winfreed.py: hopefully more solid parsing of Content-Disposition header

[winfreed.git] / winfreed.py

diff --git a/winfreed.py b/winfreed.py
index 13b7580..16fb1c7 100755 (executable)
--- a/winfreed.py
+++ b/winfreed.py
@@ -66,25 +66,28 @@ def show(pkg):
 def download(pkg):
     # the "%s" in URLs are meant to be replaced with LANGCODE
     try:
-        url = pkg['URL'] % LANGCODE
+        pkg_url = pkg['URL'] % LANGCODE
     except:
-        url = pkg['URL']
+        pkg_url = pkg['URL']
         pass
 
-    response = urllib2.urlopen(url)
+    response = urllib2.urlopen(pkg_url)
+    url = response.geturl()
 
     filename = ""
-    if 'Content-Disposition' in  response.info():
-        # Use the filename the server tells us if any,
-        # re pattern from http://stackoverflow.com/questions/8035900
-        content_disposition = response.info().getheader('Content-Disposition').strip()
-        filename = re.findall("filename=(\S+)", content_disposition)[0]
 
-    if filename == "":
-        filename = urllib2.unquote(os.path.basename(response.geturl()))
+    # From http://paste.pound-python.org/show/9545/
+    # TODO: use a proper module to parse HTTP headers
+    if response.info().has_key('Content-Disposition') and len(response.info()['Content-Disposition'].split('filename=')) > 1:
+        # If the response has Content-Disposition, we take file name from it
+        filename = response.info()['Content-Disposition'].split('filename=')[1].decode('utf-8')
+        if filename[0] == '"' or filename[0] == "'":
+            filename = urllib2.unquote(filename.split('"')[1])
+    else:
+        filename =  urllib2.unquote(url.split('/')[-1].decode('utf_8'))
 
     if filename == "":
-        sys.stderr.write("Debug (%s): filename: %s url: %s\n" %(pkg['package_name'], filename, response.geturl()))
+        sys.stderr.write("Debug (%s): filename: %s url: %s\n" % (pkg['package_name'], filename, url))
         return
 
     destfile = os.path.join(OUTPUT_DIR, filename)