Feature request: README file pointing from bld/run back to case directory #369
Labels
Comments
|
fixed by #739 |
pesieber
pushed a commit
to pesieber/cime
that referenced
this issue
Mar 15, 2023
Now that we're doing interpolation pretty much all the time, we are covering the cases that these originally tested. Two examples are: - Crop to non-crop: covered by ERP_Ld5.f10_f10_musgs.I1850Clm50Bgc.cheyenne_gnu.clm-default1850 (uses finidat clmi.I1850Clm50BgcCrop.1366-01-01.0.9x1.25_gx1v6_simyr1850_c180424.nc) - Non-crop to crop: covered by SMS_Ld5.f10_f10_musgs.I1850Clm45BgcCrop.cheyenne_gnu.clm-crop1850 (uses finidat clmi.I1850Clm45BgcGs.0901-01-01.0.9x1.25_gx1v6_simyr1850_c180204.nc) In addition, we cover the mindist logic via unit test. What's not covered by unit test is determining if all variables are initialized reasonably - particularly important for the non-crop to crop case. But I'm not sure how we can really be confident of that even if we have a system test covering this (other than the standard way of seeing that baseline comparisons haven't changed... but the issue is: if new behavior is introduced that changes answers for crops, then we could expect baseline comparisons to change, and so could miss some issue). That is to say: I'm not sure it's worth having system tests that explicitly test these cases: it's probably enough to have the unit tests plus incidental coverage via existing system tests. Resolves ESMCI#369
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This is low priority; I'm mainly opening this issue to start a discussion of whether others would find this useful, and if so, what it should look like.
In debugging someone else's case, I often know where their run directory is, but not their case directory or source code directory. The only way I know to find these is to find a Filepath file in one of the bld directories; this works, but it's awkward.
It would be helpful if there were a README-paths file (or something else) in the bld and/or run directories pointing to the $CASEROOT and $SRCROOT for this case. Maybe there are some other paths that would be useful while we're at it, like $DOUT_S_ROOT (of course, you can find this out if you know $CASEROOT, but it could be convenient to have this directly available from the run directory).
At first I was thinking that this file should be a sibling of bld and run, so we'd have:
/glade/scratch/sacks/mycase
README
bld/
run/
But I don't know if that will work, since bld and run can theoretically be in different paths. So maybe it's best for this to just go in both the bld and run directories? Or just the run directory? In that case, it might also be useful to have a pointer to $EXEROOT in the README in the run directory.
The text was updated successfully, but these errors were encountered: