Pulp

This is running now:

import redmine

redmine_api_key = open('redmine-verifier-key').read().strip()

redmine_api = redmine.Redmine('https://pulp.plan.io', key=redmine_api_key)

def url(id):
    return 'https://pulp.plan.io/issues/{}'.format(id)

def filter_issue(issue):
    # if the issue isn't a story or task, skip it
    # can only filter issues by one tracker type, so filter here
    if issue.tracker.id not in (tracker_i, tracker_s):
        return True

    # if the issue is already verified, skip it
    # .filtering by the value of the "Verified" custom field seems broken, so filter here
    for field in issue.custom_fields:
        if field.id == field_v and field.value == "1":
            return True

    return False

# these resources don't support filtering, so ghettofilter out the things we want
for status in redmine_api.issue_status.all():
    if status.name == 'CLOSED - CURRENTRELEASE':
        # ccr = closed current release
        status_ccr = status.id
    elif status.name == 'VERIFIED':
        status_v = status.id

for tracker in redmine_api.tracker.all():
    if tracker.name == 'Issue':
        tracker_i = tracker.id
    elif tracker.name == 'Story':
        tracker_s = tracker.id

for custom_field in redmine_api.custom_field.all():
    if custom_field.name == 'Verified':
        field_v = custom_field.id

# filter outs issues that aren't current CLOSED - CURRENTRELEASE, and aren't marked verified
filter_kwargs = { 
    'status_id': status_ccr,
}
closed_issues = redmine_api.issue.filter(**filter_kwargs)

# known closed issue for debugging
# closed_issues = [redmine_api.issue.get(161)]
errors = []

for issue in filter(filter_issue, closed_issues):
    # re-get the issue with journals to get status history
    issue = redmine_api.issue.get(issue.id, include='journals')

    print('processing {}'.format(url(issue.id)))

    for issue_change in reversed(issue.journals):
        for detail in issue_change.details:
            if detail['name'] == 'status_id':
                old_value, new_value = (int(v) for v in (detail['old_value'], detail['new_value']))
                break

        try:
            # these might not get set in the detail loop, so go around again if we have to
            old_value, new_value
            break
        except NameError:
            # heh
            continue
    else:
        # this should never happen
        print(' Unable to determine last status change for {}?'.format(url(issue.id)))

    if old_value == status_v and new_value == status_ccr:
        # issue's last status transition was from VERIFIED to C - CR, mark verified
        print(' {} marked verified.'.format(url(issue.id)))

        # actually do the thing
        try:
            redmine_api.issue.update(issue.id, custom_fields=[{'id': field_v, 'value': '1'}])
        except redmine.exception.BaseRedmineException as ex:
            print(' {} error on update'.format(url(issue.id)))
            errors.append((issue, ex))
    else:
        print(' {} not marked verified.'.format(url(issue.id)))

    # a newline for good luck
    print()

for issue, ex in errors:
    print('{} raised {} in processing and needs manual intervention'.format(
        issue.id, type(ex).__name__)
    )

As you might imagine, the "redmine_api_key" file contains a redmine api key. :)

The script should be idempotent, which is useful in case we need to run this again during the transition from using the VERIFIED status to using the Verified custom field.

(Edited Feb 2 with minor script tweaks.)

The script failed with a ValidationError on #792, so I'll tweak it and try again later.

This is very nice. I was not aware of the journals option. Thank you for this. Extra points for having the user reflect it was done w/ automation by using the pulpbot name.

Thanks :)

I updated the script to continue if errors occur, and log them after the fact. I also tweaked some of the filtering logic to prevent hitting the API to grab the journals before filtering units out based on info that was already available.

As of now, issues from 1-2564 have been processed and retroactively verified where appropriate. Since the process docs have not yet been updated (#2565) to reflect the proper usage, I plan to leave this open and run it again once the process is "officially" changed.