data: indexer processes

* TaskBuildIndex starts seperate indexer processes and communicates via shared memory
* indexer processes get restarted if they fail
* indexer processes are killed when the app closes or crashes
* added utility class SharedMemory to allocate and access shared memory, providing basic data structures
* added SharedMemoryGarbageCollector for keeping track of running instances and cleaning up shared memory in case of a crash
* show errors for source files that crashed during indexing
* added basic exception handling to task scheduling
* added option to turn off processes and use threads in preferences, but still use shared memory

fortune cookie message = Good news from someone dear is coming soon.
This commit is contained in:
Eberhard Graether
2017-05-04 15:50:59 +02:00
parent 9beea94fd8
commit 706119ebcc
109 changed files with 3351 additions and 1261 deletions
+188 -57
View File
@@ -1,16 +1,30 @@
#include "data/indexer/TaskBuildIndex.h"
#include "utility/AppPath.h"
#include "utility/file/FileRegister.h"
#include "utility/file/FileRegisterStateData.h"
#include "utility/logging/FileLogger.h"
#include "utility/scheduling/Blackboard.h"
#include "utility/UserPaths.h"
#include "utility/utilityApp.h"
#include "Application.h"
#include "component/view/DialogView.h"
#include "data/indexer/IndexerFactory.h"
#include "data/indexer/IndexerCommandList.h"
#include "data/indexer/IndexerComposite.h"
#include "data/indexer/interprocess/InterprocessIndexer.h"
#include "data/StorageProvider.h"
#include "component/view/DialogView.h"
#include "utility/file/FileRegister.h"
#include "utility/file/FileRegisterStateData.h"
#include "utility/scheduling/Blackboard.h"
#include "Application.h"
#include "settings/ApplicationSettings.h"
#if _WIN32
const std::string TaskBuildIndex::s_processName("sourcetrail_indexer.exe");
#else
const std::string TaskBuildIndex::s_processName("sourcetrail_indexer");
#endif
TaskBuildIndex::TaskBuildIndex(
unsigned int processCount,
std::shared_ptr<IndexerCommandList> indexerCommandList,
std::shared_ptr<StorageProvider> storageProvider,
std::shared_ptr<FileRegisterStateData> fileRegisterStateData
@@ -18,89 +32,206 @@ TaskBuildIndex::TaskBuildIndex(
: m_indexerCommandList(indexerCommandList)
, m_storageProvider(storageProvider)
, m_fileRegisterStateData(fileRegisterStateData)
, m_interprocessIndexerCommandManager(Application::getUUID(), 0, true)
, m_interprocessIndexingStatusManager(Application::getUUID(), 0, true)
, m_processCount(processCount)
, m_interrupted(false)
, m_lastCommandCount(0)
{
m_indexer = IndexerFactory::getInstance()->createCompositeIndexerForAllRegisteredModules();
}
void TaskBuildIndex::doEnter(std::shared_ptr<Blackboard> blackboard)
{
std::lock_guard<std::mutex> lock(blackboard->getMutex());
int indexerCount = 0;
if (blackboard->get("indexer_count", indexerCount))
{
indexerCount++;
blackboard->set("indexer_count", indexerCount);
std::lock_guard<std::mutex> lock(blackboard->getMutex());
blackboard->set("indexer_count", (int)m_processCount);
}
// move indexer commands to shared memory
m_lastCommandCount = m_indexerCommandList->size();
m_interprocessIndexerCommandManager.setIndexerCommands(m_indexerCommandList->getAllCommands());
std::string logFilePath;
Logger* logger = LogManager::getInstance()->getLoggerByType("FileLogger");
if (logger)
{
logFilePath = dynamic_cast<FileLogger*>(logger)->getLogFilePath().str();
}
bool multiProcess = ApplicationSettings::getInstance()->getMultiProcessIndexingEnabled();
// start indexer processes
for (unsigned int i = 0; i < m_processCount; i++)
{
const int processId = i + 1; // 0 remains reserved for the main process
m_interprocessIntermediateStorageManagers.push_back(
std::make_shared<InterprocessIntermediateStorageManager>(Application::getUUID(), processId, true)
);
if (multiProcess)
{
m_processThreads.push_back(new std::thread(&TaskBuildIndex::runIndexerProcess, this, processId, logFilePath));
}
else
{
m_processThreads.push_back(new std::thread(&TaskBuildIndex::runIndexerThread, this, processId));
}
}
}
Task::TaskState TaskBuildIndex::doUpdate(std::shared_ptr<Blackboard> blackboard)
{
std::shared_ptr<IndexerCommand> indexerCommand = m_indexerCommandList->consumeCommand();
size_t commandCount = m_interprocessIndexerCommandManager.indexerCommandCount();
if (commandCount != m_lastCommandCount)
{
updateIndexingDialog(blackboard, m_interprocessIndexingStatusManager.getCurrentlyIndexedSourceFilePath());
m_lastCommandCount = commandCount;
}
if (!indexerCommand)
if (commandCount == 0)
{
return STATE_FAILURE;
}
else
else if (m_interrupted)
{
{
std::lock_guard<std::mutex> lock(blackboard->getMutex());
std::lock_guard<std::mutex> lock(blackboard->getMutex());
blackboard->set("interrupted_indexing", true);
int sourceFileCount = 0;
blackboard->get("source_file_count", sourceFileCount);
int indexedSourceFileCount = 0;
blackboard->get("indexed_source_file_count", indexedSourceFileCount);
if (std::shared_ptr<DialogView> dialogView = Application::getInstance()->getDialogView())
{
dialogView->updateIndexingDialog(
indexedSourceFileCount, sourceFileCount, indexerCommand->getSourceFilePath().str()
);
}
}
// file register only copies the DileRegisterStateData
std::shared_ptr<FileRegister> fileRegister = std::make_shared<FileRegister>(
*(m_fileRegisterStateData.get()), indexerCommand->getIndexedPaths(), indexerCommand->getExcludedPath()
);
std::shared_ptr<IntermediateStorage> storage = m_indexer->index(indexerCommand, fileRegister);
if (storage)
{
// only write back the changes made to FileRegisterStateData if the indexer actually succeeded
m_fileRegisterStateData->inject(fileRegister->getStateData());
m_storageProvider->insert(storage);
std::lock_guard<std::mutex> lock(blackboard->getMutex());
int indexedSourceFileCount = 0;
blackboard->get("indexed_source_file_count", indexedSourceFileCount);
blackboard->set("indexed_source_file_count", indexedSourceFileCount + 1);
}
// clear indexer commands, this causes the indexer processes to return when finished with respective current indexer commands
m_interprocessIndexerCommandManager.clearIndexerCommands();
return STATE_FAILURE;
}
return (m_indexer->interrupted() ? STATE_FAILURE : STATE_SUCCESS);
fetchIntermediateStorages(blackboard);
const int SLEEP_TIME_MS = 100;
std::this_thread::sleep_for(std::chrono::milliseconds(SLEEP_TIME_MS));
return STATE_RUNNING;
}
void TaskBuildIndex::doExit(std::shared_ptr<Blackboard> blackboard)
{
std::lock_guard<std::mutex> lock(blackboard->getMutex());
int indexerCount = 0;
if (blackboard->get("indexer_count", indexerCount))
for (auto processThread : m_processThreads)
{
indexerCount--;
blackboard->set("indexer_count", indexerCount);
processThread->join();
delete processThread;
}
m_processThreads.clear();
fetchIntermediateStorages(blackboard);
std::vector<FilePath> crashedFiles = m_interprocessIndexingStatusManager.getCrashedSourceFilePaths();
if (crashedFiles.size())
{
std::shared_ptr<IntermediateStorage> is = std::make_shared<IntermediateStorage>();
for (auto path : crashedFiles)
{
is->addError("The translation unit threw an exception during indexing. Please check if the source file "
"conforms to the specified language standard and all necessary options are defined within your project "
"setup.", path, 1, 1, true, true);
LOG_INFO_STREAM(<< "crashed translation unit: " << path.str());
}
m_storageProvider->insert(is);
}
std::lock_guard<std::mutex> lock(blackboard->getMutex());
blackboard->set("indexer_count", 0);
}
void TaskBuildIndex::doReset(std::shared_ptr<Blackboard> blackboard)
{
}
void TaskBuildIndex::terminate()
{
m_interrupted = true;
utility::killRunningProcesses();
}
void TaskBuildIndex::handleMessage(MessageInterruptTasks* message)
{
m_indexer->interrupt();
m_interrupted = true;
}
void TaskBuildIndex::runIndexerProcess(int processId, const std::string& logFilePath)
{
FilePath indexerProcessPath(AppPath::getAppPath() + s_processName);
if (!indexerProcessPath.exists())
{
m_interrupted = true;
LOG_ERROR("Cannot start indexer process because executable is missing at \"" + indexerProcessPath.str() + "\"");
return;
}
std::string command = indexerProcessPath.str();
command += " " + std::to_string(processId);
command += " " + Application::getUUID();
command += " \"" + AppPath::getAppPath() + "\"";
command += " \"" + UserPaths::getUserDataPath().str() + "\"";
if (logFilePath.size())
{
command += " \"" + logFilePath + "\"";
}
int result = 1;
while (result != 0 && !m_interrupted)
{
result = utility::executeProcessAndGetExitCode(command.c_str(), "", -1);
LOG_INFO_STREAM(<< "Indexer process " << processId << " returned with " + std::to_string(result));
}
}
void TaskBuildIndex::runIndexerThread(int processId)
{
InterprocessIndexer indexer(Application::getUUID(), processId);
indexer.work();
}
void TaskBuildIndex::fetchIntermediateStorages(std::shared_ptr<Blackboard> blackboard)
{
int newlyIndexedCount = 0;
for (std::shared_ptr<InterprocessIntermediateStorageManager> storageManager: m_interprocessIntermediateStorageManagers)
{
while (int storageCount = storageManager->getIntermediateStorageCount())
{
LOG_INFO_STREAM(<< storageManager->getProcessId() << " - storage count: " << storageCount);
m_storageProvider->insert(storageManager->popIntermediateStorage());
++newlyIndexedCount;
updateIndexingDialog(blackboard, m_interprocessIndexingStatusManager.getCurrentlyIndexedSourceFilePath());
}
}
if (newlyIndexedCount > 0)
{
std::lock_guard<std::mutex> lock(blackboard->getMutex());
int indexedSourceFileCount = 0;
blackboard->get("indexed_source_file_count", indexedSourceFileCount);
blackboard->set("indexed_source_file_count", indexedSourceFileCount + newlyIndexedCount);
}
}
void TaskBuildIndex::updateIndexingDialog(std::shared_ptr<Blackboard> blackboard, const FilePath& sourcePath)
{
// TODO: factor in unindexed files...
int sourceFileCount = 0;
int indexedSourceFileCount = 0;
{
std::lock_guard<std::mutex> lock(blackboard->getMutex());
blackboard->get("source_file_count", sourceFileCount);
blackboard->get("indexed_source_file_count", indexedSourceFileCount);
}
if (std::shared_ptr<DialogView> dialogView = Application::getInstance()->getDialogView())
{
dialogView->updateIndexingDialog(
indexedSourceFileCount, sourceFileCount, sourcePath.str()
);
}
}