Issue 189

[o-smirnov]

Please review and merge #207 first, as this is based off that branch.

Adds bindings of non-current paths into singularity, fixing #189 which tripped up @Athanaseus a few times.

Also adds access_parent_dir and write_parent_dir fields to Parameter. I quickly realized this was needed for "funny" stuff like CASA flagman (because if it's accessing foo/bar.MS, it also needs access to foo/ itself, so it can create the flagversions table alongside the MS).

[o-smirnov]

Do not merge yet -- see a potential bug.

[o-smirnov]

OK seems fine now. But let my recipe run on ilifu for a few more steps to make sure the various directory bindings work as intended.

[SpheMakh]

Could you also add this test; the file is empty? tests/stimela_tests/test_slurm_config.yml

[o-smirnov]

Could you also add this test; the file is empty? tests/stimela_tests/test_slurm_config.yml

It's empty by design. The user can replace it with a non-empty config file to specify non-default slurm options. We don't have an _incude_if_exists statement, so it's easier to have an empty file.

[o-smirnov]

A few more small fixes incoming... let me run an E2E workflow on ilifu before we merge (and maybe wait for @Athanaseus to finish his test on the pleiades cluster).

[o-smirnov]

OK, I'm happy with how it works. All filesystems appear to be mounting correctly. @Athanaseus or @SpheMakh please review and merge. I'll start other fixes on a branch off this one.

The merge-base changed after approval.

The merge-base changed after approval.

[o-smirnov]

@Athanaseus looks like you reviewed, then cancelled your review -- what's up?

The merge-base changed after approval.

[Athanaseus]

When I approve, it does exactly that. Not sure why.
@Athanaseus dismissed review on this pull request.

[Athanaseus]

ok i see now, a file changed, and GitHub won't let me approve without reviewing it.

The merge-base changed after approval.

[o-smirnov]

That's odd, my last commit was yesterday, so why is it dismissing your review today? Could you please try approving again?

[Athanaseus]

LGTM!

The merge-base changed after approval.

[o-smirnov]

OK I think the review cancellation is a github bug. I'll force-merge.