distrodata: Take account proxies on distrodata tasks

Proxies defined in the enviroment where not taken into account
on the distrodata tasks. This commit implied passing the datastore
into the distro_check library and context manager for the
urllib.urlopen function.

One way to run distrodata tasks is using 'universe' as target and the
'all' distrodata task:

      $ bitbake universe -c distrodataall
      $ bitbake universe -c distro_checkall
      $ bitbake universe -c checklicenseall

Logs are located under TMPDIR/log

[YOCTO #7567]

Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

2 files changed, 69 insertions, 50 deletions

diff --git a/meta/classes/distrodata.bbclass b/meta/classes/distrodata.bbclass
index e6eb3f32b3..5a4c1b6faf 100644
--- a/meta/classes/distrodata.bbclass
+++ b/meta/classes/distrodata.bbclass
@@ -33,7 +33,7 @@ python do_distrodata_np() {
         tmpdir = d.getVar('TMPDIR', True)
         distro_check_dir = os.path.join(tmpdir, "distro_check")
         datetime = localdata.getVar('DATETIME', True)
-        dist_check.update_distro_data(distro_check_dir, datetime)
+        dist_check.update_distro_data(distro_check_dir, datetime, localdata)
 
         if pn.find("-native") != -1:
             pnstripped = pn.split("-native")
@@ -118,7 +118,7 @@ python do_distrodata() {
         tmpdir = d.getVar('TMPDIR', True)
         distro_check_dir = os.path.join(tmpdir, "distro_check")
         datetime = localdata.getVar('DATETIME', True)
-        dist_check.update_distro_data(distro_check_dir, datetime)
+        dist_check.update_distro_data(distro_check_dir, datetime, localdata)
 
         pn = d.getVar("PN", True)
         bb.note("Package Name: %s" % pn)
@@ -406,7 +406,7 @@ python do_distro_check() {
     bb.utils.mkdirhier(logpath)
     result_file = os.path.join(logpath, "distrocheck.csv")
     datetime = localdata.getVar('DATETIME', True)
-    dc.update_distro_data(distro_check_dir, datetime)
+    dc.update_distro_data(distro_check_dir, datetime, localdata)
 
     # do the comparison
     result = dc.compare_in_distro_packages_list(distro_check_dir, d)
diff --git a/meta/lib/oe/distro_check.py b/meta/lib/oe/distro_check.py
index 8ed5b0ec80..b3419ce03a 100644
--- a/meta/lib/oe/distro_check.py
+++ b/meta/lib/oe/distro_check.py
@@ -1,7 +1,28 @@
-def get_links_from_url(url):
+from contextlib import contextmanager
+@contextmanager
+def create_socket(url, d):
+    import urllib
+    socket = urllib.urlopen(url, proxies=get_proxies(d))
+    try:
+        yield socket
+    finally:
+        socket.close()
+
+def get_proxies(d):
+    import os
+    proxykeys = ['HTTP_PROXY', 'http_proxy',
+                 'HTTPS_PROXY', 'https_proxy',
+                 'FTP_PROXY', 'ftp_proxy',
+                 'FTPS_PROXY', 'ftps_proxy',
+                 'NO_PROXY', 'no_proxy',
+                 'ALL_PROXY', 'all_proxy']
+    proxyvalues = map(lambda key: d.getVar(key, True), proxykeys)
+    return dict(zip(proxykeys, proxyvalues))
+
+def get_links_from_url(url, d):
     "Return all the href links found on the web location"
 
-    import urllib, sgmllib
+    import sgmllib
     
     class LinksParser(sgmllib.SGMLParser):
         def parse(self, s):
@@ -24,19 +45,18 @@ def get_links_from_url(url):
             "Return the list of hyperlinks."
             return self.hyperlinks
 
-    sock = urllib.urlopen(url)
-    webpage = sock.read()
-    sock.close()
+    with create_socket(url,d) as sock:
+        webpage = sock.read()
 
     linksparser = LinksParser()
     linksparser.parse(webpage)
     return linksparser.get_hyperlinks()
 
-def find_latest_numeric_release(url):
+def find_latest_numeric_release(url, d):
     "Find the latest listed numeric release on the given url"
     max=0
     maxstr=""
-    for link in get_links_from_url(url):
+    for link in get_links_from_url(url, d):
         try:
             release = float(link)
         except:
@@ -70,7 +90,7 @@ def clean_package_list(package_list):
     return set.keys()
 
 
-def get_latest_released_meego_source_package_list():
+def get_latest_released_meego_source_package_list(d):
     "Returns list of all the name os packages in the latest meego distro"
 
     package_names = []
@@ -82,11 +102,11 @@ def get_latest_released_meego_source_package_list():
     package_list=clean_package_list(package_names)
     return "1.0", package_list
 
-def get_source_package_list_from_url(url, section):
+def get_source_package_list_from_url(url, section, d):
     "Return a sectioned list of package names from a URL list"
 
     bb.note("Reading %s: %s" % (url, section))
-    links = get_links_from_url(url)
+    links = get_links_from_url(url, d)
     srpms = filter(is_src_rpm, links)
     names_list = map(package_name_from_srpm, srpms)
 
@@ -96,44 +116,44 @@ def get_source_package_list_from_url(url, section):
 
     return new_pkgs
 
-def get_latest_released_fedora_source_package_list():
+def get_latest_released_fedora_source_package_list(d):
     "Returns list of all the name os packages in the latest fedora distro"
-    latest = find_latest_numeric_release("http://archive.fedoraproject.org/pub/fedora/linux/releases/")
+    latest = find_latest_numeric_release("http://archive.fedoraproject.org/pub/fedora/linux/releases/", d)
 
-    package_names = get_source_package_list_from_url("http://archive.fedoraproject.org/pub/fedora/linux/releases/%s/Fedora/source/SRPMS/" % latest, "main")
+    package_names = get_source_package_list_from_url("http://archive.fedoraproject.org/pub/fedora/linux/releases/%s/Fedora/source/SRPMS/" % latest, "main", d)
 
 #    package_names += get_source_package_list_from_url("http://download.fedora.redhat.com/pub/fedora/linux/releases/%s/Everything/source/SPRMS/" % latest, "everything")
-    package_names += get_source_package_list_from_url("http://archive.fedoraproject.org/pub/fedora/linux/updates/%s/SRPMS/" % latest, "updates")
+    package_names += get_source_package_list_from_url("http://archive.fedoraproject.org/pub/fedora/linux/updates/%s/SRPMS/" % latest, "updates", d)
 
     package_list=clean_package_list(package_names)
         
     return latest, package_list
 
-def get_latest_released_opensuse_source_package_list():
+def get_latest_released_opensuse_source_package_list(d):
     "Returns list of all the name os packages in the latest opensuse distro"
-    latest = find_latest_numeric_release("http://download.opensuse.org/source/distribution/")
+    latest = find_latest_numeric_release("http://download.opensuse.org/source/distribution/",d)
 
-    package_names = get_source_package_list_from_url("http://download.opensuse.org/source/distribution/%s/repo/oss/suse/src/" % latest, "main")
-    package_names += get_source_package_list_from_url("http://download.opensuse.org/update/%s/rpm/src/" % latest, "updates")
+    package_names = get_source_package_list_from_url("http://download.opensuse.org/source/distribution/%s/repo/oss/suse/src/" % latest, "main", d)
+    package_names += get_source_package_list_from_url("http://download.opensuse.org/update/%s/rpm/src/" % latest, "updates", d)
 
     package_list=clean_package_list(package_names)
     return latest, package_list
 
-def get_latest_released_mandriva_source_package_list():
+def get_latest_released_mandriva_source_package_list(d):
     "Returns list of all the name os packages in the latest mandriva distro"
-    latest = find_latest_numeric_release("http://distrib-coffee.ipsl.jussieu.fr/pub/linux/MandrivaLinux/official/")
-    package_names = get_source_package_list_from_url("http://distrib-coffee.ipsl.jussieu.fr/pub/linux/MandrivaLinux/official/%s/SRPMS/main/release/" % latest, "main")
+    latest = find_latest_numeric_release("http://distrib-coffee.ipsl.jussieu.fr/pub/linux/MandrivaLinux/official/", d)
+    package_names = get_source_package_list_from_url("http://distrib-coffee.ipsl.jussieu.fr/pub/linux/MandrivaLinux/official/%s/SRPMS/main/release/" % latest, "main", d)
 #    package_names += get_source_package_list_from_url("http://distrib-coffee.ipsl.jussieu.fr/pub/linux/MandrivaLinux/official/%s/SRPMS/contrib/release/" % latest, "contrib")
-    package_names += get_source_package_list_from_url("http://distrib-coffee.ipsl.jussieu.fr/pub/linux/MandrivaLinux/official/%s/SRPMS/main/updates/" % latest, "updates")
+    package_names += get_source_package_list_from_url("http://distrib-coffee.ipsl.jussieu.fr/pub/linux/MandrivaLinux/official/%s/SRPMS/main/updates/" % latest, "updates", d)
 
     package_list=clean_package_list(package_names)
     return latest, package_list
 
-def find_latest_debian_release(url):
+def find_latest_debian_release(url, d):
     "Find the latest listed debian release on the given url"
 
     releases = []
-    for link in get_links_from_url(url):
+    for link in get_links_from_url(url, d):
         if link[:6] == "Debian":
             if ';' not in link:
                 releases.append(link)
@@ -143,16 +163,15 @@ def find_latest_debian_release(url):
     except:
         return "_NotFound_"
 
-def get_debian_style_source_package_list(url, section):
+def get_debian_style_source_package_list(url, section, d):
     "Return the list of package-names stored in the debian style Sources.gz file"
-    import urllib
-    sock = urllib.urlopen(url)
-    import tempfile
-    tmpfile = tempfile.NamedTemporaryFile(mode='wb', prefix='oecore.', suffix='.tmp', delete=False)
-    tmpfilename=tmpfile.name
-    tmpfile.write(sock.read())
-    sock.close()
-    tmpfile.close()
+    with create_socket(url,d) as sock:
+        webpage = sock.read()
+        import tempfile
+        tmpfile = tempfile.NamedTemporaryFile(mode='wb', prefix='oecore.', suffix='.tmp', delete=False)
+        tmpfilename=tmpfile.name
+        tmpfile.write(sock.read())
+        tmpfile.close()
     import gzip
     bb.note("Reading %s: %s" % (url, section))
 
@@ -165,41 +184,41 @@ def get_debian_style_source_package_list(url, section):
 
     return package_names
 
-def get_latest_released_debian_source_package_list():
+def get_latest_released_debian_source_package_list(d):
     "Returns list of all the name os packages in the latest debian distro"
-    latest = find_latest_debian_release("http://ftp.debian.org/debian/dists/")
+    latest = find_latest_debian_release("http://ftp.debian.org/debian/dists/", d)
     url = "http://ftp.debian.org/debian/dists/stable/main/source/Sources.gz" 
-    package_names = get_debian_style_source_package_list(url, "main")
+    package_names = get_debian_style_source_package_list(url, "main", d)
 #    url = "http://ftp.debian.org/debian/dists/stable/contrib/source/Sources.gz" 
 #    package_names += get_debian_style_source_package_list(url, "contrib")
     url = "http://ftp.debian.org/debian/dists/stable-proposed-updates/main/source/Sources.gz" 
-    package_names += get_debian_style_source_package_list(url, "updates")
+    package_names += get_debian_style_source_package_list(url, "updates", d)
     package_list=clean_package_list(package_names)
     return latest, package_list
 
-def find_latest_ubuntu_release(url):
+def find_latest_ubuntu_release(url, d):
     "Find the latest listed ubuntu release on the given url"
     url += "?C=M;O=D" # Descending Sort by Last Modified
-    for link in get_links_from_url(url):
+    for link in get_links_from_url(url, d):
         if link[-8:] == "-updates":
             return link[:-8]
     return "_NotFound_"
 
-def get_latest_released_ubuntu_source_package_list():
+def get_latest_released_ubuntu_source_package_list(d):
     "Returns list of all the name os packages in the latest ubuntu distro"
-    latest = find_latest_ubuntu_release("http://archive.ubuntu.com/ubuntu/dists/")
+    latest = find_latest_ubuntu_release("http://archive.ubuntu.com/ubuntu/dists/", d)
     url = "http://archive.ubuntu.com/ubuntu/dists/%s/main/source/Sources.gz" % latest
-    package_names = get_debian_style_source_package_list(url, "main")
+    package_names = get_debian_style_source_package_list(url, "main", d)
 #    url = "http://archive.ubuntu.com/ubuntu/dists/%s/multiverse/source/Sources.gz" % latest
 #    package_names += get_debian_style_source_package_list(url, "multiverse")
 #    url = "http://archive.ubuntu.com/ubuntu/dists/%s/universe/source/Sources.gz" % latest
 #    package_names += get_debian_style_source_package_list(url, "universe")
     url = "http://archive.ubuntu.com/ubuntu/dists/%s-updates/main/source/Sources.gz" % latest
-    package_names += get_debian_style_source_package_list(url, "updates")
+    package_names += get_debian_style_source_package_list(url, "updates", d)
     package_list=clean_package_list(package_names)
     return latest, package_list
 
-def create_distro_packages_list(distro_check_dir):
+def create_distro_packages_list(distro_check_dir, d):
     pkglst_dir = os.path.join(distro_check_dir, "package_lists")
     if not os.path.isdir (pkglst_dir):
         os.makedirs(pkglst_dir)
@@ -220,7 +239,7 @@ def create_distro_packages_list(distro_check_dir):
     begin = datetime.now()
     for distro in per_distro_functions:
         name = distro[0]
-        release, package_list = distro[1]()
+        release, package_list = distro[1](d)
         bb.note("Distro: %s, Latest Release: %s, # src packages: %d" % (name, release, len(package_list)))
         package_list_file = os.path.join(pkglst_dir, name + "-" + release)
         f = open(package_list_file, "w+b")
@@ -231,7 +250,7 @@ def create_distro_packages_list(distro_check_dir):
     delta = end - begin
     bb.note("package_list generatiosn took this much time: %d seconds" % delta.seconds)
 
-def update_distro_data(distro_check_dir, datetime):
+def update_distro_data(distro_check_dir, datetime, d):
     """
         If distro packages list data is old then rebuild it.
         The operations has to be protected by a lock so that
@@ -258,7 +277,7 @@ def update_distro_data(distro_check_dir, datetime):
         if saved_datetime[0:8] != datetime[0:8]:
             bb.note("The build datetime did not match: saved:%s current:%s" % (saved_datetime, datetime))
             bb.note("Regenerating distro package lists")
-            create_distro_packages_list(distro_check_dir)
+            create_distro_packages_list(distro_check_dir, d)
             f.seek(0)
             f.write(datetime)