There aren't currently tests for the image widgets (widgets descending from BaseImageWidget) that ensures they display the correct image based on the file path in a prompt. We now have tests that they display an image default however (#3642) so we should add the "normal" case to these tests as well.

Some suggestions to placeholders:

• A placeholder to show Player Caused Deaths
• A placeholder to show Mob K/D
• A placeholder to show Mob Caused Deaths
• A placeholder to show recent player kills (ex: player_recent_kill_<1-10>)
• A placeholder to show recent deaths (ex: player_recent_death_<1-10>)

A list of all available placeholders on the github wiki.

Describe the bug:
Insight graph download buttons have two issues:

1. They are unclear and the arrows are all over the place (little blue ones)
2. They do not actually "do" anything.

To Reproduce:
Steps to reproduce the behavior:

1. augur.osshealth.io
2. select "insig

wq templates are not designed to work with Django form classes, but this is not necessarily clear from the documentation.

• Django REST Framework serializers handle processing and validating form submissions.
• Mustache HTML templates handle rendering and displaying forms. Unlike Django forms, the field

• Stetho for SQLite and UI layout debugging
• Rx debugging methods

Using the same underlying value for multiple choices is spec compliant, but it's usually a mistake. We should warn users.

Original issue: https://forum.opendatakit.org/t/problem-with-underlying-value-untitled/18984

Sentry can use refs to git commits to enhance the error reports we get when something fails. For that to work, Sentry needs to receive some information:

• a list of refs belonging to each release
  or
• the most recent commit of the release
  or
• the most recent commit of the release and the most recent commit of the previous release

[Sentry docs with more info](https://docs.sentry

See screenshot:

There are a number of places where JavaRosa throws RuntimeExceptions. That can make it hard for clients to differentiate between exceptions that came from JavaRosa and exceptions that came from elsewhere. As a result, Collect does awkward things like catch RuntimeExceptions and wrap them in its own JavaRosaException type: https://github.com/opendatakit/collect/blob/master/collect_app/src/mai

Software and hardware versions

Android vx.x.x, device used...

Problem description

When transferring blank forms if all are already available at the receiver, then the resulting dialog is blank

Steps to reproduce the problem

Send blank forms which are already available in the receiver

Expected behavior

Result dialog should display that Forms are already available

In CoughRecorderScreen.tsx

https://opendatakit.github.io/xforms-spec/client-audit-logs says "node: for events that refer to a specific node, the XPath path of that node". Add something like "In the case of repeats, position predicates are always included even if only one repeat instance currently exists."

Once we add a signature if we press edit, we are taken back to the SignatureActivity and the signature and text are displayed. The save button is enabled before we have made any changes.

Expected:
Save button is disabled until we make any changes.
If the drawing is empty we should disable the clear button.

Currently the class and field name generation assumes repeat groups will always be expressed in plural form (ending with an "s"). Figure out a way to refactor this so the plurality (and specifically "s") is optional.

Currently, the sign in window on WPF looks peculiar on high-DPI displays. E.g.

Microsoft has described how to fix this issue: https://docs.microsoft.com/en-us/windows/communitytoolkit/controls/wpf-winforms/webview#high-d