# r34 save controls and feedback

The analysis settings and literature associations for a private workflow belong
to the same saved workflow revision. Use **Save analysis and literature settings**
to save the analysis/reporting settings, selected literature and each document's
role together. The former two save buttons are replaced by this single action.
Saving these settings does not regenerate an existing report; use the report's
regeneration action when a new report is required.

Save controls identify the form they save and provide explicit feedback:

- **Unsaved changes** means the current form differs from the last saved values.
- **Saving…** means the save request is in progress. Repeated clicks do not create
  duplicate concurrent saves.
- **Saved** means that the application accepted the displayed values for that
  form. Editing the form again returns it to **Unsaved changes**.
- A failed save shows an error and keeps nonsensitive edits available to correct
  or retry. Credential and secret fields are cleared and must be entered again. Values from an old account, project or workflow cannot mark the newly
  selected scope as saved.

A button's blue background identifies a primary action. It does not
mean that the settings are saved or that the button is selected. Read the save
status beside the relevant form. Settings forms across the standalone workbench,
assistant and hosted account/storage controls follow this convention.

Downloading a configuration or exporting a report view is a separate action.
Its message confirms that the browser download was started; the application
cannot confirm completion of the browser's file dialog or a write to the user's
disk. Exported report view settings describe presentation choices and do not
alter simulation results or scientific evidence.

This remains release **r34**. The focused receipt in
`validation_r34_save_controls/` records the explicit tests and runtime listed in
`release_tools/validate_r34_save_controls.py`. Browser checks exercise local
standalone and hosted application servers. Native Windows execution, production
storage and live-provider scientific evaluation are not established by these
checks. Scientific qualification statuses and pending external work are retained.

The prior upload-maintenance receipt (293 Python cases and five JavaScript
commands) and autonomous receipt (1,495 Python parent cases, 130 subtests and four
JavaScript commands) remain unchanged historical evidence, each bound to its
original source. They are not included in this release's fresh focused count.
Every original archive member is retained. The original release maps and scoped
source originals are preserved in
`validation_r34_save_controls/original_files/`.

Verify the extracted delivery with:

```sh
python release_tools/validate_r34_save_controls.py verify --package
```