[PATCH] deb822: Don't interpret comment lines

[John Wright]

Lines beginning with '#' are ignored, and blocks of comments don't cause
empty paragraphs to be generated.

Closes: #632306
---
 lib/debian/deb822.py |   30 +++++++++++++++++++---
 tests/test_deb822.py |   66 +++++++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 89 insertions(+), 7 deletions(-)

diff --git a/lib/debian/deb822.py b/lib/debian/deb822.py
index a0cad69..afe89d9 100644
--- a/lib/debian/deb822.py
+++ b/lib/debian/deb822.py
@@ -53,7 +53,8 @@ class TagSectionWrapper(object, UserDict.DictMixin):
         self.__section = section
 
     def keys(self):
-        return self.__section.keys()
+        return [key for key in self.__section.keys()
+                if not key.startswith('#')]
 
     def __getitem__(self, key):
         s = self.__section.find_raw(key)
@@ -302,8 +303,11 @@ class Deb822(Deb822Dict):
         if _have_apt_pkg and use_apt_pkg and isinstance(sequence, file):
             parser = apt_pkg.TagFile(sequence)
             for section in parser:
-                yield cls(fields=fields, _parsed=TagSectionWrapper(section),
-                          encoding=encoding)
+                paragraph = cls(fields=fields,
+                                _parsed=TagSectionWrapper(section),
+                                encoding=encoding)
+                if paragraph:
+                    yield paragraph
 
         else:
             iterable = iter(sequence)
@@ -316,6 +320,22 @@ class Deb822(Deb822Dict):
 
     ###
 
+    @staticmethod
+    def _skip_useless_lines(sequence):
+        """Yields only lines that do not begin with '#'.
+
+        Also skips any blank lines at the beginning of the input.
+        """
+        at_beginning = True
+        for line in sequence:
+            if line.startswith('#'):
+                continue
+            if at_beginning:
+                if not line.rstrip('\r\n'):
+                    continue
+                at_beginning = False
+            yield line
+
     def _internal_parser(self, sequence, fields=None):
         single = re.compile("^(?P<key>\S+)\s*:\s*(?P<data>\S.*?)\s*$")
         multi = re.compile("^(?P<key>\S+)\s*:\s*$")
@@ -328,7 +348,9 @@ class Deb822(Deb822Dict):
 
         curkey = None
         content = ""
-        for line in self.gpg_stripped_paragraph(sequence):
+
+        for line in self.gpg_stripped_paragraph(
+                self._skip_useless_lines(sequence)):
             m = single.match(line)
             if m:
                 if curkey:
diff --git a/tests/test_deb822.py b/tests/test_deb822.py
index 891f4cd..b1d0595 100755
--- a/tests/test_deb822.py
+++ b/tests/test_deb822.py
@@ -236,6 +236,47 @@ iD8DBQFIGWQO0UIZh3p4ZWERAug/AJ93DWD9o+1VMgPDjWn/dsmPSgTWGQCeOfZi
 -----END PGP SIGNATURE-----
 '''
 
+UNPARSED_PARAGRAPHS_WITH_COMMENTS = '''\
+# Leading comments should be ignored.
+
+Source: foo
+Section: bar
+# An inline comment in the middle of a paragraph should be ignored.
+Priority: optional
+Homepage: http://www.debian.org/
+
+# Comments in the middle shouldn't result in extra blank paragraphs either.
+
+# Ditto.
+
+# A comment at the top of a paragraph should be ignored.
+Package: foo
+Architecture: any
+Description: An awesome package
+  # This should still appear in the result.
+  Blah, blah, blah. # So should this.
+# A comment at the end of a paragraph should be ignored.
+
+# Trailing comments shouldn't cause extra blank paragraphs.
+'''
+
+PARSED_PARAGRAPHS_WITH_COMMENTS = [
+    deb822.Deb822Dict([
+        ('Source', 'foo'),
+        ('Section', 'bar'),
+        ('Priority', 'optional'),
+        ('Homepage', 'http://www.debian.org/'),
+    ]),
+    deb822.Deb822Dict([
+        ('Package', 'foo'),
+        ('Architecture', 'any'),
+        ('Description', 'An awesome package\n'
+            '  # This should still appear in the result.\n'
+            '  Blah, blah, blah. # So should this.'),
+    ]),
+]
+
+
 class TestDeb822Dict(unittest.TestCase):
     def make_dict(self):
         d = deb822.Deb822Dict()
@@ -396,10 +437,10 @@ class TestDeb822(unittest.TestCase):
                 self.assertWellParsed(d, PARSED_PACKAGE)
             self.assertEqual(count, 2)
 
-    def _test_iter_paragraphs(self, file, cls, **kwargs):
+    def _test_iter_paragraphs(self, filename, cls, **kwargs):
         """Ensure iter_paragraphs consistency"""
         
-        f = open(file)
+        f = open(filename)
         packages_content = f.read()
         f.close()
         # XXX: The way multivalued fields parsing works, we can't guarantee
@@ -409,10 +450,12 @@ class TestDeb822(unittest.TestCase):
 
         s = StringIO()
         l = []
-        for p in cls.iter_paragraphs(open(file), **kwargs):
+        f = open(filename)
+        for p in cls.iter_paragraphs(f, **kwargs):
             p.dump(s)
             s.write("\n")
             l.append(p)
+        f.close()
         self.assertEqual(s.getvalue(), packages_content)
         if kwargs["shared_storage"] is False:
             # If shared_storage is False, data should be consistent across
@@ -726,6 +769,23 @@ Description: python modules to work with Debian-related data formats
             self.assertEqual(p2['uploaders'],
                              u'Frank Küster <frank at debian.org>')
 
+    def _test_iter_paragraphs_comments(self, paragraphs):
+        self.assertEqual(len(paragraphs), len(PARSED_PARAGRAPHS_WITH_COMMENTS))
+        for i in range(len(paragraphs)):
+            self.assertWellParsed(paragraphs[i],
+                                  PARSED_PARAGRAPHS_WITH_COMMENTS[i])
+
+    def test_iter_paragraphs_comments_use_apt_pkg(self):
+        paragraphs = list(deb822.Deb822.iter_paragraphs(
+            UNPARSED_PARAGRAPHS_WITH_COMMENTS.splitlines(), use_apt_pkg=True))
+        self._test_iter_paragraphs_comments(paragraphs)
+
+    def test_iter_paragraphs_comments_native(self):
+        paragraphs = list(deb822.Deb822.iter_paragraphs(
+            UNPARSED_PARAGRAPHS_WITH_COMMENTS.splitlines(), use_apt_pkg=False))
+        self._test_iter_paragraphs_comments(paragraphs)
+
+
 class TestPkgRelations(unittest.TestCase):
 
     def test_packages(self):
-- 
1.7.4.4