UNIT_MainWindow_TEST fails in Jenkins but not on local machine

[osrf-migration]

Original report (archived issue) by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).

---

Here's an example test failure with error message:

Unknown file:0 :: Received a fatal error.

The strange thing is that this test passes on my local machine.

I have been able to see failures of this test in my overnight testing with the gz_build_test.bash script. They have the following console output:

61: Test command: /tmp/gazebo_build/source/build/gazebo/gui/UNIT_MainWindow_TEST "-xml" "-o" "/tmp/gazebo_build/source/build/test_results/UNIT_MainWindow_TEST.xml"
61: Test timeout computed to be: 240
61: 
61: Msg Waiting for masterMsg Connected to gazebo master @ http://127.0.0.1:11345
61: Msg Publicized address: 192.168.1.183
61: Dbg AvgPrewireframe [0] AvgPostWireframe[0]
61: SharePercentChange[1.40118] ShareMaxPercentChange[2]
61: ResPercentChange[3.06243]ResMaxPercentChange[4]

For this case, I believe the failure is related to the line that prints the average color before and after wireframe is turned on:

61: Dbg AvgPrewireframe [0] AvgPostWireframe[0]

since there is a QVerify statement that expects these values to be different.

In the morning, the script is still running, and once I log in to the machine with the monitor turned on, the test starts passing. So perhaps some hardware component is disabled when the screen is turned off.

[osrf-migration]

Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).

---

• set component to "testing"

[osrf-migration]

Original comment by Jose Luis Rivero (Bitbucket: Jose Luis Rivero, GitHub: j-rivero).

---

I was looking at this issue sometime during last weeks. I was unable to reproduce it using the same development machine, logging into base system and run the tests from there. I think it should be something related to run tests from a chroot, but not sure.

[osrf-migration]

Original comment by Jose Luis Rivero (Bitbucket: Jose Luis Rivero, GitHub: j-rivero).

---

I made a chroot in my local machine and tried to run the test from there. It works just fine.

[osrf-migration]

Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).

---

It fails for me when I run overnight and my monitor is off. Can you try it with your monitor disconnected? ie. ssh in from another machine?

[osrf-migration]

Original comment by Jose Luis Rivero (Bitbucket: Jose Luis Rivero, GitHub: j-rivero).

---

I've launched the test with a sleep of 5 seconds, I've turned off the monitor and test is running fine. Even disconnecting the cable.

[osrf-migration]

Original comment by Jose Luis Rivero (Bitbucket: Jose Luis Rivero, GitHub: j-rivero).

---

During the last in person discussion @nkoenig @scpeters and me arrived at the following conclusions regards to this bug:

• This test is based on the image rendering at the screen. If the screen is off, the image is completely black. We need to find a way of disabling this kind of tests for jenkins.
• It would be nice have the same test not depending on screen rendering.

[osrf-migration]

Original comment by Jose Luis Rivero (Bitbucket: Jose Luis Rivero, GitHub: j-rivero).

---

• changed state from "new" to "resolved"

Done in pull request #764

[osrf-migration]

Original comment by Nate Koenig (Bitbucket: Nathan Koenig).

---

• set version to "all"

[osrf-migration]

Original comment by Nate Koenig (Bitbucket: Nathan Koenig).

---

• changed state from "resolved" to "closed"