summaryrefslogtreecommitdiffstats
path: root/src/corelib/concurrent
Commit message (Collapse)AuthorAgeFilesLines
* Build fix for mingwThierry Bastian2009-07-141-7/+1
| | | | | we should include qt_windows.h and not windows.h because we have to define WINVER to 0x500.
* doc: Use \inheaderfile to show which include file to include.Martin Smith2009-07-131-1/+1
| | | | | | | Letting qdoc output it automatically produces the wrong header file name. Task-number: 217268 210171 220664
* Doc: Explicitly declare the module for each namespace to prevent strangeDavid Boddie2009-07-131-0/+1
| | | | | | | omissions when qdoc is used in certain Windows environments. Task-number: 256415 Reviewed-by: Trust Me
* De-export the ThreadEngineBarrier class and de-inline the call to acquire.Morten Sørvig2009-06-262-2/+13
| | | | | | | | | | | | | | | (ThreadEngineBarrier was implemented in the header as ThreadEngineSemaphore in Qt 4.4 and Qt 4.5, and exported for a brief time in the Qt master branch. ) Unfortunately the inline call to acquire means that applications compiled with Qt 4.4. and Qt 4.5 will contain a copy of the old implementation. This is BC as long as the memory layout of the ThreadEngineBarrier class stays the same, which means that we keep the QMutex member variable. The old version of acquire will no longer be thread-safe since the new version doesn't lock the mutex, but this is fortunately not a problem; the only place it's called is in startAsynchronously() before the worker threads are started.
* Add Q_CORE_EXPORT to ThreadEngineBarrierRhys Weatherley2009-06-241-1/+1
| | | | | | | | The ThreadEngine template uses functions from ThreadEngineBarrier, so ThreadEngineBarrier must be exported to other libraries/programs that instantiate the template. Reviewed-by: trustme
* Rename ThreadEngineSemaphore -> ThreadEngineBarrierMorten Sørvig2009-06-242-85/+93
| | | | Also move the implementation to the .cpp file.
* Improve QtConcurrent scalability.Morten Sørvig2009-06-241-24/+62
| | | | | Reduce lock contention QtConcurrent by swiching the ThreadEngineSemaphore class over to a QAtomic-based implementation.
* Improve QThreadPool scalability.Morten Sørvig2009-06-241-2/+8
| | | | | | Reduce lock contention: - Skip locking in the accessor functions (where we can) - exit early in tryStart (before locking the mutex) when the threadpool is running at max capacity.
* Merge license header changes from 4.5Volker Hilsheimer2009-06-1637-74/+74
|\
| * Update license headers as requested by the marketing department.Jason McDonald2009-06-1637-74/+74
| | | | | | | | Reviewed-by: Trust Me
* | Removed some export to symbols that don't need itThierry Bastian2009-05-251-2/+2
|/ | | | ...hopefully
* Use isNull() for strings instead of comparing against QString().Jesper Thomschütz2009-05-141-1/+1
| | | | | | foo == QString() should be foo.isNull(). Fixes 7 warnings in the Norwegian Breakfast Network Reviewed-by: Samuel
* Fix memory leak in QFutureMorten Sørvig2009-05-051-0/+2
| | | | | | | | | The held results were not cleared by QFutureInterface::operator=(QFutureInterface) Add call to resultStore().clear(), similar to the destructor. This needs to be done in the header since we know the template type here. Task-number: 252208 Reviewed-by: brad
* Reduce the number of calls to QThread::idealThreadCount() in QtConcurrent.Morten Sørvig2009-03-261-4/+6
| | | | | | | | This results in a syscall and is very slow. Make the call once and cache the value. Task-number: 244718 Reviewed-by: TrustMe
* Long live Qt 4.5!Lars Knoll2009-03-2338-0/+11514