Better handle proposed changes to documents belonging to closed working groups

[ietf-svn-bot]

owner:jennifer@painless-security.com resolution_fixed type_defect | by barryleiba@computer.org

---

1. Attempts to post a -00 document for a closed WG should be denied outright.
2. Attempts to post a > -00 document for a closed WG should go to the area's current ADs for approval.
3. Attempts to flag an individual document as replacing a document in a closed WG should go to the area's current ADs for approval.
4. Attempts to flag an individual document as replacing a document in an active WG should go to the WG's chairs for approval.

---

Issue migrated from trac:3058 at 2022-03-04 07:58:13 +0000

[ietf-svn-bot]

@rjsparks@nostrum.com edited the issue description

[ietf-svn-bot]

@rjsparks@nostrum.com edited the issue description

[ietf-svn-bot]

@rjsparks@nostrum.com changed status from new to accepted

[ietf-svn-bot]

@jennifer@painless-security.com changed status from accepted to assigned

[ietf-svn-bot]

@jennifer@painless-security.com changed owner from `` to jennifer@painless-security.com

[ietf-svn-bot]

@jennifer@painless-security.com commented

---

I have a few questions after looking in to this a bit:

1. Is a new DraftSubmissionStateName needed (or likely to be needed) to represent the state where AD approval is required? Or can the grp-appr be reused, just using ADs in place of WG chair when the group is closed?

2. Does the grp-appr state currently work as designed? Tracing through the code, I am puzzled. In submission_status.html, there is a notice that indicates the "Post" button will notify group chairs to get approval when requires_group_approval is True. However, the code in the submission_status() view seems to directly post the draft if the submitter is logged in and is an author. The grp-appr state is only entered if the submitter is not a logged-in author - in that case, it is set in send_confirmation_emails().

3. I believe posting a new -00 document for a closed WG is already denied - is that possible, or am I looking in the wrong place?

4. Does "flag an individual document as replacing..." mean selecting a replacement document on the submission_status() view? Should this also affect setting it as a replacement by editing a posted draft's "replaces" property?

[ietf-svn-bot]

@rjsparks@nostrum.com commented

---

1. Yes, please create and use a new state for AD approval needed (I suggest 'ad-appr').

2. If there's not a test that proves it, please create one? I think, though that it is working as designed - the place where grp-appr gets set in the workflow is in ietf/submit/utils line 800. It may be that this decision should get refactored - having it be a side-effect of sending_confirmation_emails is non-intuitive.

3. It could be, but is there a test that proves it?

4. I'm not sure I follow the question. The concept here is a new one fwiw. The idea is that if someone submits a new individual draft and suggests that it replace an active wg draft, no 'replaces' RelatedDocument object would get created when the individual draft is posted even if the submitter was an author of both drafts. Instead, the chairs would get nudged to evaluate the situation and manually add the replaces relationship. I'm not sure I remember the current state of plumbing that handles "somebody needs to evaluate and confirm these replaces relationships" when they are supplied on the submission form in general - I suspect there's not much. If there's not something already there to leverage, I think just emailing the suggestion to the chairs is enough.

[ietf-svn-bot]

@jennifer@painless-security.com commented

---

Thank you for the replies (and Barry, who replied via email). Some of the issue here stems from not being familiar with the procedure being implemented by the code.

For 2), to confirm: the desired behavior is that the author of the previous (or current for a -00) can flag as a replacement without review, but a replacement marked by a non-author will be reviewed by the group chair. Is that correct?

For 3), there's no test, and the test Barry suggested by email (uploading a -00 draft for a closed group) did result in a rejection. This clears it up - the description was not requesting a change, just stating the complete behavior. Thanks.

For 4), what I was trying to ask is whether editing an existing posted doc (the 'Edit' button by the 'Replaces' field) to add a replacement should be subject to the same review as during the initial submission. I think the answer is no because this edit capability is only available to secretariat / ADs / document managers. (Maybe an email should be sent?)

[ietf-svn-bot]

@jennifer@painless-security.com commented

---

One more thing: does this only apply to groups of type 'WG', or to any group in a closed state?

And, er, which group states are closed? I'd guess:

closed: abandon, conclude
not closed: active, bof, proposed

what about bof-conc, dormant, unknown, and replaced? I'd guess these are all closed, but I'm not sure how they're used.

[ietf-svn-bot]

@rjsparks@nostrum.com commented

---

Yes, this is only for groups with type_id=='wg'. Stated another way, this is the policy the IESG is setting for documents in the IETF stream that go through the working group process (specifically, drafts that would have their group relationship pointing to a group type_id=='wg').

For closed, use "not in ('active','bof','proposed')"

The dormant state is not often used, and is unlikely to be used with groups of type wg. dormant groups are not closed, but are also not expected to be doing active work. At the moment we only have one dormant group ('mentor', which is of type 'team'). But were a wg made dormant, the submission behavior described above for closed groups makes the most sense.

bof-conc is used when a group starts as a bof and is concluded with nothing else happening.

'replaced' is used when working groups are reorganized and replaced with another working group. Submissions behavior here could include the chairs of the groups that replaced this group, but this is enough of a rare condition that the behavior for closed groups is fine - the AD can make sure things land in the right place.

'unknown' is a state used with modeling pre-tracker groups - there are none that are type 'wg', and we will not have an 'wg' groups in state 'unknown' going forward.

[ietf-svn-bot]

@rjsparks@nostrum.com commented

---

Replying to ietf-svn-conversion/datatracker#3058 (comment:7):

Thank you for the replies (and Barry, who replied via email). Some of the issue here stems from not being familiar with the procedure being implemented by the code.

For 2), to confirm: the desired behavior is that the author of the previous (or current for a -00) can flag as a replacement without review, but a replacement marked by a non-author will be reviewed by the group chair. Is that correct?

For 3), there's no test, and the test Barry suggested by email (uploading a -00 draft for a closed group) did result in a rejection. This clears it up - the description was not requesting a change, just stating the complete behavior. Thanks.

For 4), what I was trying to ask is whether editing an existing posted doc (the 'Edit' button by the 'Replaces' field) to add a replacement should be subject to the same review as during the initial submission. I think the answer is no because this edit capability is only available to secretariat / ADs / document managers. (Maybe an email should be sent?)

At 2) - No, that does not describe the desired behavior. Lets stop saying "flag as", it's not precise enough. Instead, we're talking about what to do with suggestions made for replacement relationships provided by the submitter of a draft revision (including an initial one). The question is when such suggestions are automatically accepted. The relevant code (as you've already discovered) is in ietf.submit.utils.update_replaces_from_submission. It does appear to currently accept the suggestion of some who is in the author list of the replaced document, and that should change if the replaced document belongs to a wg (open or closed) as above.

At 3) - please add a test.

At 4) - There's not a need for a change of behavior here - as you note, the actions are only available to people with "do anything" privileges in the datatracker (at least as far as managing documents of type 'draft' goes), and email about such changes is already sent.

[ietf-svn-bot]

@jennifer@painless-security.com commented

---

Ok, I think this is starting to be clearer.

At 2), it sounds like the desired behavior when a draft revision includes a suggested replacement is...

• For an individual draft suggested to replace an individual draft, previous author can approve. This is automatic if the submitter is the author, draft goes to aut-appr state if not.
• For any draft (individual or belonging to a WG) suggested to replace a draft belonging to a WG, the WG chair (or higher) must approve. This is automatic if the submitter is the WG chair (or higher). If not, the draft goes to grp-appr if the WG is open or to ad-appr if the WG is closed.

Am I on the right track? I believe this is consistent with the original description (items 3 and 4 in the ticket). The main thing I'm unsure of is the statement of an "individual document" replacing a WG document. Does there need to be a distinction between submitted drafts with group type individ vs those with wg groups as far as this logic goes?

Also, I'll add tests for this and my question 3.

[ietf-svn-bot]

@jennifer@painless-security.com changed _comment0 which not transferred by tractive

[ietf-svn-bot]

@rjsparks@nostrum.com commented

---

Your description looks correct to me. And individual document means "not a working group document", which is modeled in the datatracker as belonging to the special group with the acronym of 'none'. That is the only group of type 'individ'. Please use acronym=='none' to identify these.

[ietf-svn-bot]

@rjsparks@nostrum.com commented

---

Well, to be closer to accurate - instead of "not a working group document" I should have said "not a formal group document" or something like it, or have skipped the phrase.

From a dump earlier this week:

In 5af8d2d095c84918a92c7038da20167cb59f861c: Counter(Document.objects.filter(type='draft').values_list('group__type_
    ...: _slug',flat=True))
Out5af8d2d095c84918a92c7038da20167cb59f861c:
Counter({'individ': 24883,
         'wg': 10108,
         'ietf': 163,
         'rg': 248,
         'ag': 21,
         'area': 1345,
         'iana': 1})

This ticket is only concerned with documents of types 'individ' and 'wg'