7 years agoHook up the ResourceMerger and ResourceWatcher for new resource notifications
Vishesh Handa [Tue, 12 Jul 2011 19:08:37 +0000 (00:38 +0530)]
Hook up the ResourceMerger and ResourceWatcher for new resource notifications

It should also emit notifications for other statements that are added

7 years agoAdded testStoreResources_createResources
Vishesh Handa [Tue, 12 Jul 2011 18:59:51 +0000 (00:29 +0530)]
Added testStoreResources_createResources

This tests the resourceCreated() signal of the ResourceWatcher.

7 years agoShare ontology data between tests
Vishesh Handa [Tue, 12 Jul 2011 16:01:34 +0000 (21:31 +0530)]
Share ontology data between tests

The DataMangementModel tests, and ResourceWatcher tests require
the ontologies to be present. They are manually inserted by a
number of model->addStatement( ) calls.

This patch moves all those calls to a common function
Nepomuk::insertOntologies, which can be called by both thest tests.

7 years agoemit visiblechanged when it should
Marco Martin [Tue, 12 Jul 2011 18:30:22 +0000 (20:30 +0200)]
emit visiblechanged when it should

7 years agoAdd files
Albert Astals Cid [Tue, 12 Jul 2011 15:43:49 +0000 (16:43 +0100)]
Add files

For files that are desktop-like but do not have a .desktop extension

7 years agoFix off by one error
Albert Astals Cid [Tue, 12 Jul 2011 13:28:06 +0000 (14:28 +0100)]
Fix off by one error

Passing 1000 to readlink means readlink can return up to 1000 in n
which will be an out of bouds access in buffer2[ n ] = 0;

7 years agoAllow addTriplet to add information about the main resource.
Sebastian Trueg [Tue, 12 Jul 2011 13:11:17 +0000 (15:11 +0200)]
Allow addTriplet to add information about the main resource.

So far we worked under the assumption that information about the main
resource would only be added through the addValue methods. That is not
always the case. There is for example the odf analyzer which does use
addTriplet instead. Thus, we need to check the subject for the path of
the main resource before continuing to map sub-resources.

7 years agoImprove storeResources file resolution
Vishesh Handa [Tue, 12 Jul 2011 09:26:17 +0000 (14:56 +0530)]
Improve storeResources file resolution

This patch does three things -
1. Use QUrl::toLocalFile, when testing if the url exists using QFileInfo
2. Before resolving all the url present in the objects. Convert them
   from the QVariant into Soprano::Node using the ClassAndPropertyTree,
   this way based on the type, the variant is converted to the correct
   node. Resolution is a lot simpler ( and correct ) this way.
3. Fix SyncResource blank node handling - When blank nodes are stored
   as QUrls they should be stored as "_:identifier" not "identifier"

This fixes testStoreResources_folder ( and strigi indexing )

7 years agoAdd testStoreResources_folder
Vishesh Handa [Tue, 12 Jul 2011 09:24:33 +0000 (14:54 +0530)]
Add testStoreResources_folder

This tests if storeResources converts urls of the form '/tmp/' to
their respective resource uris automatically. Even when the '/tmp/'
is passed as a string and not a QUrl

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Tue, 12 Jul 2011 08:36:11 +0000 (10:36 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agocrash-- move after first call to applet()
Marco Martin [Mon, 11 Jul 2011 20:44:05 +0000 (22:44 +0200)]
crash-- move after first call to applet()

7 years agoinsert a catalog named as the plugin name
Marco Martin [Mon, 11 Jul 2011 19:23:32 +0000 (21:23 +0200)]
insert a catalog named as the plugin name

7 years agoMake the ResourceWatcher's signature match its adaptors
Vishesh Handa [Mon, 11 Jul 2011 16:56:29 +0000 (22:26 +0530)]
Make the ResourceWatcher's signature match its adaptors

This way the signals can propogate. We have to use QDbusVariant
instead of QVariant.

7 years agoFix VHanda's commit breaking. resourceAdded requires not QDBusVariant.
Shaun Reich [Mon, 11 Jul 2011 15:41:39 +0000 (11:41 -0400)]
Fix VHanda's commit breaking. resourceAdded requires not QDBusVariant.

7 years agoMake the Resource Watcher work
Vishesh Handa [Mon, 11 Jul 2011 13:00:21 +0000 (18:30 +0530)]
Make the Resource Watcher work

1. User QDbusVariant instead of QVaraint
2. Generate the adaptor via qt4_add_dbus_adaptor - I have no no idea why
   just registering the object over the sessions bus and exporting all the
   signals doesn't work.

   doesn't work = Crashes with QMetaObject::activate faults.

7 years agoMerge branch 'nepomuk/megerRefactoring'
Vishesh Handa [Mon, 11 Jul 2011 10:32:25 +0000 (16:02 +0530)]
Merge branch 'nepomuk/megerRefactoring'

7 years agoDocumentation++
Vishesh Handa [Mon, 11 Jul 2011 10:26:39 +0000 (15:56 +0530)]

7 years agoRemove the Transaction Model
Vishesh Handa [Mon, 11 Jul 2011 10:05:54 +0000 (15:35 +0530)]
Remove the Transaction Model

The ResourceMerger does all its checks without modifying the model.
Thus the Transaction Model is no longer required.

This should improve performance a lot, each statement that is being
pushed no longer needs to be logged.

7 years agoDo not allow graph merging to fail
Vishesh Handa [Mon, 11 Jul 2011 10:00:08 +0000 (15:30 +0530)]
Do not allow graph merging to fail

This is based on the assumption that if 2 valid graphs are
combined, then the new one is also a valid graph.

If someone can find me a real world example where that is not the
case, then I'll revert this patch.

7 years agoFix the tests
Vishesh Handa [Mon, 11 Jul 2011 09:53:01 +0000 (15:23 +0530)]
Fix the tests

Some of these tests tried to create custom resource uri like
"nepomuk:/res/m/Res1". This is no longer allowed.

All resource uris must be created by the DMS.

7 years agoDo not allow the creation of custom nepomuk uris
Vishesh Handa [Mon, 11 Jul 2011 09:51:46 +0000 (15:21 +0530)]
Do not allow the creation of custom nepomuk uris

Uri of the form "nepomuk:/anything" will not be allowed, unless
they already exist. All uris must be created by the Data management

7 years agoAdd nao:created and nao:lastModified in the ResourceMerger
Vishesh Handa [Mon, 11 Jul 2011 09:42:26 +0000 (15:12 +0530)]
Add nao:created and nao:lastModified in the ResourceMerger

This makes the design a lot cleaner. Prior to this, they were
added in storeResources as extra statements.

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Mon, 11 Jul 2011 08:55:46 +0000 (10:55 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoFaster Cardinality checks
Vishesh Handa [Mon, 11 Jul 2011 08:24:46 +0000 (13:54 +0530)]
Faster Cardinality checks

This should be theoretically faster, but I haven't run any benchmarks.

7 years agoCode cleanup
Vishesh Handa [Mon, 11 Jul 2011 07:19:48 +0000 (12:49 +0530)]
Code cleanup

Throw away a lot of redundant code.

7 years agoDo all the graphs checks before pushing any statement
Vishesh Handa [Mon, 11 Jul 2011 06:55:42 +0000 (12:25 +0530)]
Do all the graphs checks before pushing any statement

This makes the algorithm used a lot simpler to understand, and
will allow us to remove the Transaction Model thereby making
mergers faster.

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Sun, 10 Jul 2011 07:21:33 +0000 (09:21 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoFix typo.
Frederik Schwarzer [Sat, 9 Jul 2011 19:42:14 +0000 (21:42 +0200)]
Fix typo.

device -> devices

7 years agoFix wording.
Frederik Schwarzer [Sat, 9 Jul 2011 19:36:41 +0000 (21:36 +0200)]
Fix wording.

7 years agoFix last commit
Christoph Feck [Sat, 9 Jul 2011 16:22:01 +0000 (18:22 +0200)]
Fix last commit


7 years agoCompile without Nepomuk
Olivier Trichet [Sat, 9 Jul 2011 15:46:34 +0000 (17:46 +0200)]
Compile without Nepomuk

CCMAIL: Ivan Čukić <>

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Sat, 9 Jul 2011 07:39:47 +0000 (09:39 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoFix compile warning
Montel Laurent [Fri, 8 Jul 2011 20:06:29 +0000 (22:06 +0200)]
Fix compile warning

7 years agoDo not initialize the ResourceManager's mainModel
Vishesh Handa [Fri, 8 Jul 2011 17:23:48 +0000 (22:53 +0530)]
Do not initialize the ResourceManager's mainModel

The ResourceIdentifier used to call Sync::ResourceIdentifier which
used to set the model to ResourceManager::instance()->mainModel(),
and then a couple of lines later, we would explicitly set the
ResourceIdentifer's model via setModel( )

Nepomuk::ResourceManager::instance()->mainModel() should never
be called in the storage service. We now set the model directly in
the ResourceIdentifier's constructor.

7 years agoMerge branch 'master' of git://
Ivan Čukić [Fri, 8 Jul 2011 16:39:24 +0000 (18:39 +0200)]
Merge branch 'master' of git://

7 years agoPorted the new activity manager to master
Ivan Čukić [Fri, 8 Jul 2011 16:37:56 +0000 (18:37 +0200)]
Ported the new activity manager to master

7 years agoCreate the subject node properly
Vishesh Handa [Fri, 8 Jul 2011 16:01:54 +0000 (21:31 +0530)]
Create the subject node properly

The node should be of its proper type, and not store the identifier
as the url like QUrl("_:adf")

7 years agoResourceMerger : Resolve the nodes in two steps
Vishesh Handa [Fri, 8 Jul 2011 15:56:58 +0000 (21:26 +0530)]
ResourceMerger : Resolve the nodes in two steps

Initially resolve all the mapped (identified) nodes, then perform
all the max cardinality and domain/range checks. Only after that
should the unmapped nodes be created.

Before this, the nodes were only resolved after the cardinality and
domain range checks. This resulted in the domain range checks
occasionally failing if the required type was in the store.

This fixes : testStoreResources_correctDomainInStore2

Resolving a node => changing it's uri to the identified resource

7 years agoCode cleanup : Use QUrl instead of KUrl
Vishesh Handa [Fri, 8 Jul 2011 13:46:26 +0000 (19:16 +0530)]
Code cleanup : Use QUrl instead of KUrl

Avoid mixing the two up unless absolutely necessary.
+ Indentation fixes due to copy pasting

7 years agoMerge Sync::ResourceMerger into the DMS ResourceMerger
Vishesh Handa [Fri, 8 Jul 2011 13:31:34 +0000 (19:01 +0530)]
Merge Sync::ResourceMerger into the DMS ResourceMerger

It's an absolute pain having to deal with the base class, specially
since I can't radically change the design. The current design is way
to complex.

I will hate myself for doing this, but it is necessary. This renders
the Sync::ResourceMerger almost obsolete.

7 years agoMerge branch 'master' of
Artem Serebriyskiy [Fri, 8 Jul 2011 12:57:03 +0000 (16:57 +0400)]
Merge branch 'master' of

7 years agoAdded testStoreResources_randomNepomukUri
Vishesh Handa [Fri, 8 Jul 2011 12:50:09 +0000 (18:20 +0530)]
Added testStoreResources_randomNepomukUri

This checks if storeResources allows us to create random uris
whose scheme is "nepomuk". We donot want to allow the user
to create their own nepomuk uris. All uris must be created by
the DataMangementModel

This test FAILS.

7 years agoCheck if the graph additionalMetadata is valid before pushing
Vishesh Handa [Fri, 8 Jul 2011 12:33:24 +0000 (18:03 +0530)]
Check if the graph additionalMetadata is valid before pushing

The graphMetadata will first be converted into a
QMulitHash<QUrl, Soprano::Node>, and then checked for validity. If
it fails, the entire merge operations fails.

This should improve the performance a little bit, as the
additionalMetadata is now being converted into a
QMultiHash<QUrl, Soprano::Node> only once.

This fixes testStoreResources_graphMetadataFail

7 years agoUse the ClassAndPropertyTree to convert QVariant to Soprano::Node
Vishesh Handa [Fri, 8 Jul 2011 12:20:05 +0000 (17:50 +0530)]
Use the ClassAndPropertyTree to convert QVariant to Soprano::Node

Added a member function toNodeHash( QHash<QUrl, QVariant>& ), that is
used to convert the additional metadata to Sopano::Nodes.

7 years agoMake streaming operators for Nepomuk::SimpleResource non-friends
Artem Serebriyskiy [Fri, 8 Jul 2011 12:01:31 +0000 (16:01 +0400)]
Make streaming operators for Nepomuk::SimpleResource non-friends

1. Streaming operators uses public API of Nepomuk::SimpleResource

7 years agoAdded testStoreResources_graphMetadataFail
Vishesh Handa [Fri, 8 Jul 2011 10:51:59 +0000 (16:21 +0530)]
Added testStoreResources_graphMetadataFail

This checks if the graph metadata is checked for validity.

7 years agoDo not recursion to delete sub-resources.
Sebastian Trueg [Fri, 8 Jul 2011 10:48:12 +0000 (12:48 +0200)]
Do not recursion to delete sub-resources.

This is the same optimization I used in removeDataByApplication.

7 years agoFixed error handling in resolveUrls.
Sebastian Trueg [Fri, 8 Jul 2011 10:42:17 +0000 (12:42 +0200)]
Fixed error handling in resolveUrls.

Previously errors were just ignored and executeQuery() would call
clearError() internally which would mean that an error created in
resolveUrl() could be lost in the next call.

This, however, required an additional method removeAllResources since
removeDataByApplication did call removeResources with a list that
might contain non-existing resource URIs which is an error case.
The new internal method does not perform any checks.

7 years agoAdded testStoreResources_correct(Domain|Range)InStore2
Vishesh Handa [Fri, 8 Jul 2011 10:31:06 +0000 (16:01 +0530)]
Added testStoreResources_correct(Domain|Range)InStore2

These tests check if the domain and range is stored are used, in
the case where the resource uri has not been provided and identification
is used.

7 years agoMerge branch 'simple_res_streaming_v2'
Artem Serebriyskiy [Fri, 8 Jul 2011 10:28:39 +0000 (14:28 +0400)]
Merge branch 'simple_res_streaming_v2'

7 years agoAdd extra methods to SimpleResource and SimpleResourceGraph
Artem Serebriyskiy [Wed, 6 Jul 2011 23:04:06 +0000 (03:04 +0400)]
Add extra methods to SimpleResource and SimpleResourceGraph

1. Add serializing operators ( streaming to QDataStream )
2. Add SimpleResourceGraph::addGraph for merging graphs
3. Add tests for these methods
4. Add operator== for SimpleResourceGraph

7 years agoThrottle the IndexCleaner the same way we do with the IndexScheduler.
Sebastian Trueg [Thu, 7 Jul 2011 15:33:23 +0000 (17:33 +0200)]
Throttle the IndexCleaner the same way we do with the IndexScheduler.

BUG: 276593

7 years agoSmall cleanup.
Sebastian Trueg [Thu, 7 Jul 2011 15:33:56 +0000 (17:33 +0200)]
Small cleanup.

7 years agoUpdate mappings for korgac.
Sergio Martins [Thu, 7 Jul 2011 20:22:30 +0000 (21:22 +0100)]
Update mappings for korgac.

korgac has it's own product now.

7 years agoFix testStoreResources_lazyCardinalities
Vishesh Handa [Thu, 7 Jul 2011 15:38:22 +0000 (21:08 +0530)]
Fix testStoreResources_lazyCardinalities

If you provide incorrect data, you cannot expect the identification to
be perfect.

If a res with nco:fullname "Superman" already exists and you try
to store a resource and provide nco:fullname "Superman", and
nco:fullname "Clark Kent", then the identification will not be
sucessful, as nco:fullname is an identiying property, and all
identifying properties must either match or not exist.

7 years agofix localIO API calls
Aaron Seigo [Thu, 7 Jul 2011 14:29:42 +0000 (16:29 +0200)]
fix localIO API calls

7 years agoImplement LazyCardinalities + A unit test
Vishesh Handa [Thu, 7 Jul 2011 11:03:53 +0000 (16:33 +0530)]
Implement LazyCardinalities + A unit test

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Thu, 7 Jul 2011 08:19:16 +0000 (10:19 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agoSVN_SILENT made messages (.desktop file)
Script Kiddy [Wed, 6 Jul 2011 08:41:04 +0000 (10:41 +0200)]
SVN_SILENT made messages (.desktop file)

7 years agobuild drag and drop plugin
Marco Martin [Tue, 5 Jul 2011 19:16:28 +0000 (21:16 +0200)]
build drag and drop plugin

7 years agodrag and drop plugin by Gregory Schlomoff
Marco Martin [Tue, 5 Jul 2011 19:13:22 +0000 (21:13 +0200)]
drag and drop plugin by Gregory Schlomoff


7 years agono monthly backup any more
Burkhard Lück [Tue, 5 Jul 2011 19:13:17 +0000 (21:13 +0200)]
no monthly backup any more

7 years agotitle of kcm doc changed to Web Browsing
Burkhard Lück [Tue, 5 Jul 2011 19:11:06 +0000 (21:11 +0200)]
title of kcm doc changed to Web Browsing

7 years agoOptimization: query the nepomukindexer app resource independantly.
Sebastian Trueg [Mon, 4 Jul 2011 12:47:53 +0000 (14:47 +0200)]
Optimization: query the nepomukindexer app resource independantly.

This highly improves the execution speed of the queries.
Something like 40sec vs. 400ms.

7 years agoAdded another unit test which checks sub-resource handling.
Sebastian Trueg [Mon, 4 Jul 2011 12:01:40 +0000 (14:01 +0200)]
Added another unit test which checks sub-resource handling.

sub-resources of sub-resources should be ignored when keeping sub-
resources due to usage from other resources.

7 years agoRemoved any legacy data removal from the filke indexer service.
Sebastian Trueg [Mon, 4 Jul 2011 11:58:26 +0000 (13:58 +0200)]
Removed any legacy data removal from the filke indexer service.

This task is now done by the data management service in method
removeDataByApplication. All that is left to do is to gather all
resources that have information stored by the file indexer, either in
the old way (via or the new
way (via nao:maintainedBy as created by the DMS).

7 years agoYet another optimization for removeDataForApplication.
Sebastian Trueg [Fri, 1 Jul 2011 23:44:15 +0000 (01:44 +0200)]
Yet another optimization for removeDataForApplication.

Querying the metadata is faster when one query per graph is performed.
I suspect that some expensive join is used to handle both filter(x in ...)

7 years agoDo not use recursion for subResource removal.
Sebastian Trueg [Fri, 1 Jul 2011 20:04:04 +0000 (22:04 +0200)]
Do not use recursion for subResource removal.

7 years agoRemove legacy indexer data in removeDataByApplication.
Sebastian Trueg [Fri, 1 Jul 2011 20:02:18 +0000 (22:02 +0200)]
Remove legacy indexer data in removeDataByApplication.

In KDE < 4.7 the file indexer used  adedicated property to mark
index graphs. Now this is done through the standard nao:maintaingApp.
With this patch we add a new special case where old indexer data is
also removed.

7 years agoMore performance optimization.
Sebastian Trueg [Fri, 1 Jul 2011 19:59:06 +0000 (21:59 +0200)]
More performance optimization.

Basic redesign of the removeDataByApplication method which does not
use nested loops anymore and tries to reduce the number of queries.

7 years agoMore and improved API documentation.
Sebastian Trueg [Thu, 30 Jun 2011 14:40:17 +0000 (16:40 +0200)]
More and improved API documentation.

7 years agoRemoved all redundant checks in the indexer.
Sebastian Trueg [Wed, 29 Jun 2011 09:36:45 +0000 (11:36 +0200)]
Removed all redundant checks in the indexer.

Now that DMS does perform all the necessary checks and conversions
there is no need for that in the indexer anymore. This makes the code
simpler and easier to maintain.

There is only one thing I had to introduce (not implemented yet though):
the LazyCardinalities flag for the storeResources method which will
make it simply drop and value that would violate a cardinality

7 years agoisShiftAsModifierAllowed() was moved to KKeyServer, update calls
Simon Persson [Sat, 2 Jul 2011 11:50:24 +0000 (19:50 +0800)]
isShiftAsModifierAllowed() was moved to KKeyServer, update calls

7 years agocall va_end()
Dirk Mueller [Fri, 1 Jul 2011 21:12:24 +0000 (23:12 +0200)]
call va_end()

7 years agoMerge branch 'nepomuk/noMoreRandomProtocols'
Vishesh Handa [Fri, 1 Jul 2011 14:10:54 +0000 (19:40 +0530)]
Merge branch 'nepomuk/noMoreRandomProtocols'


7 years agoHandle legacy uris in StoreResources
Vishesh Handa [Fri, 1 Jul 2011 11:02:15 +0000 (16:32 +0530)]
Handle legacy uris in StoreResources

Uris like "res:/" must be supported if they already exist in the

7 years agoAdd testStoreResources_legacy
Vishesh Handa [Fri, 1 Jul 2011 10:59:23 +0000 (16:29 +0530)]
Add testStoreResources_legacy

This tests support for urls like "res:/", which should not be allowed,
but we have to support them if they already exist.

7 years agoAdd 2 tests -
Vishesh Handa [Fri, 1 Jul 2011 09:04:26 +0000 (14:34 +0530)]
Add 2 tests -

testStoreResources_fileExists - It makes sure, that storeResources
checks that a file exists during resolution

testStoreResources_ontology - No ontology modifications should be made

Surprisingly both pass. I'm not sure if I should be happy or not.

7 years agoNow nepomukservicestub prints more informational errors
Artem Serebriyskiy [Sun, 26 Jun 2011 17:09:55 +0000 (21:09 +0400)]
Now nepomukservicestub prints more informational errors

When nepomukservicestub fails to create an instance of plugin, it
outputs error message recived from KService

7 years agoDo not stat local files if not requested.
Sebastian Trueg [Thu, 30 Jun 2011 18:23:32 +0000 (20:23 +0200)]
Do not stat local files if not requested.

7 years agoTerminate Nepomuk::clearLegacyIndexedDataForUrls properly
Vishesh Handa [Thu, 30 Jun 2011 12:39:14 +0000 (18:09 +0530)]
Terminate Nepomuk::clearLegacyIndexedDataForUrls properly

Before this patch, this function would always return true if the list
was not empty.

7 years agoLower the priorty of the "nepomukindexer"
Vishesh Handa [Thu, 30 Jun 2011 11:37:14 +0000 (17:07 +0530)]
Lower the priorty of the "nepomukindexer"

This patch makes a huge difference, and makes the system
quite useable when nepomuk is indexing files.

7 years agoDonot only test for "akonadi" in storeResources
Vishesh Handa [Thu, 30 Jun 2011 10:59:13 +0000 (16:29 +0530)]
Donot only test for "akonadi" in storeResources

Renamed testStoreResources_akonadi -> testStoreResources_kioProtocols

We now test for all kio supported protocols, apart from "file" and "nepomuk".

7 years agoMake describeResources handle all the uri states
Vishesh Handa [Thu, 30 Jun 2011 10:47:30 +0000 (16:17 +0530)]
Make describeResources handle all the uri states

This fixes test_describeResources()

7 years agoFix some tests
Vishesh Handa [Thu, 30 Jun 2011 10:38:26 +0000 (16:08 +0530)]
Fix some tests

Some of the tests use uris like "res:/A", which are not acceptable unless
they already exist in the Nepomuk repository. Use uris of the form

Unlike the last commit which changed this, this patch does not use the
"nepomuk" scheme everywhere as in some places, it is first directly added in
the model, and then the dms function are used.

"class:/type*" have been added as RDFS:Classes

7 years agoFix resolveUrl
Vishesh Handa [Thu, 30 Jun 2011 10:37:19 +0000 (16:07 +0530)]
Fix resolveUrl

UnknownUris should be resolved if the uri already exists in the Nepomuk
repository. This is useful for legacy data.

7 years agoDMS optimization round 2:
Sebastian Trueg [Wed, 29 Jun 2011 15:08:17 +0000 (17:08 +0200)]
DMS optimization round 2:

Much faster query for the metadata graphs to be deleted.

7 years agoDMS optimization round 1:
Sebastian Trueg [Wed, 29 Jun 2011 15:06:17 +0000 (17:06 +0200)]
DMS optimization round 1:

Query optimization:
* Do not use an extra count query for the metadata props. Just use the
* Split the main query in two which allows for way more optimization in
  Virtuoso resulting in much faster query execution.

7 years agoAdded another unit test.
Sebastian Trueg [Wed, 29 Jun 2011 14:59:43 +0000 (16:59 +0200)]
Added another unit test.

This one checks if storeResources does gracefully handle the same
property value encoded in different ways (string and int for example).

7 years agoMerge branch 'master' of git://
Shaun Reich [Wed, 29 Jun 2011 05:42:15 +0000 (01:42 -0400)]
Merge branch 'master' of git://

7 years agoAdd some D-BUS debug methods for KUiServer.
Shaun Reich [Wed, 29 Jun 2011 05:39:46 +0000 (01:39 -0400)]
Add some D-BUS debug methods for KUiServer.

Finally off my TODO list. But this makes it easier to diagnose some
issues. Otherwise it's just blind fire really. This way you know if
kuiserver has some services registered, as well as which JobViews at
which objectPath(s).

7 years agoRevert "Fix loads of tests"
Vishesh Handa [Wed, 29 Jun 2011 04:48:18 +0000 (10:18 +0530)]
Revert "Fix loads of tests"

This reverts commit 458bcbbffbc8fe00737bfa77969492b04d3edd1b.

7 years agoMake global shortcuts involving shift+tab work
Simon Persson [Wed, 29 Jun 2011 03:10:19 +0000 (11:10 +0800)]
Make global shortcuts involving shift+tab work

When entering (ctrl,alt,meta)+shift+tab as a new shortcut
KKeySequenceWidget gets "backtab" as the key pressed but records "tab".
kglobalaccel gets the keypress event as "backtab" which is correct but
only triggers shortcuts with "backtab" in them and such shortcuts are
not possible for the user to enter.

This commit makes kglobalaccel look for both tab and backtab when
triggered, same as Qt does. By triggering on both it also avoids
breaking peoples alt+shift+tab on upgrade.

BUG: 174142
BUG: 258467
BUG: 261296
BUG: 274006
REVIEW: 101523

7 years agoFix global shortcuts that include symbols produced with shift key
Simon Persson [Wed, 29 Jun 2011 02:56:03 +0000 (10:56 +0800)]
Fix global shortcuts that include symbols produced with shift key

KKeySequenceWidget (used to enter shortcuts) removes shift from the
recorded shortcut if the symbol produced from that key is different
when shift is used. kglobalaccel needs to include shift in the grab in
order to be triggered on this class of shortcuts, and then in the
keypress event handler it also needs to strip the shift again before
checking which shortcut was just triggered.
BUG: 179504
BUG: 197548
BUG: 215030
REVIEW: 101520

7 years agoAdded two more test methods to the resource watcher test.
Sebastian Trueg [Tue, 28 Jun 2011 16:01:30 +0000 (18:01 +0200)]
Added two more test methods to the resource watcher test.

7 years agoFixed nao:prefLabel and nao:description ranges in createResource.
Sebastian Trueg [Tue, 28 Jun 2011 15:55:46 +0000 (17:55 +0200)]
Fixed nao:prefLabel and nao:description ranges in createResource.

7 years agoAdded unit tests for createResource.
Sebastian Trueg [Tue, 28 Jun 2011 15:48:55 +0000 (17:48 +0200)]
Added unit tests for createResource.

This lead to the discovery of a bug: the types parameter was not
verified, ie. unknown types were allowed.

7 years agokio_sftp: Fix keyboard interactive prompt.
Andreas Schneider [Mon, 27 Jun 2011 17:20:37 +0000 (19:20 +0200)]
kio_sftp: Fix keyboard interactive prompt.

7 years agoAdd testStoreResources_akonadi
Vishesh Handa [Mon, 27 Jun 2011 13:46:02 +0000 (19:16 +0530)]
Add testStoreResources_akonadi

This is a very simple test that checks if the "akonadi:item=5" url,
is stored as the nie:url, and not the resource uri.

It passes :)

7 years agofix testStoreResources_itemUris
Vishesh Handa [Mon, 27 Jun 2011 13:21:12 +0000 (18:51 +0530)]
fix testStoreResources_itemUris

Uris of the form 'testuri:?item=' should not be allowed, as
'testUri' is a not a recognized protocol.