Enable Live Listener to request data from the start of the run

[Ross Miller]

Allow the SNS Live Listener to receive data all the way back to the beginning of the run when it first connects.

[Ross Miller]

Development branch is "feature/6661_LiveListenerDataStart"

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: c53ffd3ff3396a80787a3fa84f4172bfc32cc4ed

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 3ca82b8317c65f2c95f8de0c377a2ae0ffb69989

[Russell Taylor]

I'm taking this to get it out of the list and sit on it until the SMS that can handle it is deployed.

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: 175ed95afab73d32dbd66893d4c04fa773afb8d5

[Ross Miller]

GUI changes to the StartLiveData dialog

Changed the layout of the radio buttons, added "UTC" to the absolute time label and added code to limit the times the edit box will accept

Refs #6661

Changeset: 3c7ba438a239f8453a22c063a5d5adf6990189f2

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d1efa7da04a0920a6b0ae601dcdc0dacb1ece22c

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 7120a3e36629d9e2db37ce56c1cfa248b1e4df08

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 17c1f7873da0334ca07c62cc6874e992cdad7f0a

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: c53ffd3ff3396a80787a3fa84f4172bfc32cc4ed

[Ross Miller]

GUI changes to the StartLiveData dialog

Changed the layout of the radio buttons, added "UTC" to the absolute time label and added code to limit the times the edit box will accept

Refs #6661

Changeset: 9380b3176f1cd71aa3673fead8fe7225afad5eb8

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 3ca82b8317c65f2c95f8de0c377a2ae0ffb69989

[Russell Taylor]

When this works, it works really nicely. However, I sometimes get a failure:

Error in execution of algorithm LoadInstrument:
nthValue(): TimeSeriesProperty 's2' is empty
Caught a runtime exception.
Exception message: nthValue(): TimeSeriesProperty 's2' is empty
Thread will exit.

What we need to do if possible is delay the loading of the instrument until the logs have been populated. We could do something around catching the exception, but that would leave some detectors in the wrong place.

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: c53ffd3ff3396a80787a3fa84f4172bfc32cc4ed

[Ross Miller]

GUI changes to the StartLiveData dialog

Changed the layout of the radio buttons, added "UTC" to the absolute time label and added code to limit the times the edit box will accept

Refs #6661

Changeset: 9380b3176f1cd71aa3673fead8fe7225afad5eb8

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 3ca82b8317c65f2c95f8de0c377a2ae0ffb69989

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: c53ffd3ff3396a80787a3fa84f4172bfc32cc4ed

[Ross Miller]

GUI changes to the StartLiveData dialog

Changed the layout of the radio buttons, added "UTC" to the absolute time label and added code to limit the times the edit box will accept

Refs #6661

Changeset: 9380b3176f1cd71aa3673fead8fe7225afad5eb8

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 3ca82b8317c65f2c95f8de0c377a2ae0ffb69989

[Ross Miller]

Delay initialization of the live listener

Don't perform the final initialization steps for the live listener until the first BankedEventPacket arrives. We don't need to complete the init steps any earlier and waiting as long as possible ensures all the device decriptor and variable value packets have been processed.

Refs #6661

Changeset: fb8a1d00f7e4727a37f5b1eeda3c89d77b1ac324

[Ross Miller]

Delay initialization of the live listener

Don't perform the final initialization steps for the live listener until the first BankedEventPacket arrives. We don't need to complete the init steps any earlier and waiting as long as possible ensures all the device decriptor and variable value packets have been processed.

Refs #6661

Changeset: fb8a1d00f7e4727a37f5b1eeda3c89d77b1ac324

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: c53ffd3ff3396a80787a3fa84f4172bfc32cc4ed

[Ross Miller]

GUI changes to the StartLiveData dialog

Changed the layout of the radio buttons, added "UTC" to the absolute time label and added code to limit the times the edit box will accept

Refs #6661

Changeset: 9380b3176f1cd71aa3673fead8fe7225afad5eb8

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 3ca82b8317c65f2c95f8de0c377a2ae0ffb69989

[Ross Miller]

Delay initialization of the live listener

Don't perform the final initialization steps for the live listener until the first BankedEventPacket arrives. We don't need to complete the init steps any earlier and waiting as long as possible ensures all the device decriptor and variable value packets have been processed.

Refs #6661

Changeset: fb8a1d00f7e4727a37f5b1eeda3c89d77b1ac324

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: c53ffd3ff3396a80787a3fa84f4172bfc32cc4ed

[Ross Miller]

GUI changes to the StartLiveData dialog

Changed the layout of the radio buttons, added "UTC" to the absolute time label and added code to limit the times the edit box will accept

Refs #6661

Changeset: 9380b3176f1cd71aa3673fead8fe7225afad5eb8

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 3ca82b8317c65f2c95f8de0c377a2ae0ffb69989

[Ross Miller]

Delay initialization of the live listener

Don't perform the final initialization steps for the live listener until the first BankedEventPacket arrives. We don't need to complete the init steps any earlier and waiting as long as possible ensures all the device decriptor and variable value packets have been processed.

Refs #6661

Changeset: fb8a1d00f7e4727a37f5b1eeda3c89d77b1ac324

[Ross Miller]

Changes to allow Live Listener to replay historic data from SMS

This is just the first cut - few more minor changes are still needed.

Refs #6661

Changeset: c53ffd3ff3396a80787a3fa84f4172bfc32cc4ed

[Ross Miller]

GUI changes to the StartLiveData dialog

Changed the layout of the radio buttons, added "UTC" to the absolute time label and added code to limit the times the edit box will accept

Refs #6661

Changeset: 9380b3176f1cd71aa3673fead8fe7225afad5eb8

[Ross Miller]

Fix GUI problems in RHEL 6 builds

The previous GUI changes used a function that wasn't available in the version of Qt that comes with RHEL6.

Refs #6661

Changeset: d71071832da6d9caf8e230f5cd75ad7d911d0f1d

[Ross Miller]

Change handling of RunStatus packets

Refs #6661

Changeset: 6ea1d522d4aa36a3f8fd0deb2fe71646424ed4f7

[Ross Miller]

Add code to validate requested start time for live listener

Added code to LoadLiveData.cpp to validate the time string for the requested start time. Also use the epoch time for the "start from now" option (instead of using the current time). The epoch time gets converted to 0 when passed to SMSD and SMSD handles that a little more efficiently than passing it the current time.

Refs #6661

Changeset: 3ca82b8317c65f2c95f8de0c377a2ae0ffb69989

[Ross Miller]

Delay initialization of the live listener

Don't perform the final initialization steps for the live listener until the first BankedEventPacket arrives. We don't need to complete the init steps any earlier and waiting as long as possible ensures all the device decriptor and variable value packets have been processed.

Refs #6661

Changeset: fb8a1d00f7e4727a37f5b1eeda3c89d77b1ac324

[Russell Taylor]

This is still locking up when the Run status is 'Stopped'. Unfortunately, that's how it's been most of the day so I can't even recheck for an actual run.

[Russell Taylor]

We need to sort out the problem where the algorithm locks up if no BankedEvent packets are received (e.g. when outside a run). Seems like this can't be achieved in time for this release.

[Ross Miller]

Don't initialize live listener without required log values

Parse the instrument definition XML looking for nodes named 'logfile' that are children of nodes named 'parameter'. Record the 'id' attribute for these nodes. Prior to executing the LoadInstrument algorithm, verify that we've received values for each of logs.

Refs #6661

Changeset: eb1f716487c2edf757b165e805639c1b33fd267d

[Ross Miller]

Don't let the live listener block the main thread indefinitely

Limit the amount of time the main thread will block inside extractData() waiting for the live listener background thread to initialize.

Refs #6661

Changeset: 1de96afed0a0f1fce8a522cfaa59d01f2027c504

[Ross Miller]

Ensure that the live listener background thread always exits

If the user canceles the live listener algorithm at just the right time, it was possible for the background thread to be stuck in a loop waiting for something from the foreground thread and thus the thread would never exit. This change adds a couple of checks to make sure the background thread won't get stuck.

Refs #6661

Changeset: 476b227e5e23b491e81e062504ea8ad43165072c

[Ross Miller]

Don't initialize live listener without required log values

Parse the instrument definition XML looking for nodes named 'logfile' that are children of nodes named 'parameter'. Record the 'id' attribute for these nodes. Prior to executing the LoadInstrument algorithm, verify that we've received values for each of logs.

Refs #6661

Changeset: eb1f716487c2edf757b165e805639c1b33fd267d

[Ross Miller]

Don't let the live listener block the main thread indefinitely

Limit the amount of time the main thread will block inside extractData() waiting for the live listener background thread to initialize.

Refs #6661

Changeset: 1de96afed0a0f1fce8a522cfaa59d01f2027c504

[Ross Miller]

Ensure that the live listener background thread always exits

If the user canceles the live listener algorithm at just the right time, it was possible for the background thread to be stuck in a loop waiting for something from the foreground thread and thus the thread would never exit. This change adds a couple of checks to make sure the background thread won't get stuck.

Refs #6661

Changeset: 476b227e5e23b491e81e062504ea8ad43165072c

[Russell Taylor]

Choosing 'Start of Run' while the instrument is taking data leads to the following:

SNSLiveEventDataListener-[Fatal] Caught an invalid packet exception in SNSLiveEventDataListener network read thread.
SNSLiveEventDataListener-[Fatal] Exception message is: VariableValue (double) packet has invalid status: 65534
SNSLiveEventDataListener-[Fatal] Thread is exiting.

Choosing to start from 'Now' is OK.

[Russell Taylor]

Hmmm - that one seems to have gone away now. Could it have been a function of being deep into a run?

However, I've now been seeing initialization timeouts when (I think) they're outside a run.

[Andrei Savici]

Merge branch 'feature/6823_DakotaChiSquared_output' into develop into 6856_ConvertToDiffractionMDWS_v2

[Ross Miller]

Merge branch 'feature/6661_LiveListenerDataStart' into develop into 6856_ConvertToDiffractionMDWS_v2

[Stuart Campbell]

Merge branch 'feature/7021_cmake_linux_distro' into develop into 6856_ConvertToDiffractionMDWS_v2

[Mathieu Doucet]

Merge branch 'feature/6516_sans_reduction_apiv2' into develop into 6856_ConvertToDiffractionMDWS_v2

[Andrei Savici]

Merge branch 'feature/6823_DakotaChiSquared_output' into develop into 6856_ConvertToDiffractionMDWS_v2

[Ross Miller]

Merge branch 'feature/6661_LiveListenerDataStart' into develop into 6856_ConvertToDiffractionMDWS_v2

[Stuart Campbell]

Merge branch 'feature/7021_cmake_linux_distro' into develop into 6856_ConvertToDiffractionMDWS_v2

[Mathieu Doucet]

Merge branch 'feature/6516_sans_reduction_apiv2' into develop into 6856_ConvertToDiffractionMDWS_v2

[Ross Miller]

More cleanups to the Live Listener init process

Added checks to see if we could complete the init process to several of the rxPacket() functions. Also, if the user requests data from the start of the current run and we're not actually in a run, we'll throw an exception with an appropriate error message.

Refs #6661

Changeset: 1c591d1d8e688d5af916e9bb1d5128634743475e

[Ross Miller]

Merge branch 'feature/6661_LiveListenerDataStart' into develop

[Russell Taylor]

The out of run behaviour is not quite what we want. It gives the following:

StartLiveData-[Notice] StartLiveData started
SNSLiveEventDataListener-[Fatal] Caught a runtime exception.
SNSLiveEventDataListener-[Fatal] Exception message: User requested data from start of current run, but system is not currently in a run.
SNSLiveEventDataListener-[Fatal] Thread will exit.
LoadLiveData-[Notice] Performing the Replace operation.
MonitorLiveData-[Notice] MonitorLiveData started
StartLiveData-[Notice] StartLiveData successful, Duration 0.11 seconds
MonitorLiveData-[Notice] Loading live data chunk 0 at 14:35:19
MonitorLiveData-[Error] Error in execution of algorithm MonitorLiveData:
MonitorLiveData-[Error] User requested data from start of current run, but system is not currently in a run.

It also hands back an empty (1 spectrum) workspace. We want the exception to come out the first time extractData() is called so that StartLiveData fails rather than apparently succeeding. We don't want to end up with a workspace.

[Ross Miller]

Don't set m_workspaceInitialized to true in exception handlers

Originally did this to prevent us from blocking indefinitely waiting for an initialization that would never happen, but that code has since changed and modifying m_workspaceInitialized is no longer desirable

Refs #6661

Changeset: a41c2e6e526282befeda84530dab2f72debf7361

[Ross Miller]

Merge branch 'feature/6661_LiveListenerDataStart' into develop

[Russell Taylor]

Works perfectly when in a run and exhibits the behaviour we decided upon when outside a run - which is to just fail the algorithm. Given that we fell back to this 'simple' behaviour, I wonder whether all of the complexity in the SNSLiveEventDataListener (e.g. XML parsing) is really necessary - the class is pretty complex as it is and its length has increased by ~30% under this ticket.

[Stuart Campbell]

This ticket has been transferred to github issue ​7507