datacommsserver/esockserver/test/TE_EsockTestSteps/inc/Connections.TestSteps.h
branchRCL_3
changeset 25 9d7ce34704c8
parent 23 cbb19216b74d
child 26 b564fb5fd78b
--- a/datacommsserver/esockserver/test/TE_EsockTestSteps/inc/Connections.TestSteps.h	Thu Aug 19 11:05:47 2010 +0300
+++ b/datacommsserver/esockserver/test/TE_EsockTestSteps/inc/Connections.TestSteps.h	Tue Aug 31 16:25:36 2010 +0300
@@ -190,73 +190,21 @@
 Note that since the test is normally executed by a thread of lower priority than ESock_IP, it relies on
 ESock_IP yielding (otherwise it won't test different phases. 
 For instance DummyMCPR will artificially yield on TCFServiceProvider::TJoinRequest.
-NetMCPR will not normally yield as is. DummyFlowFactory will yield on serving a TFindOrCreateRequest.
+NetMCPR will not normally yield as is. 
 
 @internalComponent
 */
 class CStartStopCrazyLoopRConnectionStep : public CStartRConnectionStep
     {
 public:
-    enum {
-    /*
-     * The test executes a loop of starting and then stopping a connection at discrete, growing intervals.
-     * It is crucial that enough iterations are executed, as otherwise the test isn't doing its job.
-     * The test will fail if this is detected. As iteration gaps are time-based the test first calibrates 
-     * itself by running CStartStopCrazyLoopRConnectionStep::CalibrateStart to determine how 
-     * much time does the start take in order to chop this span further into ELoops intevals. 
-     * Calibration isn't uterrly precise and it can be that a connection startup executed within
-     * the loop completes quicker than during calibration. It has been in fact observed 
-     * in the overnight builds (about 7.5%). A number of things can be done to accomdate for that.
-     * The approach chosen is to: 
-     * (1) Plan for EPlannedLoops (chop down the CStartStopCrazyLoopRConnectionStep::CalibrateStart 
-     *     into EPlannedLoops
-     * (2) Insist on excuting at least EMandatoryLoops (the test will fail if less iterations are executed)
-     * (3) Keep on iterating beyond EPlannedLoops until the snooze slides of the connection 
-     *     startup time (the calbrated connection startup time equally might have been unrepresentativelly short)
-     */     
-    EPlannedLoops = 15,
-    EMandatoryLoops = 14,
-    };
-    
     CStartStopCrazyLoopRConnectionStep(CCEsockTestBase*& aEsockTest);
-    virtual ~CStartStopCrazyLoopRConnectionStep();
-    
     TVerdict doSingleTestStep();
     TInt CalibrateStart();
-    virtual TInt doLoopStep(TInt aTimerVal);
-    
-protected:
-    RTimer iTimer;
-    TRequestStatus* ipConnectionStartStatus;
     };
 
 _LIT(KStartStopCrazyLoopRConnectionStep,"StartStopCrazyLoopRConnectionStep");
 
 /**
-Class implementing StartCloseCrazyLoopRConnectionStep
-The test does a loop, where every iteration:
-(1) Starts the connection
-(2) Snoozes for an <interval>
-(3) Closes the connection injecting a cancel at that stage of the connection start
-(4) Increases the <interval> a bit so that next time round the cancel is injected at a different stage.
-
-Note that since the test is normally executed by a thread of lower priority than ESock_IP, it relies on
-ESock_IP yielding (otherwise it won't test different phases. 
-For instance DummyMCPR will artificially yield on TCFServiceProvider::TJoinRequest.
-NetMCPR will not normally yield as is. DummyFlowFactory will yield on serving a TFindOrCreateRequest.
-
-@internalComponent
-*/
-class CStartCloseCrazyLoopRConnectionStep : public CStartStopCrazyLoopRConnectionStep
-    {
-public:
-    CStartCloseCrazyLoopRConnectionStep(CCEsockTestBase*& aEsockTest);
-    virtual TInt doLoopStep(TInt aTimerVal);
-    };
-
-_LIT(KStartCloseCrazyLoopRConnectionStep,"StartCloseCrazyLoopRConnectionStep");
-
-/**
 Class implementing openrconnectionStep
 
 @internalComponent