@aashish24 this is a great candidate for testing out the travis build issue - the lean build never touches the spyder source, so there is absolutely no reason why this pull request should have reported an error.

@aashish24 @jbeezley @doutriaux1 should I tweak and resubmit this PR for testing through the new build system? As noted above spyder is never even touched in the lean builds that were configured last time this PR was updated...

@durack1 don't worry about it. Can it wait until 2.2 happens?

@aashish24 @doutriaux1 if it was possible it'd be great to get this patch update into 2.2.. But if you guys are trying to simplify things 2.3 would be ok I suppose..

@durack1 we have a strict deadline which is due in next few days. You wait is really appreciated. We will get it merged as soon as we tag 2.2

@doutriaux1 @aashish24 I just had a go at rebasing this off UV-CDAT/uvcdat release branch, I'm doubtful this will build.. So please ignore this..

@doutriaux1 @aashish24 ok wow, it did (not all that surprising as it's now basically just the release branch, with the spyder update which is never built)..

@doutriaux1 I'm not sure you folks feel comfortable about merging such changes for 2.2?

@durack1 I'll try a full build again once #1267 has been approved. If it works then I guess i'll merge it in.

Ok great, I'll do the same rebasing for #1129, #1130 and #1131 and see how these also go..

@doutriaux1 you might want to cherry pick the one file that's changed CMake/cdat_modules/spyder_pkg.cmake

@doutriaux1 @aashish24 @jbeezley this PR should be deprecated by #1300 which seems to have triggered the buildbot testing

@aashish24 @jbeezley @doutriaux1 it still seems those 3 errors, DVD3D image lines, DODS and netcdf version being picked up errors are still occurring: https://open.cdash.org/viewTest.php?onlyfailed&buildid=3808629

@aashish24 @jbeezley it seems that netcdf version issue is only a problem on garant. The test-laptop only has the DVD3D image line and the DODS issue.. https://open.cdash.org/viewTest.php?onlyfailed&buildid=3808646

I fixed that in #1298, which isn't in master. test-laptop is compiling libcf against a system installed lapack. garant doesn't have a system installed lapack, so it builds without regridding.

@doutriaux1 @aashish24 should #1298 also be merged into master then?

@durack1 looks like @jbeezley had it merged into another PR.

@doutriaux1 closing this and will reissue a new PR rebased off current master