Revision bump adds upstream patch to improve perfomance on resource identification spotted by scarabeus.

Package-Manager: portage-2.2.0_alpha98/cvs/Linux i686

4 files changed, 177 insertions, 9 deletions

diff --git a/kde-base/nepomuk/ChangeLog b/kde-base/nepomuk/ChangeLog
index 851a5b3c5863..2c83719847c0 100644
--- a/kde-base/nepomuk/ChangeLog
+++ b/kde-base/nepomuk/ChangeLog
@@ -1,6 +1,14 @@
 # ChangeLog for kde-base/nepomuk
 # Copyright 1999-2012 Gentoo Foundation; Distributed under the GPL v2
-# $Header: /var/cvsroot/gentoo-x86/kde-base/nepomuk/ChangeLog,v 1.119 2012/03/31 13:52:13 maekke Exp $
+# $Header: /var/cvsroot/gentoo-x86/kde-base/nepomuk/ChangeLog,v 1.120 2012/04/04 11:01:59 johu Exp $
+
+*nepomuk-4.8.1-r1 (04 Apr 2012)
+
+  04 Apr 2012; Johannes Huber <johu@gentoo.org>
+  +files/nepomuk-4.8.1-performance.patch, +nepomuk-4.8.1-r1.ebuild,
+  -nepomuk-4.8.1.ebuild:
+  Revision bump adds upstream patch to improve perfomance on resource
+  identification spotted by scarabeus.
 
   31 Mar 2012; Markus Meier <maekke@gentoo.org> nepomuk-4.8.1.ebuild:
   add ~arm, bug #402989
diff --git a/kde-base/nepomuk/Manifest b/kde-base/nepomuk/Manifest
index 64bb7ae4804c..495196102624 100644
--- a/kde-base/nepomuk/Manifest
+++ b/kde-base/nepomuk/Manifest
@@ -1,16 +1,21 @@
 -----BEGIN PGP SIGNED MESSAGE-----
-Hash: SHA1
+Hash: SHA512
 
+AUX nepomuk-4.8.1-performance.patch 5917 RMD160 33f9cd5c04cf4f929ac5b33bbdfe59b232f66e0c SHA1 bbe57e6d0f89f79d96cc5204287a20e5e507c708 SHA256 591f4596c2a04d8c049dcb04dab14d6365378df55b0d1952d10d0df7e31450f0
 DIST kde-runtime-4.7.4.tar.bz2 6026460 RMD160 0b35753ef335268d2705a61218fa92c88f13e4da SHA1 bf5c266b7748cda44cc3a2fb231a2d6dde2b09f6 SHA256 8d1bd0712ca8f9c6f4700f481a0201938165026b75e05fe7b0c5b9a30dd9dd1a
 DIST kde-runtime-4.8.1.tar.xz 5761688 RMD160 15122a1e39a1db63345742583c14d98b0e3e96bc SHA1 82b57dc38335716fb382a665b536a9aece4684d6 SHA256 1f2dae51075f93ffd357968e3805527cad32a92b12a8e8bbb1d135eef679c51b
 EBUILD nepomuk-4.7.4.ebuild 640 RMD160 70ac26c7d8c589dea6c043b2c8c92e77a22e2c04 SHA1 3d56b1e5303b4dc563630ff0e06d4e8aa93c82c0 SHA256 38452035bb36ce5987fd8f2e00a40b2a7ce92b33bf925e0c24f54e36aea93ce6
-EBUILD nepomuk-4.8.1.ebuild 630 RMD160 befe5ac9c96dbb271802414b785d8403982bae50 SHA1 23e3d940532cc4bb15195d78f44a8f0d42126b47 SHA256 838d7aa85c3536733d449ae3ce63c9b7f690c41244155a5a13d38607d126f9fd
-MISC ChangeLog 14553 RMD160 ba14108477120bb5d378d5def25b2428734b5bdd SHA1 01200cdc9c8cb6d25564e1005f2c38cd7fcfe56d SHA256 4582ab1a3683c47ed43fac3261db6942120963abf9fc57eb83b92e13d00e3f32
+EBUILD nepomuk-4.8.1-r1.ebuild 682 RMD160 cadc0f63d12ed81df9439fbf261882cc303efcb5 SHA1 3be04cc4d59d35b2d049f34f45bdad8207e4300a SHA256 d54079bdb35324bbea72209d7e2c93e2f2ac58f4c832aeddd5dd3b7d7d01927f
+MISC ChangeLog 14835 RMD160 aed75f9bcfce59df475634aa65a5b3e8e3c052ef SHA1 369439d4ba1f90eaeaca081a0b676b84294d1842 SHA256 0010ad525ab904d07f69bdd7b184fe04fe82b2413a93727bf90fd10fd95d0f9b
 MISC metadata.xml 156 RMD160 ecce3b981f150c45ae1e84e2d208e678d6124259 SHA1 b64f7c0b4e5db816d82ad19848f72118af129d35 SHA256 2f4da28506b9d4185f320f67a6191d30c7a921217ed4447ed46ea0bc4aefc79a
 -----BEGIN PGP SIGNATURE-----
-Version: GnuPG v2.0.17 (GNU/Linux)
+Version: GnuPG v2.0.18 (GNU/Linux)
 
-iEYEARECAAYFAk93DLkACgkQkKaRLQcq0GKyPgCeKTprIUXB0Yxd7Xyf0btwu1PQ
-KOgAn1ngijzARerqwxcjkL/CmiVCwKIJ
-=NCsl
+iQEcBAEBCgAGBQJPfCpCAAoJEO+t9ga+3I3a6jcH/Ajck8ejwXXkseXFPQKpefBf
+/gXaVKm+T+c7LTc9BWOwDdTKis/2Dv8Y6wuGMBmI3ndwky19YyuvX3VjZVjJ+t3C
+pUUnRNEZsO2oIF679MAq/eZPLO9xvq8tH541RHfgD914YZardncxxKV9OEuIHkME
+qmEJqnbA+6Oa3+584E9yWWLLRjyEN6AhsR7eorvYnQVdvfTbG3g+vVGL4F3fJQdo
+SXpGmy5mpgsGzUe+9/4tQYH6j3HCO+9qXTL/q79esqk8IdYeZkF3sCSYJCQnUkB/
+ch2IJmo2STBAlSOuf/mJJCvFF6h+wNhMwnQ14ro44zoyaGSz1fNYs42hjpie3Cw=
+=sU3k
 -----END PGP SIGNATURE-----
diff --git a/kde-base/nepomuk/files/nepomuk-4.8.1-performance.patch b/kde-base/nepomuk/files/nepomuk-4.8.1-performance.patch
new file mode 100644
index 000000000000..06d01e7702db
--- /dev/null
+++ b/kde-base/nepomuk/files/nepomuk-4.8.1-performance.patch
@@ -0,0 +1,151 @@
+commit 754275eda610dce1160286a76339353097d8764c
+Author: Sebastian Trueg <trueg@kde.org>
+Date:   Fri Mar 9 17:17:48 2012 +0100
+
+    Backport from nepomuk-core: improved performance on res identification.
+    
+    BUG: 289932
+    FIXED-IN: 4.8.2
+
+diff --git a/nepomuk/services/backupsync/lib/resourceidentifier.cpp b/nepomuk/services/backupsync/lib/resourceidentifier.cpp
+index c1a9919..894372c 100644
+--- a/nepomuk/services/backupsync/lib/resourceidentifier.cpp
++++ b/nepomuk/services/backupsync/lib/resourceidentifier.cpp
+@@ -31,6 +31,7 @@
+ #include <Soprano/Statement>
+ #include <Soprano/Graph>
+ #include <Soprano/Node>
++#include <Soprano/BindingSet>
+ #include <Soprano/StatementIterator>
+ #include <Soprano/QueryResultIterator>
+ #include <Soprano/Model>
+@@ -176,19 +177,18 @@ bool Nepomuk::Sync::ResourceIdentifier::runIdentification(const KUrl& uri)
+         return false;
+     }
+ 
+-    QString query;
+-
+     QStringList identifyingProperties;
+     QHash<KUrl, Soprano::Node> identifyingPropertiesHash;
+ 
+     QHash< KUrl, Soprano::Node >::const_iterator it = res.constBegin();
+     QHash< KUrl, Soprano::Node >::const_iterator constEnd = res.constEnd();
++    QList<Soprano::Node> requiredTypes;
+     for( ; it != constEnd; it++ ) {
+         const QUrl & prop = it.key();
+ 
+         // Special handling for rdf:type
+         if( prop == RDF::type() ) {
+-            query += QString::fromLatin1(" ?r a %1 . ").arg( it.value().toN3() );
++            requiredTypes << it.value().uri();
+             continue;
+         }
+ 
+@@ -219,6 +219,10 @@ bool Nepomuk::Sync::ResourceIdentifier::runIdentification(const KUrl& uri)
+         return false;
+     }
+ 
++
++    // construct the identification query
++    QString query = QLatin1String("select distinct ?r where { ");
++
+     //
+     // Optimization:
+     // If there is only one identifying property using all that optional and filter stuff
+@@ -235,7 +239,7 @@ bool Nepomuk::Sync::ResourceIdentifier::runIdentification(const KUrl& uri)
+                            QString::number( numIdentifyingProperties++ ) );
+         }
+ 
+-        // Make sure atleast one of the identification properties has been matched
++        // Make sure at least one of the identification properties has been matched
+         // by adding filter( bound(?o1) || bound(?o2) ... )
+         query += QString::fromLatin1("filter( ");
+         for( int i=0; i<numIdentifyingProperties-1; i++ ) {
+@@ -247,43 +251,68 @@ bool Nepomuk::Sync::ResourceIdentifier::runIdentification(const KUrl& uri)
+         query += QString::fromLatin1("?r %1 %2 . ").arg(Soprano::Node::resourceToN3(identifyingPropertiesHash.constBegin().key()),
+                                                          identifyingPropertiesHash.constBegin().value().toN3());
+     }
+-    query += QLatin1String("}");
+ 
+-    // Construct the entire query
+-    QString queryBegin = QString::fromLatin1("select distinct ?r count(?p) as ?cnt "
+-    "where { ?r ?p ?o. filter( ?p in (%1) ).")
+-    .arg( identifyingProperties.join(",") );
+-
+-    query = queryBegin + query + QString::fromLatin1(" order by desc(?cnt)");
++    //
++    // For performance reasons we add a limit even though this could mean that we
++    // miss a resource to identify since we check the types below.
++    //
++    query += QLatin1String("} LIMIT 100");
+ 
+-    kDebug() << query;
+ 
+     //
+-    // Only store the results which have the maximum score
++    // Fetch a score for each result.
++    // We do this in a separate query for performance reasons.
+     //
+-    QSet<KUrl> results;
+-    int score = -1;
++    QMultiHash<int, KUrl> resultsScoreHash;
++    int maxScore = -1;
+     Soprano::QueryResultIterator qit = d->m_model->executeQuery( query, Soprano::Query::QueryLanguageSparql );
+     while( qit.next() ) {
+-        //kDebug() << "RESULT: " << qit["r"] << " " << qit["cnt"];
++        const Soprano::Node r(qit["r"]);
++
++        //
++        // Check the type requirements. Experiments have shown this to mean a substantial
++        // performance boost as compared to doing it in the main query.
++        //
++        if(!requiredTypes.isEmpty() ) {
++            query = QLatin1String("ask where { ");
++            foreach(const Soprano::Node& type, requiredTypes) {
++                query += QString::fromLatin1("%1 a %2 . ").arg(r.toN3(), type.toN3());
++            }
++            query += QLatin1String("}");
++            if(!d->m_model->executeQuery(query, Soprano::Query::QueryLanguageSparql).boolValue()) {
++                continue;
++            }
++        }
++
++
++        const int score = d->m_model->executeQuery(QString::fromLatin1("select count(?p) as ?cnt where { "
++                                                                       "%1 ?p ?o. filter( ?p in (%2) ) . }")
++                                                   .arg( r.toN3(),
++                                                         identifyingProperties.join(",") ),
++                                                   Soprano::Query::QueryLanguageSparql)
++                          .allBindings().first()["cnt"].literal().toInt();
+ 
+-        int count = qit["cnt"].literal().toInt();
+-        if( score == -1 ) {
+-            score = count;
++        if( maxScore < score ) {
++            maxScore = score;
+         }
+-        else if( count < score )
+-            break;
+ 
+-        results << qit["r"].uri();
++        resultsScoreHash.insert(score, r.uri());
+     }
+ 
++    //
++    // Only get the results which have the maximum score
++    //
++    QSet<KUrl> results = QSet<KUrl>::fromList(resultsScoreHash.values(maxScore));
++
++
+     //kDebug() << "Got " << results.size() << " results";
+     if( results.empty() )
+         return false;
+ 
+     KUrl newUri;
+-    if( results.size() == 1 )
++    if( results.size() == 1 ) {
+         newUri = *results.begin();
++    }
+     else {
+         kDebug() << "DUPLICATE RESULTS!";
+         newUri = duplicateMatch( res.uri(), results );
diff --git a/kde-base/nepomuk/nepomuk-4.8.1.ebuild b/kde-base/nepomuk/nepomuk-4.8.1-r1.ebuild
index eb1a6c42229f..705c403b13db 100644
--- a/kde-base/nepomuk/nepomuk-4.8.1.ebuild
+++ b/kde-base/nepomuk/nepomuk-4.8.1-r1.ebuild
@@ -1,6 +1,6 @@
 # Copyright 1999-2012 Gentoo Foundation
 # Distributed under the terms of the GNU General Public License v2
-# $Header: /var/cvsroot/gentoo-x86/kde-base/nepomuk/nepomuk-4.8.1.ebuild,v 1.3 2012/03/31 13:52:13 maekke Exp $
+# $Header: /var/cvsroot/gentoo-x86/kde-base/nepomuk/nepomuk-4.8.1-r1.ebuild,v 1.1 2012/04/04 11:01:59 johu Exp $
 
 EAPI=4
 
@@ -22,3 +22,7 @@ RDEPEND="${DEPEND}"
 
 RESTRICT="test"
 # bug 392989
+
+PATCHES=(
+	"${FILESDIR}/${P}-performance.patch"
+)