Merge branch 'master' of git://projects.archlinux.org/archweb

Conflicts:
	packages/urls.py
	templates/packages/details.html

2 files changed, 68 insertions, 51 deletions

diff --git a/devel/management/commands/rematch_packager.py b/devel/management/commands/rematch_packager.py
new file mode 100644
index 00000000..ba6e6a54
--- /dev/null
+++ b/devel/management/commands/rematch_packager.py
@@ -0,0 +1,64 @@
+# -*- coding: utf-8 -*-
+"""
+rematch_packager command
+
+Match all packages with a packager_str but NULL packager_id to a packager if we
+can find one.
+
+Usage: ./manage.py rematch_packager
+"""
+
+from django.core.management.base import NoArgsCommand
+
+import sys
+import logging
+
+from devel.utils import UserFinder
+from main.models import Package
+
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s -> %(levelname)s: %(message)s',
+    datefmt='%Y-%m-%d %H:%M:%S',
+    stream=sys.stderr)
+logger = logging.getLogger()
+
+class Command(NoArgsCommand):
+    help = "Runs a check on all active mirror URLs to determine if they are reachable via IPv4 and/or v6."
+
+    def handle_noargs(self, **options):
+        v = int(options.get('verbosity', None))
+        if v == 0:
+            logger.level = logging.ERROR
+        elif v == 1:
+            logger.level = logging.INFO
+        elif v == 2:
+            logger.level = logging.DEBUG
+
+        return match_packager()
+
+def match_packager():
+    finder = UserFinder()
+    logger.info("getting all unmatched packages")
+    package_count = matched_count = 0
+    unknown = set()
+
+    for package in Package.objects.filter(packager__isnull=True):
+        logger.debug("package %s, packager string %s",
+                package.pkgname, package.packager_str)
+        package_count += 1
+        user = finder.find(package.packager_str)
+        if user:
+            package.packager = user
+            logger.debug("  found user %s" % user.username)
+            package.save()
+            matched_count += 1
+        else:
+            unknown.add(package.packager_str)
+
+    logger.info("%d packages checked, %d newly matched",
+            package_count, matched_count)
+    logger.debug("unknown packagers:\n%s",
+            "\n".join(unknown))
+
+# vim: set ts=4 sw=4 et:
diff --git a/devel/management/commands/reporead.py b/devel/management/commands/reporead.py
index baf7fee1..470b785d 100644
--- a/devel/management/commands/reporead.py
+++ b/devel/management/commands/reporead.py
@@ -16,7 +16,6 @@ Example:
 from django.core.management.base import BaseCommand, CommandError
 from django.contrib.auth.models import User
 from django.db import transaction
-from django.db.models import Q
 
 from collections import defaultdict
 import io
@@ -28,6 +27,7 @@ import logging
 from datetime import datetime
 from optparse import make_option
 
+from devel.utils import UserFinder
 from main.models import Arch, Package, PackageDepend, PackageFile, Repo
 from packages.models import Conflict, Provision, Replacement
 
@@ -130,55 +130,6 @@ class Pkg(object):
         return u'%s-%s' % (self.ver, self.rel)
 
 
-def find_user(userstring):
-    '''
-    Attempt to find the corresponding User object for a standard
-    packager string, e.g. something like
-        'A. U. Thor <author@example.com>'.
-    We start by searching for a matching email address; we then move onto
-    matching by first/last name. If we cannot find a user, then return None.
-    '''
-    if userstring in find_user.cache:
-        return find_user.cache[userstring]
-    matches = re.match(r'^([^<]+)? ?<([^>]*)>', userstring)
-    if not matches:
-        return None
-
-    user = None
-    name = matches.group(1)
-    email = matches.group(2)
-
-    def user_email():
-        return User.objects.get(email=email)
-    def profile_email():
-        return User.objects.get(userprofile__public_email=email)
-    def user_name():
-        # yes, a bit odd but this is the easiest way since we can't always be
-        # sure how to split the name. Ensure every 'token' appears in at least
-        # one of the two name fields.
-        name_q = Q()
-        for token in name.split():
-            # ignore quoted parts; e.g. nicknames in strings
-            if re.match(r'^[\'"].*[\'"]$', token):
-                continue
-            name_q &= (Q(first_name__icontains=token) |
-                    Q(last_name__icontains=token))
-        return User.objects.get(name_q)
-
-    for matcher in (user_email, profile_email, user_name):
-        try:
-            user = matcher()
-            break
-        except (User.DoesNotExist, User.MultipleObjectsReturned):
-            pass
-
-    find_user.cache[userstring] = user
-    return user
-
-# cached mappings of user strings -> User objects so we don't have to do the
-# lookup more than strictly necessary.
-find_user.cache = {}
-
 DEPEND_RE = re.compile(r"^(.+?)((>=|<=|=|>|<)(.*))?$")
 
 def create_depend(package, dep_str, optional=False):
@@ -231,6 +182,8 @@ def create_multivalued(dbpkg, repopkg, db_attr, repo_attr):
     for name in getattr(repopkg, repo_attr):
         collection.create(name=name)
 
+finder = UserFinder()
+
 def populate_pkg(dbpkg, repopkg, force=False, timestamp=None):
     db_score = 1
 
@@ -249,7 +202,7 @@ def populate_pkg(dbpkg, repopkg, force=False, timestamp=None):
     dbpkg.build_date = repopkg.builddate
     dbpkg.packager_str = repopkg.packager
     # attempt to find the corresponding django user for this string
-    dbpkg.packager = find_user(repopkg.packager)
+    dbpkg.packager = finder.find(repopkg.packager)
 
     if timestamp:
         dbpkg.flag_date = None