The variable name stuff has been merged (ish) and develop is broken

[guillaumevernieres]

Background

Our develop branch is broken since the merge the of the "variable name" PR's in a few of the JCSDA repositories.

Temporary solution

I prepared a branch with a JEDI bundle that points to specific commits here:
https://github.com/NOAA-EMC/GDASApp/tree/feature/pre-ioda

[RussTreadon-NOAA]

Need to update hash for vader. CMakeLists.txt has

 # Core JEDI repositories
   ecbuild_bundle( PROJECT oops  GIT "https://github.com/jcsda/oops.git"  TAG 2ddcd606908d158f282e852483dfc861e84094ec )
   ecbuild_bundle( PROJECT vader GIT "https://github.com/jcsda/vader.git" TAG 64e94e91856954277d8ace4a0b3c455705565d61 )
   ecbuild_bundle( PROJECT saber GIT "https://github.com/jcsda/saber.git" TAG 64e94e91856954277d8ace4a0b3c455705565d61 )

Note that vader and saber have same hash. build.sh fails because it can not find the specified vader hash. Change vader hash to bbc9e9f329b1d93d09e463b2dccbdbfc964d05d4. Haven't completed build yet so can't guarantee it works.

[guillaumevernieres]

Oops, let me go look for that hash @RussTreadon-NOAA . Sorry about that!

[guillaumevernieres]

here's the vader commit that I built with @RussTreadon-NOAA : 6a07daff8b893697ebad7e8329427eb9f4a26d19

[RussTreadon-NOAA]

Refreshed my local copy of feature/pre-ioda. Now build.sh fails with

CMake Error at test/CMakeLists.txt:8 (file):
  file DOWNLOAD HASH mismatch

    for file: [/scratch1/NCEPDEV/da/Russ.Treadon/git/global-workflow/test/sorc/gdas.cd/build/test/gdasapp-fix-782aca.tgz]
      expected hash: [782aca0e0ff579557877129f242d0aa4c4252620233622ecacc1deaa524ca8ed]
        actual hash: [e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855]
             status: [28;"Timeout was reached"]



CMake Error at test/soca/CMakeLists.txt:8 (file):
  file COPY cannot find
  "/scratch1/NCEPDEV/da/Russ.Treadon/git/global-workflow/test/sorc/gdas.cd/build/test/testdata/iced.2019-04-15-43200.nc":
  No such file or directory.


CMake Error at test/atm/CMakeLists.txt:7 (file):
  file failed to create symbolic link
  '/scratch1/NCEPDEV/da/Russ.Treadon/git/global-workflow/test/sorc/gdas.cd/build/test/testdata/gdas.t06z.adpsfc.tm00.bufr_d':
  No such file or directory

I don't understand this message. A check of ftp.emc.ncep.noaa.gov shows that the tarball we want, gdasapp-fix-782aca.tgz, is present. Why does the build say it found e3b0c44? A manual "ftp ftp.emc.ncep.noaa.gov" hangs. We seem to be having network issues at NCWCP. Perhaps this is the real culprit. I'm stuck. I can't build feature/pre-ioda on Hera.

[guillaumevernieres]

Damn. I'm going to try to rebuild that branch from scratch @RussTreadon-NOAA .

[RussTreadon-NOAA]

NCWCP network issues are clearing up. The Orion and Hera builds of feature/pre-ioda are currently running and in the component compilation step.

[RussTreadon-NOAA]

Hera build completed with WORKFLOW_BUILD=ON. All 40 test_gdasapp ctests pass. Still waiting for Orion build with WORKFLOW_BUILD=ON to finish.

[ShastriPaturi]

I could successfully build on hera with feature/pre-ioda for GDASApp and run a fcst.

[RussTreadon-NOAA]

The Orion WORKFLOW_BUILD=ON build from feature/pre-ioda is complete. All test_gdasapp ctests successfully ran to completion.

[guillaumevernieres]

Thanks for testing @RussTreadon-NOAA ! Do you want to just copy the hash commits to your branch and we merge the bugfix with your open pr?

[RussTreadon-NOAA]

This works @guillaumevernieres , but PR #278 likely won't ready to merge until Monday (1/23). If we want the updated CMakeLists.txt in develop today (1/20), a PR based off this issue is the quickest route. I opt for this. This is easy and clean. I can review and approve.