[DBWG] DBWG-2: proposal to auto-generate contents of the mandatory "changed" field in db objects.
benm at workonline.africa
Wed Aug 5 11:41:24 UTC 2020
On 08/05, Michel ODOU wrote:
> Hi Frank,
> On 05/08/2020 13:22, Frank Habicht wrote:
> > my personal (non-chair) opinion is that this is dangerous.
> > Can we get consensus that this should not be allowed?
> > I believe all creations and updates of domain objects should enforce
> > that there is a mnt-by.
> The mnt-by attribute is mandatory since June 2014. It is not possible to
> create/update a domain object without a mnt-by attribute. However, we
> still have many unprotected domain objects in the database that have not
> changed since.
So, we need ideas for safely backfilling those old objects. Anyone?
> > From below I understand that there was no enforcement that later
> > creations/updates also had maintainers for these objects, correct?
> There is no enforcement that later *updates* have a mnt-by attribute for
> person and role objects because, as explained in my previous mail, the
> mnt-by attribute remains optional on these objects, for historical
> reasons (you need a person or a role to create a maintainer because
> admin-c is mandatory - if at the same time you need a mntner to create
> the person/role object, then we have a cyclic dependency. a way to break
> it is to make the mnt-by optional in the person and role objects).
I find it a little strange that anyone would choose to make mnt-by
optional, rather than admin-c. But that's not the issue.
Surely the solution is simply both be created, if necessary, in the same
I know WHOIS is old and clunky, but that can't be all that hard?
> However, since September 2017, when a person or role object is *created*
> without a mnt-by attribute, the WHOIS will generate one for the user.
> This still applies.
But yet, mnt-by can be subsequently removed, by accident?!
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 833 bytes
Desc: not available
More information about the DBWG