I don't know if there is any other file you need to edit when adding a config option: some JSON schema maybe?

There is a JSON schema somewhere, I can probably dig it up if you can't find it. Look for previous PRs that added new config options.

But is there any way we can try-except this, or check if we're on AFS instead of relying on manual user intervention?

There doesn't seem to be an easy reliable way of detecting AFS (or, in general, telling what filesystem is for a given path). As for try-except: it's a valid idea, but will make the code less readable because of nested try-except:

try:
    chmod(perms_with_setgid)
except PermissionError:
    try:
        chmod(permis_without_setgid)
    except PermissionError as e:
        raise e

An alternative solution could be creating a "probe" directory at the start of spack install, trying to set S_ISGID flag, and remembering the result - but that is a very extensive change IMO.

Why is Spack setting the setgid bit? On some filesystems it's used to control "BSD" vs "ATT" group semantics:

When the bit is set for a directory, the set of files in that directory will have the same group as the group of the parent directory, and not that of the user who created those files. This is used for file sharing since they can be now modified by all the users who are part of the group of the parent directory.

from here on geeksforgeeks

but those permissions are inherited from the parent dir, so if someone wants those semantics then they should be able to get them by setting the SETGID bit on the parent directory.

Is there something else going on?

@hartzell @becker33 any progress?

@iarspider -- I'm not a doer on this, just was trying to clear up my confusion about how it was being used.

Ping @becker33 @scheibelp !

Bouncing close/open to restart tests