SNF4CGP

履歴

adeniz 95eaa859f9 Updated version to 1.4.1. git-svn-id: https://svn.microneil.com/svn/SNF4CGP/trunk@60 59e8e3e7-56fa-483b-b4b4-fa6ab0af3dfc		8年前
..
Command.hpp	Fixed bad cin bug - if cin went bad the executive process would loop, rapidly generate UNKNOWN commands, and produce FAILURE outputs. This would happen when CGP was disabling the plugin (Specs say the plugin should get QUIT first, but it didn't). The fix creates a new Command::NOTGOOD that is produced by getCommand() if cin.good() becomes false. The executive process responds to Command::NOTGOOD by announcing the input has gone bad and then shutting down cleanly.	15年前
ConfigurationEngine.cpp	Fixed bug where Reject action told CGP REJECTED. Quixotically CGPs API indicates that to Reject a message it needs to see ERROR, and that when it sees REJECTED message processing is actually postponed. A look at CGP logs indicates that REJECTED messages are scanned again later.	14年前
ConfigurationEngine.hpp	Fixed bug where Reject action told CGP REJECTED. Quixotically CGPs API indicates that to Reject a message it needs to see ERROR, and that when it sees REJECTED message processing is actually postponed. A look at CGP logs indicates that REJECTED messages are scanned again later.	14年前
ExecutiveProcess.cpp	Changed log and startup message prefix from "SNF4CGP" to "CGPSNF".	15年前
ExecutiveProcess.hpp	Fixed potential race condition where the ExecutiveProcess might send a notification message through cout while the OutputProcessor was busy using cout also. The result would be (at best) corrupted output. Now the ExecutiveProcess and the OutputProcessor both use OutputProcessor::sendString() which protects cout with a mutex.	15年前
InputProcessor.cpp	Fixed bad cin bug - if cin went bad the executive process would loop, rapidly generate UNKNOWN commands, and produce FAILURE outputs. This would happen when CGP was disabling the plugin (Specs say the plugin should get QUIT first, but it didn't). The fix creates a new Command::NOTGOOD that is produced by getCommand() if cin.good() becomes false. The executive process responds to Command::NOTGOOD by announcing the input has gone bad and then shutting down cleanly.	15年前
InputProcessor.hpp	Refactored a bit to solve circular references in module includes.	15年前
JobPool.cpp	Corrected bug where Hold action cannot complete because the reader is not available. Previous "improvement" was based on a bad assumption that responses were emitted during the job. Actually, responses are emitted after the job is finalized and sent to the output processor so that theory was incorrect. Earlier problem was related to something else & this code was ok.	11年前
JobPool.hpp	Fixed bug where Reject action told CGP REJECTED. Quixotically CGPs API indicates that to Reject a message it needs to see ERROR, and that when it sees REJECTED message processing is actually postponed. A look at CGP logs indicates that REJECTED messages are scanned again later.	14年前
OutputProcessor.cpp	Bug fix: doFile now closes message reader immediately after scan so that the file is not open when CGP resumes processing.	11年前
OutputProcessor.hpp	Fixed potential race condition where the ExecutiveProcess might send a notification message through cout while the OutputProcessor was busy using cout also. The result would be (at best) corrupted output. Now the ExecutiveProcess and the OutputProcessor both use OutputProcessor::sendString() which protects cout with a mutex.	15年前
ScannerPool.cpp	Fixed bug - PlatformVersion was not reported. PlatformVersion is now propagated through the initialization sequence to the scanner pool and ultimately the rulebase so it can be reported properly.	15年前
ScannerPool.hpp	Fixed bug - PlatformVersion was not reported. PlatformVersion is now propagated through the initialization sequence to the scanner pool and ultimately the rulebase so it can be reported properly.	15年前
WorkerPool.cpp	Corrected shutdown deadlock in Workers.	15年前
WorkerPool.hpp	Corrected shutdown deadlock in Workers.	15年前
main.cpp	Updated version to 1.4.1.	8年前