2.4.1rc test

[doutriaux1]

@aashish24 @chaosphere2112 @danlipsa @sankhesh

please build and merge as soon as possible. Thanks.

[doutriaux1]

goes with: CDAT/uvcdat-testdata#125

[aashish24]

@doutriaux1 why comment out tests? The release user will see won't be running tests so I don't know a strong need for this? thoughts?

[doutriaux1]

because they do not work on every machine? And like that it will look cleaner to Dean.

[aashish24]

because they do not work on every machine? And like that it will look cleaner to Dean.

I am fine with it if could you create an issue for these (I believe I have the issue created for continents but not sure about the other test).

[aashish24]

@doutriaux1 I see 5 commits but only changes to one file. Does this sounds about right to you?

[doutriaux1]

@aashish24 squash them when you merge it in master. I had to play with the threshold and the tag point of the uvcmetrics. Please more importantly try it and run ctest on your machine

[aashish24]

@aashish24 squash them when you merge it in master. I had to play with the threshold and the tag point of the uvcmetrics. Please more importantly try it and run ctest on your machine

Okay sounds good. I will do a clean build in next hour or so and will update.

[doutriaux1]

thanks @aashish24

[aashish24]

@doutriaux1 build is running as I am typing.

[chaosphere2112]

@doutriaux1 Running ctests now

[chaosphere2112]

I ran tests against the wrong testdata branch, so I'm running the diags again (Everything else passed except test_vcs_boxfill_robinson_wrap, which looks like a threshold issue).

[painter1]

Looking at some failures from the dashboard, I see several cases where the newest plots have a thicker line around the plot border than the baseline plots.

[chaosphere2112]

Diags tests (Doing another that will push to the dashboard):

The following tests FAILED:
    562 - diags_test_04 (Failed)
    563 - diags_test_41 (Failed)
    564 - diags_test_05 (Failed)
    565 - diags_test_06 (Failed)
    566 - diags_test_07 (Failed)
    567 - diags_test_08 (Failed)
    568 - diags_test_09 (Failed)
    570 - diags_test_11 (Failed)
    572 - diags_test_13 (Failed)
    573 - diags_test_15 (Failed)

[chaosphere2112]

@painter1 Yeah, those can generally be fixed by increasing the threshold setting for the checkimage call that compares the .png files

[aashish24]

I am on the vpn so it took longer to download and build. About to submit test results.

[aashish24]

548 - CDMS_Test_multiple_formats (Failed)
560 - diags_test_02 (Failed)
561 - diags_test_03 (Failed)
562 - diags_test_04 (Failed)
563 - diags_test_41 (Failed)
564 - diags_test_05 (Failed)
565 - diags_test_06 (Failed)
566 - diags_test_07 (Failed)
567 - diags_test_08 (Failed)
568 - diags_test_09 (Failed)
569 - diags_test_10 (Failed)
570 - diags_test_11 (Failed)
571 - diags_test_12 (Failed)
572 - diags_test_13 (Failed)
573 - diags_test_15 (Failed)

[aashish24]

Pushing it on the dashboard as well.

[doutriaux1]

@chaosphere2112 and I took a look, looks like VTK is not antialiasing lines on mac... Not going to delay the release for this. I'd say we're clear.

[doutriaux1]

I meant VTK is no respecting the "no-antialiasing" on mac. It IS antialiasing on mac whereas it shouldn't be.