Merge lp:~vila/bzr-builddeb/904704-safe-decode-changelogs into lp:bzr-builddeb

[Vincent Ladeuil]

While locally testing mass-import I came across import failures during the
parsing of changelog (part of python-debian).

It turns out I'm locally using oneiric with python-debian-0.1.20ubuntu2
while jubany is still using 0.1.14ubuntu2.

Based on a discussion on IRC, this fixes the issue in a way that should be
compatible with all versions.

I've tested locally against python-debian-0.1.20ubuntu2.

[James Westby (james-w)]

review: Approve

=== modified file 'import_dsc.py'
--- import_dsc.py	2011-12-02 11:53:30 +0000
+++ import_dsc.py	2011-12-15 12:54:25 +0000
@@ -997,12 +997,8 @@
         changelog_path = os.path.join(self.tree.basedir, 'debian',
                 'changelog')
         if os.path.exists(changelog_path):
-            f = open(changelog_path)
-            try:
-                changelog_contents = f.read()
-            finally:
-                f.close()
-            changelog = Changelog(file=changelog_contents, max_blocks=1)
+            changelog = self.get_changelog_from_source(
+                self.tree.basedir, max_blocks=1)
         message, authors, thanks, bugs = \
                 get_commit_info_from_changelog(changelog, self.branch)
         if message is None:
@@ -1077,10 +1073,17 @@
         # FIXME: What about other versions ?
         return { None: parents }
 
-    def get_changelog_from_source(self, dir):
+    def get_changelog_from_source(self, dir, max_blocks=None):
         cl_filename = os.path.join(dir, "debian", "changelog")
+        content = open(cl_filename).read()
+        # Older versions of python-debian were accepting various encodings in
+        # the changelog. This is not true with 0.1.20ubuntu2 at least which
+        # force an 'utf-8' encoding. This leads to failures when trying to
+        # parse old changelogs. Using safe_decode() below will fallback to
+        # iso-8859-1 (latin_1) in this case.
+        content = safe_decode(content)
         cl = Changelog()
-        cl.parse_changelog(open(cl_filename).read(), strict=False)
+        cl.parse_changelog(content, strict=False, max_blocks=max_blocks)
         return cl
 
     def _fetch_from_branch(self, branch, revid):

=== modified file 'tests/test_import_dsc.py'
--- tests/test_import_dsc.py	2011-12-02 11:53:30 +0000
+++ tests/test_import_dsc.py	2011-12-15 12:54:25 +0000
@@ -1889,6 +1889,18 @@
         builder.build()
         self.db1.import_package(builder.dsc_name())
 
+    def test_parsechangelog_with_latin1(self):
+        self.build_tree_contents(
+            [('unstable/debian/',),
+             ('unstable/debian/changelog', u'''\
+package (1.0-1) UNRELEASED; urgency=low
+
+  * Lots of work.
+
+ -- Jo\xe9 Master <[e-mail]>  Thu,  2 Dec 2004 16:59:43 +0100
+'''.encode('latin_1'))])
+        cl = self.db1.get_changelog_from_source(self.db1.tree.basedir)
+
 
 class OneZeroSourceExtractorTests(tests.TestCaseInTempDir):