tarball interface encode check encoding is in UTF8

installsystems/image.py

@@ -651,7 +651,7 @@ class PackageImage(Image):
         '''
         desc = {}
         # check format
-        img_format = self._tarball.get_str("format")
+        img_format = self._tarball.get_utf8("format")
         try:
             if float(img_format) >= math.floor(float(self.format)) + 1.0:
                 raise Exception()
@@ -660,7 +660,7 @@ class PackageImage(Image):
         desc["format"] = img_format
         # check description
         try:
-            img_desc = self._tarball.get_str("description.json")
+            img_desc = self._tarball.get_utf8("description.json")
             desc.update(json.loads(img_desc))
             self.check_image_name(desc["name"])
             self.check_image_version(desc["version"])
@@ -674,7 +674,7 @@ class PackageImage(Image):
             raise Exception(u"Invalid description: %s" % e)
         # try to load changelog
         try:
-            img_changelog = self._tarball.get_str("changelog")
+            img_changelog = self._tarball.get_utf8("changelog")
             desc["changelog"] = Changelog(img_changelog)
         except KeyError:
             desc["changelog"] = Changelog("")
@@ -747,7 +747,7 @@ class PackageImage(Image):
             warn(u"No file matching %s" % filename)
         for filename in filelist:
             arrow(filename)
-            out(self._tarball.get_str(filename))
+            out(self._tarball.get_utf8(filename))
 
     def download(self, directory, force=False, image=True, payload=False):
         '''

installsystems/tarball.py

@@ -25,6 +25,9 @@ class Tarball(tarfile.TarFile):
         ti.mtime = int(time.time())
         ti.uid = ti.gid = 0
         ti.uname = ti.gname = "root"
+        # unicode char is encoded in UTF-8, has changelog must be in UTF-8
+        if isinstance(content, unicode):
+            content = content.encode("UTF-8")
         ti.size = len(content) if content is not None else 0
         self.addfile(ti, StringIO.StringIO(content))
 
@@ -36,6 +39,15 @@ class Tarball(tarfile.TarFile):
         fd = self.extractfile(ti)
         return fd.read() if fd is not None else ""
 
+    def get_utf8(self, name):
+        '''
+        Return an unicode string from a file encoded in UTF-8 inside tarball
+        '''
+        try:
+            return unicode(self.get_str(name), "UTF-8")
+        except UnicodeDecodeError:
+            raise Exception(u"Invalid UTF-8 character in %s" % name)
+
     def getnames(self, re_pattern=None, glob_pattern=None, dir=True):
         names = super(Tarball, self).getnames()
         # regexp matching