KeywordsManager: use lists where order matters

Also remove x.lstrip('~') code from from getRawMissingKeywords() since
I don't understand it (was only recently added, and we can add back
later if appropriate).

diff --git a/pym/portage/package/ebuild/_config/KeywordsManager.py b/pym/portage/package/ebuild/_config/KeywordsManager.py
index d153e956d3a85f5339acac3730dd5d7b6c5b2653..247dcb43a9390fc9eefca296d6c8b777a3b4b01c 100644 (file)
--- a/pym/portage/package/ebuild/_config/KeywordsManager.py
+++ b/pym/portage/package/ebuild/_config/KeywordsManager.py
@@ -118,11 +118,11 @@ class KeywordsManager(object):
 
                mygroups = self.getKeywords(cpv, slot, keywords, repo)
                # Repoman may modify this attribute as necessary.
-               pgroups = set(global_accept_keywords.split())
+               pgroups = global_accept_keywords.split()
 
                unmaskgroups = self.getPKeywords(cpv, slot, repo,
                                global_accept_keywords)
-               pgroups.update(unmaskgroups)
+               pgroups.extend(unmaskgroups)
 
                # Hack: Need to check the env directly here as otherwise stacking
                # doesn't work properly as negative values are lost in the config
@@ -130,7 +130,9 @@ class KeywordsManager(object):
                egroups = backuped_accept_keywords.split()
 
                if unmaskgroups or egroups:
-                       pgroups = self._getEgroups(egroups, pgroups.copy())
+                       pgroups = self._getEgroups(egroups, pgroups)
+               else:
+                       pgroups = set(pgroups)
 
                return self._getMissingKeywords(cpv, pgroups, mygroups)
 
@@ -166,18 +168,16 @@ class KeywordsManager(object):
 
                mygroups = self.getKeywords(cpv, slot, keywords, repo)
                # Repoman may modify this attribute as necessary.
-               pgroups = set(global_accept_keywords.split())
+               pgroups = global_accept_keywords.split()
 
                # Hack: Need to check the env directly here as otherwise stacking
                # doesn't work properly as negative values are lost in the config
                # object (bug #139600)
-               # we want to use the environment keywords here,
-               # but stripped to it's base arch
-               # we want the raw keywords needed to be accepted from the ebuild
                if backuped_accept_keywords:
                        egroups = self._getEgroups(backuped_accept_keywords.split(),
-                                       pgroups.copy())
-                       pgroups = set([x.lstrip('~') for x in egroups])
+                                       pgroups)
+               else:
+                       pgroups = set(pgroups)
 
                missing = self._getMissingKeywords(cpv, pgroups, mygroups)
 
@@ -193,10 +193,11 @@ class KeywordsManager(object):
                @rtype: List
                @return: list of KEYWORDS that have been accepted
                """
-               mygroups.update(egroups)
+               mygroups = list(mygroups)
+               mygroups.extend(egroups)
                inc_pgroups = set()
                for x in mygroups:
-                       if x.startswith("-"):
+                       if x[:1] == "-":
                                if x == "-*":
                                        inc_pgroups.clear()
                                else: