Playback Scrubber

[JShep1]

This is a base implementation draft mainly to get feedback on the gui portion. Below is a small demo with only the scrubber bar itself. Any enhancements or usability improvement recommendations would be appreciated. @ColeOSRF @nkoenig @chapulina

So far I am going to add:

• Start and end times on either side of the slide bar
• The current time displayed above the slider marker
• Manually enter time to jump to

[ColeOSRF]

After you add the current time it might be good to be able to click on that and type in a specific time. Could be useful for getting the time right where you want it without having to fiddle with the slider to get small increments. Speaking of small increments maybe an arrow on each side of the slider that you can click on and it will move the time by a small increment. Not sure the best way to determine how big it should jump each time you press the arrow. Perhaps a percentage of the time? Maybe the user should be able to set the increments?

[JShep1]

After you add the current time it might be good to be able to click on that and type in a specific time. Could be useful for getting the time right where you want it without having to fiddle with the slider to get small increments. Speaking of small increments maybe an arrow on each side of the slider that you can click on and it will move the time by a small increment. Not sure the best way to determine how big it should jump each time you press the arrow. Perhaps a percentage of the time? Maybe the user should be able to set the increments?

Yeah entering the time would definitely be useful. I think we're going to rely on the World Control buttons on the lower left to control the stepping through the log playback. Perhaps this prompts a step back button to accompany the play and step forward buttons

[JShep1]

Ready for another review, here's a preview

[JShep1]

Depends on gazebosim/gz-math#152

[adlarkin]

I just wanted to comment that this seems to work well with #351. This is definitely a great addition, thanks @JShep1!

[chapulina]

The INTEGRATION_log_system test failures look suspicious:

2020-09-17T19:34:09.1303931Z /github/workspace/test/integration/log_system.cc:745: Failure
2020-09-17T19:34:09.1304402Z Value of: common::exists(logFile)
2020-09-17T19:34:09.1304708Z   Actual: false
2020-09-17T19:34:09.1304964Z Expected: true
2020-09-17T19:34:09.1305924Z �[1;33m[Wrn] [Filesystem.cc:288] �[0m�[1;33mFailed to open file [�[0m�[1;33m/github/workspace/build/test/test_logs/test_logs_record/state.tlog�[0m�[1;33m]: �[0m�[1;33mNo such file or directory�[0m�[1;33m�[0m
2020-09-17T19:34:09.1306845Z /github/workspace/test/integration/log_system.cc:750: Failure
2020-09-17T19:34:09.1307359Z Value of: common::exists(logPlaybackFile)
2020-09-17T19:34:09.1307721Z   Actual: false
2020-09-17T19:34:09.1307963Z Expected: true
2020-09-17T19:34:09.1308316Z Failed to open the requested sqlite3 database
2020-09-17T19:34:09.1308809Z /github/workspace/test/integration/log_system.cc:773: Failure
2020-09-17T19:34:09.1309713Z Expected: (batch.end()) != (recordedIter), actual: 8-byte object <90-85 47-7B B7-55 00-00> vs 8-byte object <A0-2D 00-8C 5E-7F 00-00>

[JShep1]

The INTEGRATION_log_system test failures look suspicious:

Agreed, is this a flaky test? Jenkins shows it as not being a new failure, unless I'm interpreting it wrong.

[chapulina]

Agreed, is this a flaky test? Jenkins shows it as not being a new failure

If you look here, it seems to be flaky because it's failed before:

https://build.osrfoundation.org/job/ignition_gazebo-ci-pr_any-ubuntu_auto-amd64/4177/testReport/junit/(root)/INTEGRATION_log_system/test_ran/history/

But if you look into each of those builds, they've all been builds related to this PR.

If you look at the history of this test for the ign-gazebo2 branch, it looks like it's been passing consistently:

https://build.osrfoundation.org/job/ignition_gazebo-ci-ign-gazebo2-bionic-amd64/149/testReport/(root)/LogSystemTest/history/

I haven't tried reproducing the failure locally yet, but I think we should take a look.

[chapulina]

I can reproduce the failures locally and the latest ign-gazebo2 build that I just triggered doesn't have them. I'm looking into it.

https://build.osrfoundation.org/job/ignition_gazebo-ci-ign-gazebo2-bionic-amd64/150/

[chapulina]

The changes in 802e92b fixed the tests for me locally. I have no idea why that would be an issue on this PR but not on ign-gazebo2 though 🤷‍♀️

[chapulina]

Phew, so it looks like log_system is fixed. And I just noticed that Gui_TEST is also a new failure that I can reproduce locally. On it.

[chapulina]

Ok, Gui_TEST is fixed in a7fbd0c.

[JShep1]

Thanks for fixing those issues! I'm assuming we'll be good to merge once CI runs now?

[chapulina]

Argh, it looks like log_system has been passing consistently on Mac:

https://build.osrfoundation.org/job/ignition_gazebo-ci-ign-gazebo2-homebrew-amd64/156/testReport/(root)/LogSystemTest/history/

And I see it passed on previous commits of this PR. Let's run it one more time just to sanity check. I retriggered it.

[chapulina]

Mac passed now! Merging! 🎉