Looking for help from the community to understand some image question behavior.
My survey works on four layers. One of those layers is a point layer with attachments enabled ('Photopoints'), which is a nested repeat (1:M). When surveys are edited from the inbox (not directly from the inbox, but through a URL parameter provided in an application) the photo is 'present' but not, and there's a little cloud symbol that appears (see below). It looks like the survey pulls the photo down so the user can see it.
My issue: I have a constraint in the survey that forces the user to upload a photo if a condition is met (it operates by counting the responses to the image question). In this case, the requirement is met, as a photo and and its necessary dependent information was submitted, but something else in the survey needs to be edited, yet the survey can no longer 'count' the responses to the image question since it's not really 'there.' Is there a workaround so the constraint can stay in place (a photo must be submitted for certain kinds of records) but to have it count the attached features as well?
What if you make a hidden calculated questions and count that instead of the photo question. Have it populate when the repeat is created, or when the initial photo is taken. That way it is relying on the photo question but not the photo itself.
Found a temporary workaround here (LINK) using the pulldata and 'mode.' At least I can turn the constraint off when edited. This is still not the full extent of the preferred behavior, and a better move would be to be able to count attachments in a given layer if the survey mode is in 'edit.'