Merge pull request #44 from alpo/unicode-poms

encode/decode UTF-8 in POMs

jip/cache.py

@@ -19,6 +19,7 @@
 # OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 # SOFTWARE.
 #
+import codecs
 
 from jip.repository import MavenRepos
 from jip.util import get_virtual_home
@@ -54,7 +55,7 @@ def download_jar(self, artifact, local_path=None):
     def download_pom(self, artifact):
         path = self.get_artifact_uri(artifact, 'pom')
         if os.path.exists(path):
-            f = open(path, 'r')
+            f = codecs.open(path, mode='r', encoding='utf-8')
             data = f.read()
             f.close()
             return data
@@ -63,7 +64,7 @@ def download_pom(self, artifact):
 
     def put_pom(self, artifact, data):
         path = self.get_artifact_uri(artifact, 'pom')
-        f = open(path, 'w')
+        f = codecs.open(path, mode='w', encoding='utf-8')
         f.write(data)
         f.close()
 

jip/maven.py

@@ -110,7 +110,7 @@ def get_element_tree(self):
             ## we use this dirty method to remove namesapce attribute so that elementtree will use default empty namespace
             pom_string = re.sub(r"<project(.|\s)*?>", '<project>', self.pom_string, 1)
             parser = ElementTree.XMLParser(target=WhitespaceNormalizer())
-            parser.feed(pom_string)
+            parser.feed(pom_string.encode('utf-8'))
             self.eletree = parser.close()
         return self.eletree
 

test/install_test.py

@@ -18,6 +18,11 @@ def testResolve(self):
         artifacts = _resolve_artifacts([junit])
         self.assertEqual(len(artifacts), 1)
 
+    def testResolveArtifactWithUmlautsInPom(self):
+        artifact = Artifact('de.l3s.boilerpipe', 'boilerpipe', '1.1.0')
+        artifacts = _resolve_artifacts([artifact])
+        self.assertEqual(len(artifacts), 1)
+
     def testExclusion(self):
         pig = Artifact('org.apache.pig', 'pig', '0.8.3')
         exclusion = [Artifact(*x.split(":")) for x in ['ant:ant', 'junit:junit','org.eclipse.jdt:core']]