Compare keywords not as exact line match but separately. This allows multiple combine...

svn path=/trunk/gentoolkit/; revision=859

diff --git a/pym/gentoolkit/eshowkw/keywords_content.py b/pym/gentoolkit/eshowkw/keywords_content.py
index 530e3f6983e8e651e9814e354198599b4c4c418e..56b856be95d07f37779fa31f0cb6c42e910ebb67 100644 (file)
--- a/pym/gentoolkit/eshowkw/keywords_content.py
+++ b/pym/gentoolkit/eshowkw/keywords_content.py
@@ -38,6 +38,11 @@ class keywords_content:
                                result.append(item)
                        return result
 
+               def __cleanKeyword(self, keyword):
+                       """Remove masked arches and hardmasks from keywords since we don't care about that."""
+                       return ["%s" % x for x in keyword.split()
+                               if x != '-*' and not x.startswith('-')]
+
                def __listRedundantAll(self, keywords):
                        """Search for redundant packages using all versions ignoring its slotting."""
                        return list(self.__compareSelected(list(keywords)))
@@ -61,26 +66,20 @@ class keywords_content:
 
                def __compareKeywordWithRest(self, keyword, keywords):
                        """Compare keywords with list of keywords."""
-                       for key in keywords:
-                               if self.__checkShadow(keyword, key):
+                       kw = self.__cleanKeyword(keyword)
+                       for kwi in keywords:
+                               kwi = self.__cleanKeyword(kwi)
+                               kw = self.__checkShadow(kw, kwi)
+                               if not kw:
                                        return True
                        return False
 
                def __checkShadow(self, old, new):
                        """Check if package version is overshadowed by other package version."""
-                       # remove -* and -arch since they are useless for us
-                       newclean = ["%s" % x for x in new.split()
-                               if x != '-*' and not x.startswith('-')]
-                       oldclean = ["%s" % x for x in old.split()
-                               if x != '-*' and not x.startswith('-')]
-
-                       tmp = set(newclean)
-                       tmp.update("~%s" % x for x in newclean
+                       tmp = set(new)
+                       tmp.update("~%s" % x for x in new
                                if not x.startswith("~"))
-                       if not set(oldclean).difference(tmp):
-                               return True
-                       else:
-                               return False
+                       return list(set(old).difference(tmp))
 
                def __init__(self, keywords, slots, ignore_slots = False):
                        """Query all relevant data for redundancy package checking"""