office-gobmx/framework/source/jobs/job.cxx
Noel Grandin 4f5f24a4a3 loplugin:ostr in framework
Change-Id: I2fdc32aa5de6a1315fe69997f9b531259aa78605
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/167339
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
2024-05-08 17:18:51 +02:00

857 lines
29 KiB
C++

/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
/*
* This file is part of the LibreOffice project.
*
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/.
*
* This file incorporates work covered by the following license notice:
*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed
* with this work for additional information regarding copyright
* ownership. The ASF licenses this file to you under the Apache
* License, Version 2.0 (the "License"); you may not use this file
* except in compliance with the License. You may obtain a copy of
* the License at http://www.apache.org/licenses/LICENSE-2.0 .
*/
#include <jobs/job.hxx>
#include <jobs/jobresult.hxx>
#include <com/sun/star/frame/Desktop.hpp>
#include <com/sun/star/frame/TerminationVetoException.hpp>
#include <com/sun/star/task/XJob.hpp>
#include <com/sun/star/task/XAsyncJob.hpp>
#include <com/sun/star/util/CloseVetoException.hpp>
#include <com/sun/star/util/XCloseBroadcaster.hpp>
#include <com/sun/star/util/XCloseable.hpp>
#include <com/sun/star/lang/DisposedException.hpp>
#include <comphelper/sequence.hxx>
#include <sal/log.hxx>
#include <comphelper/diagnose_ex.hxx>
#include <utility>
#include <vcl/svapp.hxx>
namespace framework{
/**
@short standard ctor
@descr It initialize this new instance. But it set some generic parameters here only.
Specialized information (e.g. the alias or service name ofthis job) will be set
later using the method setJobData().
@param xContext
reference to the uno service manager
@param xFrame
reference to the frame, in which environment we run
(May be null!)
*/
Job::Job( /*IN*/ const css::uno::Reference< css::uno::XComponentContext >& xContext ,
/*IN*/ css::uno::Reference< css::frame::XFrame > xFrame )
: m_aJobCfg (xContext )
, m_xContext (xContext )
, m_xFrame (std::move(xFrame ))
, m_bListenOnDesktop (false )
, m_bListenOnFrame (false )
, m_bListenOnModel (false )
, m_bPendingCloseFrame (false )
, m_bPendingCloseModel (false )
, m_eRunState (E_NEW )
{
}
/**
@short standard ctor
@descr It initialize this new instance. But it set some generic parameters here only.
Specialized information (e.g. the alias or service name ofthis job) will be set
later using the method setJobData().
@param xContext
reference to the uno service manager
@param xModel
reference to the model, in which environment we run
(May be null!)
*/
Job::Job( /*IN*/ const css::uno::Reference< css::uno::XComponentContext >& xContext ,
/*IN*/ css::uno::Reference< css::frame::XModel > xModel )
: m_aJobCfg (xContext )
, m_xContext (xContext )
, m_xModel (std::move(xModel ))
, m_bListenOnDesktop (false )
, m_bListenOnFrame (false )
, m_bListenOnModel (false )
, m_bPendingCloseFrame (false )
, m_bPendingCloseModel (false )
, m_eRunState (E_NEW )
{
}
/**
@short superfluous!
@descr Releasing of memory and reference must be done inside die() call.
Otherwise it's a bug.
*/
Job::~Job()
{
}
/**
@short set (or delete) a listener for sending dispatch result events
@descr Because this object is used in a wrapped mode ... the original listener
for such events can't be registered here directly. Because the
listener expect to get the original object given as source of the event.
That's why we get this source here too, to fake(!) it at sending time!
@param xListener
the original listener for dispatch result events
@param xSourceFake
our user, which got the registration request for this listener
*/
void Job::setDispatchResultFake( /*IN*/ const css::uno::Reference< css::frame::XDispatchResultListener >& xListener ,
/*IN*/ const css::uno::Reference< css::uno::XInterface >& xSourceFake )
{
SolarMutexGuard g;
// reject dangerous calls
if (m_eRunState != E_NEW)
{
SAL_INFO("fwk", "Job::setJobData(): job may still running or already finished");
return;
}
m_xResultListener = xListener;
m_xResultSourceFake = xSourceFake;
}
void Job::setJobData( const JobData& aData )
{
SolarMutexGuard g;
// reject dangerous calls
if (m_eRunState != E_NEW)
{
SAL_INFO("fwk", "Job::setJobData(): job may still running or already finished");
return;
}
m_aJobCfg = aData;
}
/**
@short runs the job
@descr It doesn't matter, if the job is an asynchronous or
synchronous one. This method returns only if it was finished
or cancelled.
@param lDynamicArgs
optional arguments for job execution
In case the represented job is a configured one (which uses static
arguments too) all information will be merged!
*/
void Job::execute( /*IN*/ const css::uno::Sequence< css::beans::NamedValue >& lDynamicArgs )
{
/* SAFE { */
class SolarMutexAntiGuard {
SolarMutexResettableGuard & m_rGuard;
public:
SolarMutexAntiGuard(SolarMutexResettableGuard & rGuard) : m_rGuard(rGuard)
{
m_rGuard.clear();
}
~SolarMutexAntiGuard()
{
m_rGuard.reset();
}
};
SolarMutexResettableGuard aWriteLock;
// reject dangerous calls
if (m_eRunState != E_NEW)
{
SAL_INFO("fwk", "Job::execute(): job may still running or already finished");
return;
}
// create the environment and mark this job as running ...
m_eRunState = E_RUNNING;
impl_startListening();
css::uno::Reference< css::task::XAsyncJob > xAJob;
css::uno::Reference< css::task::XJob > xSJob;
css::uno::Sequence< css::beans::NamedValue > lJobArgs = impl_generateJobArgs(lDynamicArgs);
// It's necessary to hold us self alive!
// Otherwise we might die by ref count ...
css::uno::Reference< css::task::XJobListener > xThis(this);
try
{
// create the job
// We must check for the supported interface on demand!
// But we prefer the synchronous one ...
m_xJob = m_xContext->getServiceManager()->createInstanceWithContext(m_aJobCfg.getService(), m_xContext);
xSJob.set(m_xJob, css::uno::UNO_QUERY);
if (!xSJob.is())
xAJob.set(m_xJob, css::uno::UNO_QUERY);
// execute it asynchronous
if (xAJob.is())
{
m_aAsyncWait.reset();
SolarMutexAntiGuard const ag(aWriteLock);
/* } SAFE */
xAJob->executeAsync(lJobArgs, xThis);
// wait for finishing this job - so this method
// does the same for synchronous and asynchronous jobs!
m_aAsyncWait.wait();
/* SAFE { */
// Note: Result handling was already done inside the callback!
}
// execute it synchron
else if (xSJob.is())
{
css::uno::Any aResult;
{
SolarMutexAntiGuard const ag(aWriteLock);
/* } SAFE */
aResult = xSJob->execute(lJobArgs);
}
/* SAFE { */
impl_reactForJobResult(aResult);
}
}
#if OSL_DEBUG_LEVEL > 0
catch(const css::uno::Exception&)
{
TOOLS_INFO_EXCEPTION("fwk", "Job::execute(): Got exception during job execution");
}
#else
catch(const css::uno::Exception&)
{}
#endif
// deinitialize the environment and mark this job as finished...
// but don't overwrite any information about STOPPED or might DISPOSED jobs!
impl_stopListening();
if (m_eRunState == E_RUNNING)
m_eRunState = E_STOPPED_OR_FINISHED;
// If we got a close request from our frame or model...
// but we disagreed with that by throwing a veto exception...
// and got the ownership...
// we have to close the resource frame or model now -
// and to disable ourself!
if (m_bPendingCloseFrame)
{
m_bPendingCloseFrame = false;
css::uno::Reference< css::util::XCloseable > xClose(m_xFrame, css::uno::UNO_QUERY);
if (xClose.is())
{
try
{
xClose->close(true);
}
catch(const css::util::CloseVetoException&) {}
}
}
if (m_bPendingCloseModel)
{
m_bPendingCloseModel = false;
css::uno::Reference< css::util::XCloseable > xClose(m_xModel, css::uno::UNO_QUERY);
if (xClose.is())
{
try
{
xClose->close(true);
}
catch(const css::util::CloseVetoException&) {}
}
}
aWriteLock.clear();
/* SAFE { */
// release this instance ...
die();
}
/**
@short kill this job
@descr It doesn't matter if this request is called from inside or
from outside. We release our internal structures and stop
every activity. After doing so - this instance will not be
usable any longer! Of course we try to handle further requests
carefully. Maybe someone else holds a reference to us ...
*/
void Job::die()
{
SolarMutexGuard g;
impl_stopListening();
if (m_eRunState != E_DISPOSED)
{
try
{
css::uno::Reference< css::lang::XComponent > xDispose(m_xJob, css::uno::UNO_QUERY);
if (xDispose.is())
{
xDispose->dispose();
m_eRunState = E_DISPOSED;
}
}
catch(const css::lang::DisposedException&)
{
m_eRunState = E_DISPOSED;
}
}
m_xJob.clear();
m_xFrame.clear();
m_xModel.clear();
m_xDesktop.clear();
m_xResultListener.clear();
m_xResultSourceFake.clear();
m_bPendingCloseFrame = false;
m_bPendingCloseModel = false;
}
/**
@short generates list of arguments for job execute
@descr There exist a set of information, which can be needed by a job.
a) it's static configuration data (Equals for all jobs. )
b) it's specific configuration data (Different for every job.)
c) some environment values (e.g. the frame, for which this job was started)
d) any other dynamic data (e.g. parameters of a dispatch() request)
We collect all this information and generate one list which include all others.
@param lDynamicArgs
list of dynamic arguments (given by a corresponding dispatch() call)
Can be empty too.
@return A list which includes all mentioned sub lists.
*/
css::uno::Sequence< css::beans::NamedValue > Job::impl_generateJobArgs( /*IN*/ const css::uno::Sequence< css::beans::NamedValue >& lDynamicArgs )
{
css::uno::Sequence< css::beans::NamedValue > lAllArgs;
/* SAFE { */
SolarMutexClearableGuard aReadLock;
// the real structure of the returned list depends from the environment of this job!
JobData::EMode eMode = m_aJobCfg.getMode();
// Create list of environment variables. This list must be part of the
// returned structure every time... but some of its members are optional!
sal_Int32 nLen = 1;
if (m_xFrame.is())
++nLen;
if (m_xModel.is())
++nLen;
if (eMode==JobData::E_EVENT)
++nLen;
css::uno::Sequence< css::beans::NamedValue > lEnvArgs(nLen);
auto plEnvArgs = lEnvArgs.getArray();
plEnvArgs[0].Name = "EnvType";
plEnvArgs[0].Value <<= m_aJobCfg.getEnvironmentDescriptor();
sal_Int32 i = 0;
if (m_xFrame.is())
{
++i;
plEnvArgs[i].Name = "Frame";
plEnvArgs[i].Value <<= m_xFrame;
}
if (m_xModel.is())
{
++i;
plEnvArgs[i].Name = "Model";
plEnvArgs[i].Value <<= m_xModel;
}
if (eMode==JobData::E_EVENT)
{
++i;
plEnvArgs[i].Name = "EventName";
plEnvArgs[i].Value <<= m_aJobCfg.getEvent();
}
// get the configuration data from the job data container ... if possible
// Means: if this job has any configuration data. Note: only really
// filled lists will be set to the return structure at the end of this method.
css::uno::Sequence< css::beans::NamedValue > lConfigArgs;
std::vector< css::beans::NamedValue > lJobConfigArgs;
if (eMode==JobData::E_ALIAS || eMode==JobData::E_EVENT)
{
lConfigArgs = m_aJobCfg.getConfig();
lJobConfigArgs = m_aJobCfg.getJobConfig();
}
aReadLock.clear();
/* } SAFE */
// Add all valid (not empty) lists to the return list
if (lConfigArgs.hasElements())
{
sal_Int32 nLength = lAllArgs.getLength();
lAllArgs.realloc(nLength+1);
auto plAllArgs = lAllArgs.getArray();
plAllArgs[nLength].Name = "Config";
plAllArgs[nLength].Value <<= lConfigArgs;
}
if (!lJobConfigArgs.empty())
{
sal_Int32 nLength = lAllArgs.getLength();
lAllArgs.realloc(nLength+1);
auto plAllArgs = lAllArgs.getArray();
plAllArgs[nLength].Name = "JobConfig";
plAllArgs[nLength].Value <<= comphelper::containerToSequence(lJobConfigArgs);
}
if (lEnvArgs.hasElements())
{
sal_Int32 nLength = lAllArgs.getLength();
lAllArgs.realloc(nLength+1);
auto plAllArgs = lAllArgs.getArray();
plAllArgs[nLength].Name = "Environment";
plAllArgs[nLength].Value <<= lEnvArgs;
}
if (lDynamicArgs.hasElements())
{
sal_Int32 nLength = lAllArgs.getLength();
lAllArgs.realloc(nLength+1);
auto plAllArgs = lAllArgs.getArray();
plAllArgs[nLength].Name = "DynamicData";
plAllArgs[nLength].Value <<= lDynamicArgs;
}
return lAllArgs;
}
/**
@short analyze the given job result and change the job configuration
@descr Note: Some results can be handled only, if this job has a valid configuration!
For "not configured jobs" (means pure services) they can be ignored.
But these cases are handled by our JobData member. We can call it every time.
It does the right things automatically. E.g. if the job has no configuration ...
it does nothing during setJobConfig()!
@param aResult
the job result for analyzing
*/
void Job::impl_reactForJobResult( /*IN*/ const css::uno::Any& aResult )
{
SolarMutexGuard g;
// analyze the result set ...
JobResult aAnalyzedResult(aResult);
// some of the following operations will be supported for different environments
// or different type of jobs only.
JobData::EEnvironment eEnvironment = m_aJobCfg.getEnvironment();
// write back the job specific configuration data ...
// If the environment allow it and if this job has a configuration!
if (
(m_aJobCfg.hasConfig() ) &&
(aAnalyzedResult.existPart(JobResult::E_ARGUMENTS))
)
{
m_aJobCfg.setJobConfig(aAnalyzedResult.getArguments());
}
// disable a job for further executions.
// Note: this option is available inside the environment EXECUTOR only
if (
// (eEnvironment == JobData::E_EXECUTION ) &&
(m_aJobCfg.hasConfig() ) &&
(aAnalyzedResult.existPart(JobResult::E_DEACTIVATE))
)
{
m_aJobCfg.disableJob();
}
// notify any interested listener with the may given result state.
// Note: this option is available inside the environment DISPATCH only
if (
(eEnvironment == JobData::E_DISPATCH ) &&
(m_xResultListener.is() ) &&
(aAnalyzedResult.existPart(JobResult::E_DISPATCHRESULT))
)
{
// Attention: Because the listener expect that the original object send this event ...
// and we nor the job are the right ones ...
// our user has set itself before. So we can fake this source address!
css::frame::DispatchResultEvent aEvent = aAnalyzedResult.getDispatchResult();
aEvent.Source = m_xResultSourceFake;
m_xResultListener->dispatchFinished(aEvent);
}
}
/**
@short starts listening for office shutdown and closing of our
given target frame (if it's a valid reference)
@descr We will register ourself as terminate listener
at the global desktop instance. That will hold us
alive and additional we get the information, if the
office wish to shutdown. If then an internal job
is running we will have the chance to suppress that
by throwing a veto exception. If our internal wrapped
job finished his work, we can release this listener
connection.
Further we are listener for closing of the (possible valid)
given frame. We must be sure, that this resource won't be gone
if our internal job is still running.
*/
void Job::impl_startListening()
{
SolarMutexGuard g;
// listening for office shutdown
if (!m_xDesktop.is() && !m_bListenOnDesktop)
{
try
{
m_xDesktop = css::frame::Desktop::create( m_xContext );
css::uno::Reference< css::frame::XTerminateListener > xThis(this);
m_xDesktop->addTerminateListener(xThis);
m_bListenOnDesktop = true;
}
catch(const css::uno::Exception&)
{
m_xDesktop.clear();
}
}
// listening for frame closing
if (m_xFrame.is() && !m_bListenOnFrame)
{
try
{
css::uno::Reference< css::util::XCloseBroadcaster > xCloseable(m_xFrame , css::uno::UNO_QUERY);
css::uno::Reference< css::util::XCloseListener > xThis(this);
if (xCloseable.is())
{
xCloseable->addCloseListener(xThis);
m_bListenOnFrame = true;
}
}
catch(const css::uno::Exception&)
{
m_bListenOnFrame = false;
}
}
// listening for model closing
if (!m_xModel.is() || m_bListenOnModel)
return;
try
{
css::uno::Reference< css::util::XCloseBroadcaster > xCloseable(m_xModel , css::uno::UNO_QUERY);
css::uno::Reference< css::util::XCloseListener > xThis(this);
if (xCloseable.is())
{
xCloseable->addCloseListener(xThis);
m_bListenOnModel = true;
}
}
catch(const css::uno::Exception&)
{
m_bListenOnModel = false;
}
}
/**
@short release listener connection for office shutdown
@descr see description of impl_startListening()
*/
void Job::impl_stopListening()
{
SolarMutexGuard g;
// stop listening for office shutdown
if (m_xDesktop.is() && m_bListenOnDesktop)
{
try
{
css::uno::Reference< css::frame::XTerminateListener > xThis(this);
m_xDesktop->removeTerminateListener(xThis);
m_xDesktop.clear();
m_bListenOnDesktop = false;
}
catch(const css::uno::Exception&)
{
}
}
// stop listening for frame closing
if (m_xFrame.is() && m_bListenOnFrame)
{
try
{
css::uno::Reference< css::util::XCloseBroadcaster > xCloseable(m_xFrame , css::uno::UNO_QUERY);
css::uno::Reference< css::util::XCloseListener > xThis(this);
if (xCloseable.is())
{
xCloseable->removeCloseListener(xThis);
m_bListenOnFrame = false;
}
}
catch(const css::uno::Exception&)
{
}
}
// stop listening for model closing
if (!(m_xModel.is() && m_bListenOnModel))
return;
try
{
css::uno::Reference< css::util::XCloseBroadcaster > xCloseable(m_xModel , css::uno::UNO_QUERY);
css::uno::Reference< css::util::XCloseListener > xThis(this);
if (xCloseable.is())
{
xCloseable->removeCloseListener(xThis);
m_bListenOnModel = false;
}
}
catch(const css::uno::Exception&)
{
}
}
/**
@short callback from any asynchronous executed job
@descr Our execute() method waits for this callback.
We have to react for the possible results here,
to kill the running job and disable the blocked condition
so execute() can be finished too.
@param xJob
the job, which was running and inform us now
@param aResult
its results
*/
void SAL_CALL Job::jobFinished( /*IN*/ const css::uno::Reference< css::task::XAsyncJob >& xJob ,
/*IN*/ const css::uno::Any& aResult )
{
SolarMutexGuard g;
// It's necessary to check this.
// May this job was cancelled by any other reason
// some milliseconds before. :-)
if (m_xJob.is() && m_xJob==xJob)
{
// react for his results
// (means enable/disable it for further requests
// or save arguments or notify listener ...)
impl_reactForJobResult(aResult);
// Let the job die!
m_xJob.clear();
}
// And let the start method "execute()" finishing it's job.
// But do it every time. So any outside blocking code can finish
// his work too.
m_aAsyncWait.set();
}
/**
@short prevent internal wrapped job against office termination
@descr This event is broadcasted by the desktop instance and ask for an office termination.
If the internal wrapped job is still in progress, we disagree with that by throwing the
right veto exception. If not - we agree. But then we must be aware, that another event
notifyTermination() can follow. Then we have no chance to do the same. Then we have to
accept that and stop our work instandly.
@param aEvent
describes the broadcaster and must be the desktop instance
@throw TerminateVetoException
if our internal wrapped job is still running.
*/
void SAL_CALL Job::queryTermination( /*IN*/ const css::lang::EventObject& )
{
SolarMutexGuard g;
// Otherwise try to close() it
css::uno::Reference< css::util::XCloseable > xClose(m_xJob, css::uno::UNO_QUERY);
if (xClose.is())
{
try
{
xClose->close(false);
m_eRunState = E_STOPPED_OR_FINISHED;
}
catch(const css::util::CloseVetoException&) {}
}
if (m_eRunState != E_STOPPED_OR_FINISHED)
{
css::uno::Reference< css::uno::XInterface > xThis(static_cast< ::cppu::OWeakObject* >(this), css::uno::UNO_QUERY);
throw css::frame::TerminationVetoException(u"job still in progress"_ustr, xThis);
}
}
/**
@short inform us about office termination
@descr Instead of the method queryTermination(), here is no chance to disagree with that.
We have to accept it and cancel all current processes inside.
It can occur only, if job was not already started if queryTermination() was called here.
Then we had not thrown a veto exception. But now we must agree with this situation and break
all our internal processes. It's not a good idea to mark this instance as non startable any longer
inside queryTermination() if no job was running too. Because that would disable this job and may
the office does not really shutdown, because another listener has thrown the suitable exception.
@param aEvent
describes the broadcaster and must be the desktop instance
*/
void SAL_CALL Job::notifyTermination( /*IN*/ const css::lang::EventObject& )
{
die();
// Do nothing else here. Our internal resources was released ...
}
/**
@short prevent internal wrapped job against frame closing
@descr This event is broadcasted by the frame instance and ask for closing.
If the internal wrapped job is still in progress, we disagree with that by throwing the
right veto exception. If not - we agree. But then we must be aware, that another event
notifyClosing() can follow. Then we have no chance to do the same. Then we have to
accept that and stop our work instandly.
@param aEvent
describes the broadcaster and must be the frame instance
@param bGetsOwnership
If it's set to <sal_True> and we throw the right veto exception, we have to close this frame later
if our internal processes will be finished. If it's set to <FALSE/> we can ignore it.
@throw CloseVetoException
if our internal wrapped job is still running.
*/
void SAL_CALL Job::queryClosing( const css::lang::EventObject& aEvent ,
sal_Bool bGetsOwnership )
{
SolarMutexGuard g;
// do nothing, if no internal job is still running ...
// The frame or model can be closed then successfully.
if (m_eRunState != E_RUNNING)
return;
// try close() first at the job.
// The job can agree or disagree with this request.
css::uno::Reference< css::util::XCloseable > xClose(m_xJob, css::uno::UNO_QUERY);
if (xClose.is())
{
xClose->close(bGetsOwnership);
// Here we can say: "this job was stopped successfully". Because
// no veto exception was thrown!
m_eRunState = E_STOPPED_OR_FINISHED;
return;
}
// try dispose() then
// Here the job has no chance for a veto.
// But we must be aware of an "already disposed exception"...
try
{
css::uno::Reference< css::lang::XComponent > xDispose(m_xJob, css::uno::UNO_QUERY);
if (xDispose.is())
{
xDispose->dispose();
m_eRunState = E_DISPOSED;
}
}
catch(const css::lang::DisposedException&)
{
// the job was already disposed by any other mechanism !?
// But it's not interesting for us. For us this job is stopped now.
m_eRunState = E_DISPOSED;
}
if (m_eRunState != E_DISPOSED)
{
// analyze event source - to find out, which resource called queryClosing() at this
// job wrapper. We must bind a "pending close" request to this resource.
// Closing of the corresponding resource will be done if our internal job finish it's work.
m_bPendingCloseFrame = (m_xFrame.is() && aEvent.Source == m_xFrame);
m_bPendingCloseModel = (m_xModel.is() && aEvent.Source == m_xModel);
// throw suitable veto exception - because the internal job could not be cancelled.
css::uno::Reference< css::uno::XInterface > xThis(static_cast< ::cppu::OWeakObject* >(this), css::uno::UNO_QUERY);
throw css::util::CloseVetoException(u"job still in progress"_ustr, xThis);
}
// No veto ...
// But don't call die() here or free our internal member.
// This must be done inside notifyClosing() only. Otherwise the
// might stopped job has no chance to return its results or
// call us back. We must give him the chance to finish it's work successfully.
}
/**
@short inform us about frame closing
@descr Instead of the method queryClosing(), here is no chance to disagree with that.
We have to accept it and cancel all current processes inside.
@param aEvent
describes the broadcaster and must be the frame or model instance we know
*/
void SAL_CALL Job::notifyClosing( const css::lang::EventObject& )
{
die();
// Do nothing else here. Our internal resources was released ...
}
/**
@short shouldn't be called normally
@descr But it doesn't matter, who called it. We have to kill our internal
running processes hardly.
@param aEvent
describe the broadcaster
*/
void SAL_CALL Job::disposing( const css::lang::EventObject& aEvent )
{
/* SAFE { */
{
SolarMutexGuard aWriteLock;
if (m_xDesktop.is() && aEvent.Source == m_xDesktop)
{
m_xDesktop.clear();
m_bListenOnDesktop = false;
}
else if (m_xFrame.is() && aEvent.Source == m_xFrame)
{
m_xFrame.clear();
m_bListenOnFrame = false;
}
else if (m_xModel.is() && aEvent.Source == m_xModel)
{
m_xModel.clear();
m_bListenOnModel = false;
}
}
/* } SAFE */
die();
// Do nothing else here. Our internal resources was released ...
}
} // namespace framework
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */