """
-    return newline.join(
-            icalendar.tools.wrap(text, lenght,
-                subsequent_indent=' ',
-                drop_whitespace=False,
-                break_long_words=True,
-                replace_whitespace=False
-                )
-            )
+    assert isinstance(text, str)
+#    text.decode('utf-8')  # try to decode, to be sure it's utf-8 or ASCII
+    l_line = len(text)
+    new_lines = []
+    start = 0
+    while True:
+        end = start + length - 1
+        chunk = text[start:end]
+        m = NEWLINE.search(chunk)
+        if m is not None and m.end() != l_line:
+            new_lines.append(text[start:start + m.start()])
+            start += m.end()
+            continue
+
+        if end >= l_line:
+            end = l_line
+        else:
+            # Check that we don't fold in the middle of a UTF-8 character:
+            # http://lists.osafoundation.org/pipermail/ietf-calsify/2006-August/001126.html
+            while True:
+                char_value = ord(text[end])
+                if char_value < 128 or char_value >= 192:
+                    # This is not in the middle of a UTF-8 character, so we
+                    # can fold here:
+                    break
+                else:
+                    end -= 1
+
+        # Recompute the chunk, since start or end may have changed.
+        chunk = text[start:end]
+        new_lines.append(chunk)
+        if end == l_line:
+            break  # Done
+        start = end
+    return (newline + ' ').join(new_lines).rstrip(' ')
+
+#    return newline.join(
+#            icalendar.tools.wrap(text, length,
+#                subsequent_indent=' ',
+#                drop_whitespace=False,
+#                break_long_words=True,
+#                replace_whitespace=False
+#                )
+#            )
+

 #################################################################
 # Property parameter stuff
@ -341,7 +373,7 @@ class Contentline(str):

    A value can also be unicode
    >>> from icalendar.prop import vText
-    >>> parts = ('SUMMARY', Parameters(), vText(u'INternational char æ ø å'))
+    >>> parts = ('SUMMARY', Parameters(), vText(u'INternational char æ ø å'))
    >>> Contentline.from_parts(parts)
    'SUMMARY:INternational char \\xc3\\xa6 \\xc3\\xb8 \\xc3\\xa5'

--- a/src/icalendar/tests/test_icalendar.py
+++ b/src/icalendar/tests/test_icalendar.py
@ -1,5 +1,7 @@
-#from interlude import interact
-import unittest, doctest, os
+# coding: utf-8
+import unittest
+import doctest
+import os
 from icalendar import (
    cal,
    caselessdict,
@ -9,45 +11,34 @@ from icalendar import (

 OPTIONFLAGS = doctest.NORMALIZE_WHITESPACE | doctest.ELLIPSIS

-class FuckYouTests(unittest.TestCase):
-    def XtestBla(self):
-        from icalendar import Calendar
-        c = Calendar()
-        c['description']=u'Paragraph one\n\nParagraph two'
-        output = c.to_ical()
-        cmp = ("BEGIN:VCALENDAR\r\nDESCRIPTION:Paragraph one\r\n \r\n "
-               "Paragraph two\r\nEND:VCALENDAR\r\n")
-        self.assertEqual(output, cmp)

-    def XtestTrailingNewline(self):
-        from icalendar.parser import Contentlines, Contentline
-        c = Contentlines([Contentline('BEGIN:VEVENT\\r\\n')])
-        output = c.to_ical()
-        self.assertEqual(output, 'BEGIN:VEVENT\\r\\n')
+class IcalendarTestCase (unittest.TestCase):

-    def XtestLongLine(self):
-        from icalendar.parser import Contentlines, Contentline
-        c = Contentlines([Contentline('BEGIN:VEVENT\\r\\n')])
-        c.append(Contentline(''.join(['123456789 ']*10)+'\\r\\n'))
-        output = c.to_ical()
-        cmp = ("BEGIN:VEVENT\\r\\n\\r\\n123456789 123456789 123456789 "
-               "123456789 123456789 123456789 123456789 1234\\r\\n 56789 "
-               "123456789 123456789 \\r\\n")
-        self.assertEqual(output, cmp)
-
-    def testHmm(self):
+    def test_long_lines(self):
        from icalendar.parser import Contentlines, Contentline
        c = Contentlines([Contentline('BEGIN:VEVENT\r\n')])
-        c.append(Contentline(''.join(['123456789 ']*10)+'\r\n'))
-        output = c.to_ical()
-        # XXX: sure? looks weird in conjunction with generated content above.
-        #cmp = ('BEGIN:VEVENT\r\n\r\n123456789 123456789 123456789 123456789 '
-        #       '123456789 123456789 123456789 1234\r\n 56789 123456789 '
-        #       '123456789 \r\n')
-        cmp = ('BEGIN:VEVENT\r\n\r\n123456789 123456789 123456789 123456789 '
-               '123456789 123456789 123456789 \r\n 123456789 123456789 '
-               '123456789 \r\n\r\n')
-        self.assertEqual(output, cmp)
+        c.append(Contentline(''.join('123456789 ' * 10) + '\r\n'))
+        self.assertEqual(
+            c.to_ical(),
+            'BEGIN:VEVENT\r\n\r\n123456789 123456789 123456789 123456789 '
+            '123456789 123456789 123456789 1234\r\n 56789 123456789 '
+            '123456789 \r\n\r\n'
+        )
+
+    def test_fold_line(self):
+        from icalendar.parser import foldline
+        self.assertRaises(AssertionError, foldline, u'привет', length=3)
+        self.assertEqual(foldline('foobar', length=4), 'foo\r\n bar')
+        self.assertEqual(
+            foldline('Lorem ipsum dolor sit amet, consectetur adipiscing elit.'
+                     ' Vestibulum convallis imperdiet dui posuere.'),
+            'Lorem ipsum dolor sit amet, consectetur adipiscing elit.'
+            ' Vestibulum conval\r\n lis imperdiet dui posuere.'
+        )
+        self.assertEqual(
+            foldline('DESCRIPTION:АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЬЫЪЭЮЯ'),
+            'DESCRIPTION:АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЬЫЪЭ\r\n ЮЯ'
+        )


 def load_tests(loader=None, tests=None, pattern=None):
@ -60,7 +51,11 @@ def load_tests(loader=None, tests=None, pattern=None):
    for docfile in ['example.txt', 'groupscheduled.txt',
                    'small.txt', 'multiple.txt', 'recurrence.txt']:
        filename = os.path.abspath(os.path.join(current_dir, docfile))
-        suite.addTest(doctest.DocFileSuite(docfile,
-            optionflags=OPTIONFLAGS,
-            globs={'__file__': filename}))
+        suite.addTest(
+            doctest.DocFileSuite(
+                docfile,
+                optionflags=OPTIONFLAGS,
+                globs={'__file__': filename}
+            )
+        )
    return suite