isofill broken

[doutriaux1]

@danlipsa I think your PR broke vcs

using commit: 507d2bc the uvcmetrics test 04 works, any commit after that I get the wrong colors for an isofill on test 04

git clone git://github.com/UV-CDAT/uvcmetrics
cd uvcmetrics
python setup.py install
cmake .
ctest -R 04 -D Experimental

leads to bad plots, look for example at: https://open.cdash.org/testDetails.php?test=484915834&build=4560808

Notice that we now have a level that shows values greater than 300 when the max is at: 294.

We fixed a similar issue before: CDAT/cdat#1947

PR for fix was : CDAT/cdat#2027

This is an URGENT matter for the diags due out this week.

[doutriaux1]

@danlipsa also I now get these warnings:

4: Warning: In /export/doutriaux1/anaconda2/conda-bld/work/Common/DataModel/vtkDataSet.cxx, line 476
4: vtkPolyData (0x3372220): Cell array GlobalIds with 1 components, has 1520 tuples but there are only 1411 cells
4: 
4: Warning: In /export/doutriaux1/anaconda2/conda-bld/work/Common/DataModel/vtkDataSet.cxx, line 476
4: vtkPolyData (0x3f11e50): Cell array GlobalIds with 1 components, has 1520 tuples but there are only 1411 cells
4: 
4: Warning: In /export/doutriaux1/anaconda2/conda-bld/work/Common/DataModel/vtkDataSet.cxx, line 476
4: vtkPolyData (0x3381390): Cell array GlobalIds with 1 components, has 1520 tuples but there are only 1411 cells
4: 
4: Warning: In /export/doutriaux1/anaconda2/conda-bld/work/Common/DataModel/vtkDataSet.cxx, line 476
4: vtkPolyData (0x453c810): Cell array GlobalIds with 1 components, has 1520 tuples but there are only 1411 cells
4: 
4: Warning: In /export/doutriaux1/anaconda2/conda-bld/work/Common/DataModel/vtkDataSet.cxx, line 476
4: vtkPolyData (0x36352f0): Cell array GlobalIds with 1 components, has 1008 tuples but there are only 934 cells
4: 
4: Warning: In /export/doutriaux1/anaconda2/conda-bld/work/Common/DataModel/vtkDataSet.cxx, line 476
4: vtkPolyData (0x44e7c60): Cell array GlobalIds with 1 components, has 1008 tuples but there are only 934 cells
4: 

[doutriaux1]

git revert -n 5680cc3be6afd00ee7ab06363df7a574a29acd6e

fixes the issue

[danlipsa]

@doutriaux1 I'll take a look. How to you build/install cdat with the split vcs?

[doutriaux1]

you can use the uvcdat branch split_repos

[doutriaux1]

or simply install directly from the vcs repo.

[doutriaux1]

don't forget to use: --old-and-unmanageable when building vcs

[danlipsa]

@doutriaux1 Sounds good. I'll take a look.

[doutriaux1]

@danlipsa after reverting your commit i get these failures
https://open.cdash.org/viewTest.php?onlyfailed&buildid=4560938

Shouldn't I get more?

[danlipsa]

@doutriaux1 You should not get any failure - assuming I started from a repo with no failures.

[danlipsa]

@doutriaux1 Everything I changed was in one commit, including the tests - this is the benefit of using just one commit 😄

[doutriaux1]

https://open.cdash.org/viewTest.php?onlyfailed&buildid=4560986

actually looks like the baselines are more wrong

[aashish24]

@danlipsa will look into it as the priority. Looks like the labels are wrong in this test "https://open.cdash.org/testDetails.php?test=484965777&build=4560986". All other tests looks reasonable to me but please post message here if we are mistaken.

[danlipsa]

@aashish24 Why do you say the labels are wrong?

[doutriaux1]

@danlipsa on the baseline of the test pointed to by @aashish24 the labels are on the right, they should be o the left, no?

[aashish24]

@doutriaux1 @danlipsa should we close this now?