cerbero: Define a custom user agent while downloading

Some places block urllib, and others block wget / curl. Set our own
agent to workaround that, and so that people can contact us in case
we're doing something they don't like while downloading things.

3 files changed, 12 insertions, 5 deletions

diff --git a/cerbero/enums.py b/cerbero/enums.py
index 6ef5cda1..4137e027 100644
--- a/cerbero/enums.py
+++ b/cerbero/enums.py
@@ -17,6 +17,9 @@
 # Boston, MA 02111-1307, USA.
 
 
+# Safest place to define this since this file imports very few modules
+CERBERO_VERSION = '1.16.2.1'
+
 class Platform:
     ''' Enumeration of supported platforms '''
     LINUX = 'linux'
diff --git a/cerbero/utils/shell.py b/cerbero/utils/shell.py
index 20247e1b..2f38c7bc 100644
--- a/cerbero/utils/shell.py
+++ b/cerbero/utils/shell.py
@@ -35,12 +35,13 @@ from pathlib import Path, PurePath
 from distutils.version import StrictVersion
 
 import cerbero.hacks
-from cerbero.enums import Platform
+from cerbero.enums import CERBERO_VERSION, Platform
 from cerbero.utils import _, system_info, to_unixpath
 from cerbero.utils import messages as m
 from cerbero.errors import FatalError
 
 
+USER_AGENT = 'GStreamer Cerbero/' + CERBERO_VERSION
 PATCH = 'patch'
 TAR = 'tar'
 
@@ -325,7 +326,7 @@ def download_wget(url, destination=None, check_cert=True, overwrite=False):
     @param destination: destination where the file will be saved
     @type destination: str
     '''
-    cmd = "wget %s " % url
+    cmd = "wget --user-agent '{}' {} ".format(USER_AGENT, url)
     path = None
     if destination is not None:
         cmd += "-O %s " % destination
@@ -366,7 +367,9 @@ def download_urllib2(url, destination=None, check_cert=True, overwrite=False):
     try:
         logging.info(destination)
         with open(destination, 'wb') as d:
-            f = urllib.request.urlopen(url, context=ctx)
+            req = urllib.request.Request(url)
+            req.add_header('User-Agent', USER_AGENT)
+            f = urllib.request.urlopen(req, context=ctx)
             d.write(f.read())
     except urllib.error.HTTPError as e:
         if os.path.exists(destination):
@@ -383,7 +386,7 @@ def download_curl(url, destination=None, check_cert=True, overwrite=False):
     @type destination: str
     '''
     path = None
-    cmd = "curl -L --fail --retry 2 "
+    cmd = "curl -L --fail --retry 2 --user-agent '{}' ".format(USER_AGENT)
     if not check_cert:
         cmd += " -k "
     if destination is not None:
diff --git a/setup.py b/setup.py
index 1ce4f572..62bd8923 100644
--- a/setup.py
+++ b/setup.py
@@ -4,6 +4,7 @@ import shutil
 from setuptools import setup, find_packages
 from setuptools.command import sdist as setuptools_sdist
 from cerbero.utils import shell
+from cerbero.enums import CERBERO_VERSION
 from distutils.dir_util import copy_tree
 import distutils.log
 
@@ -127,7 +128,7 @@ class extended_sdist(setuptools_sdist.sdist):
 
 setup(
     name = "cerbero",
-    version = '1.16.2.1',
+    version = CERBERO_VERSION,
     author = "Andoni Morales",
     author_email = "amorales@fluendo.com",
     description = ("Multi platform build system for Open Source projects"),