Two failing tests in the build-namelist tester since ctsm5.3.014
Brief summary of bug
UPDATE also posted below: The tests that started failing with ctsm5.3.032 have been resolved.
These two tests start failing with ctsm5.3.032
3200/3264 < FAIL> <Test Id: 3200> <Desc: -res 0.9x1.25 -mask gx1v7 -sim_year 1850 -envxml_dir . -lnd_tuning_mod clm5_0_CRUJRA2024 -bgc sp -phys clm5_0: lnd_in file exists>
3203/3264 < FAIL> <Test Id: 3203> <Desc: -res 0.9x1.25 -mask gx1v7 -sim_year 1850 -envxml_dir . -lnd_tuning_mod clm5_0_CRUJRA2024 -bgc bgc -phys clm5_0: lnd_in file exists>
NOTE: Below we realize there's an additional one: (failing at least by ctsm5.3.069)
1220/3394 < FAIL> <Test Id: 1220> <Desc: inventoryfileDNE>
For a total of three tests that fail
General bug information
CTSM version you are using: ctsm5.3.032
Does this bug cause significantly incorrect results in the model's science? No
Configurations affected: f09_gx1v7 for 1850 with clm5_0 and CRUJRA2024
Details of bug
New tests added in ctsm5.3.032 fail.
This may also indicate a problem for cases with this setup as well, but it's unclear.
Important output or errors that show the problem
# Failed test '-res 0.9x1.25 -mask gx1v7 -sim_year 1850 -envxml_dir . -lnd_tuning_mod clm5_0_CRUJRA2024 -bgc sp -phys clm5_0: lnd_in file exists'
# at NMLTest/CompFiles.pm line 60.
/glade/derecho/scratch/erik/ctsm5.3.032/bld/unit_testers/lnd_in does NOT exist
/bin/cp: cannot stat '/glade/derecho/scratch/erik/ctsm5.3.032/bld/unit_testers/lnd_in': No such file or directory
-res+0.9x1.25+-mask+gx1v7+-sim_year+1850+-envxml_dir+.+-lnd_tuning_mod+clm5_0_CRUJRA2024+-bgc+sp namelists for -phys+clm5_0
/bin/cat: lnd_in.-phys+clm5_0.-res+0.9x1.25+-mask+gx1v7+-sim_year+1850+-envxml_dir+.+-lnd_tuning_mod+clm5_0_CRUJRA2024+-bgc+sp: No such file or directory
ERROR : CLM build-namelist::CLMBuildNamelist::setup_logic_initial_conditions() : use_init_interp is NOT synchronized with init_interp_attributes in the namelist_defaults file, this should be corrected there
# Failed test '-res 0.9x1.25 -mask gx1v7 -sim_year 1850 -envxml_dir . -lnd_tuning_mod clm5_0_CRUJRA2024 -bgc bgc -phys clm5_0: lnd_in file exists'
# at NMLTest/CompFiles.pm line 60.
/glade/derecho/scratch/erik/ctsm5.3.032/bld/unit_testers/lnd_in does NOT exist
/bin/cp: cannot stat '/glade/derecho/scratch/erik/ctsm5.3.032/bld/unit_testers/lnd_in': No such file or directory
-res+0.9x1.25+-mask+gx1v7+-sim_year+1850+-envxml_dir+.+-lnd_tuning_mod+clm5_0_CRUJRA2024+-bgc+bgc namelists for -phys+clm5_0
/bin/cat: lnd_in.-phys+clm5_0.-res+0.9x1.25+-mask+gx1v7+-sim_year+1850+-envxml_dir+.+-lnd_tuning_mod+clm5_0_CRUJRA2024+-bgc+bgc: No such file or directory
ERROR : CLM build-namelist::CLMBuildNamelist::se
I'm also seeing another test fail: not ok 1202 - inventoryfileDNE. It looks like maybe this is failing because there are two unit test cases with the same name:
https://github.com/ESCOMP/CTSM/blob/37204fbc251cd3077dc3751628b75c25d95cf1a0/bld/unit_testers/build-namelist_test.pl#L1133-L1148
Note I came upon this testing #3372 and confirmed it is present in ctsm5.3.069.
I'm also seeing another test fail:
not ok 1202 - inventoryfileDNE. It looks like maybe this is failing because there are two unit test cases with the same name: [...]
I have traced this one back to ctsm5.3.014 (.013 passes). I changed the second instance of the repeating testname to useFATESLUH2fileDNE and found that now both fail:
not ok 1092 - useFATESLUH2fileDNE
[...]
not ok 1103 - inventoryfileDNE
I will update the title of this issue from "Three failing tests" to "Four failing tests"
ctsm5.3.032 is documented in #2956 and a linked CDEPS PR ctsm5.3.014 is documented in #2888 and numerous linked PRs
#3501 resolves the two failures associated with #2956 that came in with ctsm5.3.032.