mkmf submodule broken for ESM4 / OMIP tag

[jkrasting]

The mkmf submodule no longer checks out using the tag dev/gfdl/2018.04.11. The error message is:

error: Server does not allow request for unadvertised object 309d50df87a455337ec3c9bd6d0f43ede348e0c6
Fetched in submodule path 'src/mkmf', but it did not contain 309d50df87a455337ec3c9bd6d0f43ede348e0c6. Direct fetching of that commit failed.

Here are the checkout commands:

git clone https://github.com/NOAA-GFDL/MOM6-examples.git
cd MOM6-examples/
git checkout dev/gfdl/2018.04.11
git submodule init
git submodule update --recursive

And the output:

Cloning into 'MOM6-examples'...
remote: Counting objects: 41819, done.
remote: Compressing objects: 100% (86/86), done.
remote: Total 41819 (delta 87), reused 75 (delta 57), pack-reused 41675
Receiving objects: 100% (41819/41819), 45.25 MiB | 26.24 MiB/s, done.
Resolving deltas: 100% (35160/35160), done.

Note: checking out 'dev/gfdl/2018.04.11'.

You are in 'detached HEAD' state. You can look around, make experimental
changes and commit them, and you can discard any commits you make in this
state without impacting any branches by performing another checkout.

If you want to create a new branch to retain commits you create, you may
do so (now or later) by using -b with the checkout command again. Example:

  git checkout -b <new-branch-name>

HEAD is now at 192fb89b... Merge branch 'jkrasting-user/jpk/OM4_05_remap_overrides' into dev/gfdl

Submodule 'src/FMS' (https://github.com/NOAA-GFDL/FMS.git) registered for path 'src/FMS'
Submodule 'src/MOM6' (https://github.com/NOAA-GFDL/MOM6.git) registered for path 'src/MOM6'
Submodule 'src/SIS2' (https://github.com/NOAA-GFDL/SIS2.git) registered for path 'src/SIS2'
Submodule 'src/atmos_null' (https://github.com/NOAA-GFDL/atmos_null.git) registered for path 'src/atmos_null'
Submodule 'src/coupler' (https://github.com/NOAA-GFDL/coupler.git) registered for path 'src/coupler'
Submodule 'src/icebergs' (https://github.com/NOAA-GFDL/icebergs.git) registered for path 'src/icebergs'
Submodule 'src/land_null' (https://github.com/NOAA-GFDL/land_null.git) registered for path 'src/land_null'
Submodule 'src/mkmf' (https://github.com/NOAA-GFDL/mkmf.git) registered for path 'src/mkmf'
Submodule 'tools/analysis/mpl-cmocean' (https://github.com/matplotlib/cmocean.git) registered for path 'tools/analysis/mpl-cmocean'
Submodule 'tools/matlab/gtools' (https://github.com/Adcroft/gtools.git) registered for path 'tools/matlab/gtools'
Submodule 'tools/python/MIDAS' (https://github.com/mjharriso/MIDAS.git) registered for path 'tools/python/MIDAS'

Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/FMS'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/MOM6'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/SIS2'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/atmos_null'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/coupler'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/icebergs'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/land_null'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/src/mkmf'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/tools/analysis/mpl-cmocean'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/tools/matlab/gtools'...
Cloning into '/lustre/f1/unswept/John.Krasting/MOM6-examples/tools/python/MIDAS'...
Submodule path 'src/FMS': checked out '34097d072ca2913668350ebd1a04957516c8fb67'
Submodule path 'src/MOM6': checked out 'b190a305031dbb6c277b8af674643e867709478a'
Submodule path 'src/SIS2': checked out 'a3be158607ac4c2cf56610c55b5f7254c76db821'
Submodule path 'src/atmos_null': checked out 'a149b0b6d2375c2221d108c44283a14dbdaaf404'
Submodule path 'src/coupler': checked out '530e9a184fd47600b0f0a2a87060466cbe13a862'
Submodule path 'src/icebergs': checked out '33603c93e5807969476bcb4a9cdc37ec2e360089'
Submodule path 'src/land_null': checked out 'b7a8d92ade6125ae7281dd1ca84078ce226440b4'
error: Server does not allow request for unadvertised object 309d50df87a455337ec3c9bd6d0f43ede348e0c6
Fetched in submodule path 'src/mkmf', but it did not contain 309d50df87a455337ec3c9bd6d0f43ede348e0c6. Direct fetching of that commit failed.

@adcroft and @underwoo - I see the mkmf repo was updated a few days ago. Did something go awry with the repo?

[underwoo]

Nothing changed. Just some cleanup.

[jkrasting]

Are you able to replicate this? Any idea why git can no longer checkout this specific commit? It seems that the mkmf directory is empty, except for the .git directory. -- ----- Dr. John Krasting, Physical Scientist Biogeochemistry, Ecosystems, and Climate Group NOAA/Geophysical Fluid Dynamics Laboratory 201 Forrestal Road, Princeton, NJ 08540 P. (609) 452-5359 | F. (609) 987-5063

[adcroft]

I can reproduce your problem.

Commit NOAA-GFDL/mkmf@309d50d does still exist (click the link).

However, that commit used to be on a branch "find.symlinks" which has been deleted. From an older clone:

git log --oneline ..origin/find.symlinks
309d50d Better fix for symlinks
8895cac list_paths now finds symlinks as well

That branch was clearly not merged because the commit is not in the history for the sole remaining branch "master":

$ git log --oneline
e22242f New template for Travis-CI (#16)
e746b66 Version 19.3.0 (#15)
44bce48 Add ability to use CPP to preprocess Fortran files (#14)
f6e28ed Updated mkmf version to 19.2.0
fc38690 Add -l option for link options
27b7a82 Update ncrc templates for latest ncrc/c3 settings
725a382 Fixed FFLAGS_REPRO options for ncrc-pgi
1848160 Clean up documentation duplication
932dd47 Merge branch 'git_version_tag'
421d030 Merge in github changes
8400755 Bring in changes in upstream repository
2af5b15 Add git file hash as cpp macro to compile command
fa104ce COPYING file now in markdown
6b10024 Merge pull request #8 from nicjhan/gnu-fpe-trap
544e7e3 Trap FPEs in GNU debug build.
1bfb7ca Merge branch 'linux.intel.fix'
c41505f Remove deprecated options in linux-intel.mk
fd35f5d Update LICENSE.md
a5647dc More POSIX conformance, ref hash in no repo
3ffb5ff Corrected the install steps
e84eae3 Add Install section to README.module
c2f56f7 Add documents and templates
02c4534 Moved mkmf to bin dir
f92fdc9 Doc and other stuff added.
6e7a1fb Import of bronx-7 mkmf

I suspect that only commits on branches are being downloaded during the clone.

@underwoo For the interim, I think we need you to restore the branch since tagged models point to it. I'll make a list of any commits you asked us to use that are not on "master". Somehow we need to find a way to get that commit into the history, perhaps by reverting it on the find.symlinks branch and then merging the branch as a no-op?

[underwoo]

Commit/branch restored.

[jkrasting]

Back to normal again. Thanks @underwoo and @adcroft! -- ----- Dr. John Krasting, Physical Scientist Biogeochemistry, Ecosystems, and Climate Group NOAA/Geophysical Fluid Dynamics Laboratory 201 Forrestal Road, Princeton, NJ 08540 P. (609) 452-5359 | F. (609) 987-5063