Use urllib.urlopen() instead of portage.fetch() for fetching metadata.dtd.

svn path=/main/trunk/; revision=14709

diff --git a/bin/repoman b/bin/repoman
index c885b34244a3bdfd49bdb5b2b627df413e83831b..703d18e473ab14efc8b602506435b06d0871f11b 100755 (executable)
--- a/bin/repoman
+++ b/bin/repoman
@@ -25,6 +25,7 @@ import sys
 import tempfile
 import time
 import platform
+import urllib
 
 from io import StringIO
 from itertools import chain
@@ -440,6 +441,10 @@ suspect_rdepend = frozenset([
        "x11-misc/imake",
 ])
 
+metadata_dtd_uri = 'http://www.gentoo.org/dtd/metadata.dtd'
+# force refetch if the local copy creation time is older than this
+metadata_dtd_ctime_interval = 60 * 60 * 24 * 7 # 7 days
+
 # file.executable
 no_exec = frozenset(["Manifest","ChangeLog","metadata.xml"])
 
@@ -765,59 +770,80 @@ desktop_pattern = re.compile(r'.*\.desktop$')
 for x in qacats:
        stats[x]=0
        fails[x]=[]
+
 xmllint_capable = False
 metadata_dtd = os.path.join(repoman_settings["DISTDIR"], 'metadata.dtd')
-if options.mode == "manifest":
-       pass
-elif not find_binary('xmllint'):
-       print(red("!!! xmllint not found. Can't check metadata.xml.\n"))
-       if options.xml_parse or repolevel==3:
-               print(red("!!!")+" sorry, xmllint is needed.  failing\n")
-               sys.exit(1)
-else:
-       #hardcoded paths/urls suck. :-/
-       must_fetch=1
-       backup_exists=0
-       try:
-               # if it's been over a week since fetching (or the system clock is fscked), grab an updated copy of metadata.dtd 
-               # clock is fscked or it's been a week. time to grab a new one.
-               ct = os.stat(metadata_dtd)[ST_CTIME]
-               if abs(time.time() - ct) > (60*60*24*7):
-                       # don't trap the exception, we're watching for errno 2 (file not found), anything else is a bug.
-                       backup_exists=1
-               else:
-                       must_fetch=0
 
-       except (OSError,IOError) as e:
-               if e.errno != 2:
-                       print(red("!!!")+" caught exception '%s' for %s/metadata.dtd, bailing" % (str(e), portage.CACHE_PATH))
-                       sys.exit(1)
+def fetch_metadata_dtd():
+       """
+       Fetch metadata.dtd if it doesn't exist or the ctime is older than
+       metadata_dtd_ctime_interval.
+       @rtype: bool
+       @returns: True if successful, otherwise False
+       """
+
+       must_fetch = True
+       metadata_dtd_st = None
+       current_time = int(time.time())
+       try:
+               metadata_dtd_st = os.stat(metadata_dtd)
+       except EnvironmentError as e:
+               if e.errno not in (errno.ENOENT, errno.ESTALE):
+                       raise
+               del e
+       else:
+               # Trigger fetch if metadata.dtd mtime is old or clock is wrong.
+               if abs(current_time - metadata_dtd_st.st_ctime) \
+                       < metadata_dtd_ctime_interval:
+                       must_fetch = False
 
        if must_fetch:
-               print() 
-               print(green("***")+" the local copy of metadata.dtd needs to be refetched, doing that now")
                print()
-               val = 0
+               print(green("***") + " the local copy of metadata.dtd " + \
+                       "needs to be refetched, doing that now")
+               print()
                try:
+                       url_f = urllib.urlopen(metadata_dtd_uri)
+                       last_modified = url_f.info().getdate('last-modified')
+                       if last_modified is not None:
+                               last_modified = time.mktime(last_modified)
+
+                       metadata_dtd_tmp = "%s.%s" % (metadata_dtd, os.getpid())
                        try:
-                               os.unlink(metadata_dtd)
-                       except OSError as e:
-                               if e.errno != errno.ENOENT:
-                                       raise
-                               del e
-                       val=portage.fetch(['http://www.gentoo.org/dtd/metadata.dtd'],repoman_settings,fetchonly=0, \
-                               try_mirrors=0)
+                               local_f = open(metadata_dtd_tmp, mode='wb')
+                               local_f.write(url_f.read())
+                               local_f.close()
+                               if last_modified is not None:
+                                       os.utime(metadata_dtd_tmp,
+                                               (int(last_modified), int(last_modified)))
+                               os.rename(metadata_dtd_tmp, metadata_dtd)
+                       finally:
+                               try:
+                                       os.unlink(metadata_dtd_tmp)
+                               except OSError:
+                                       pass
 
-               except SystemExit as e:
-                       raise  # Need to propogate this
-               except Exception as e:
+                       url_f.close()
+
+               except EnvironmentError as e:
                        print()
-                       print(red("!!!")+" attempting to fetch 'http://www.gentoo.org/dtd/metadata.dtd', caught")
-                       print(red("!!!")+" exception '%s' though." % str(e))
-                       val=0
-               if not val:
+                       print(red("!!!")+" attempting to fetch '%s', caught" % metadata_dtd_uri)
+                       print(red("!!!")+" exception '%s' though." % (e,))
                        print(red("!!!")+" fetching new metadata.dtd failed, aborting")
-                       sys.exit(1)
+                       return False
+
+       return True
+
+if options.mode == "manifest":
+       pass
+elif not find_binary('xmllint'):
+       print(red("!!! xmllint not found. Can't check metadata.xml.\n"))
+       if options.xml_parse or repolevel==3:
+               print(red("!!!")+" sorry, xmllint is needed.  failing\n")
+               sys.exit(1)
+else:
+       if not fetch_metadata_dtd():
+               sys.exit(1)
        #this can be problematic if xmllint changes their output
        xmllint_capable=True