Running down a test failure in conda-forge/libnetcdf-feedstock#154, ncdump/tst_bom.sh and ncdump/tst_nccopy4.sh failed, with odd diffs. They were running concurrently, and it turns out the problem is that both test scripts write temporary output to tmp.cdl.
I'll also note that tmp.cdl is mentioned in ncdump/run_back_comp_tests.sh (but commented out). The file is also removed (but not created?) in ncdump/tst_nccopy5.sh.
There are also some occurrences in nczarr_test/run_nccopyz.sh and nczarr_test/run_nczarr_fill.sh. I'm not sure if that's potentially problematic or not.
Running down a test failure in conda-forge/libnetcdf-feedstock#154,
ncdump/tst_bom.shandncdump/tst_nccopy4.shfailed, with odd diffs. They were running concurrently, and it turns out the problem is that both test scripts write temporary output totmp.cdl.I'll also note that
tmp.cdlis mentioned inncdump/run_back_comp_tests.sh(but commented out). The file is also removed (but not created?) inncdump/tst_nccopy5.sh.There are also some occurrences in
nczarr_test/run_nccopyz.shandnczarr_test/run_nczarr_fill.sh. I'm not sure if that's potentially problematic or not.