Prompt Service

The nsIPromptService provides methods for opening various types of prompts. See the interface documentation for a list of prompt types. Every prompt method has 3 different versions:

  • Prompt by window: This is considered the legacy way of prompting and only works if the window you want to prompt for is in the same process. Only supports window prompts.

  • Prompt by browsing context (synchronous): Use a browsing context as parent for the prompt. Works cross process from parent and content process.

  • Prompt by browsing context (asynchronous): Returns a promise which resolves once the prompt closes.

The synchronous prompt methods use call by reference (XPCOM inout or out) to return the updated prompt arguments to the caller. When prompting async the arguments are passed in by value. Prompt results are returned in an nsIPropertyBag when the Promise resolves.

Note

If you don’t provide a parent window or browsing context the prompt service will fallback to a window prompt. The same goes for browsing contexts of chrome windows, because there is no clear association to a browser / tab.

Examples

JavaScript Sync

Here is an example of opening a confirm prompt from JavaScript. We are in the parent process and we want to show a tab prompt:

// Get the browsing context for the currently selected tab
let browsingContext = gBrowser.selectedBrowser.browsingContext;

// Specify prompt type, can be MODAL_TYPE_TAB, MODAL_TYPE_CONTENT,
// MODAL_TYPE_WINDOW
let modalType = Services.prompt.MODAL_TYPE_TAB;

// Object for checkbox state to pass by reference.
let check = { value: false };

// Prompt synchronously and store result
let confirmed = Services.prompt.confirmCheckBC(browsingContext, modalType,
"My Title", "Hello, World!", "Check this box if you agree", check);

// check.value now contains final checkbox state
// confirmed is a boolean which indicates whether the user pressed ok (true)
// or cancel (false)
console.debug("User checked checkbox?", check.value);
console.debug("User confirmed prompt?", confirmed);

JavaScript Async

The same prompt as above, but called async:

// Get the browsing context for the currently selected tab
let browsingContext = gBrowser.selectedBrowser.browsingContext;

// Specify prompt type, can be MODAL_TYPE_TAB, MODAL_TYPE_CONTENT,
// MODAL_TYPE_WINDOW
let modalType = Services.prompt.MODAL_TYPE_TAB;

// Note that the checkbox state variable is not an object in this case,
because we get the checkbox result via the result object.
let check = false;

// Prompt asynchronously and await result
let propBag = await Services.prompt.asyncConfirmCheck(browsingContext,
                                                      modalType,
                                                      "My Title",
                                                      "Hello, World!",
                                                      "Check this box if you agree",
                                                      check);

let ok = propBag.getProperty("ok");
let checked = propBag.getProperty("checked");

// ok is the boolean indicating if the user clicked "ok" (true) or
// "cancel" (false).
// checked is a boolean indicating the final checkbox state
console.debug("User checked checkbox?", checked);
console.debug("User confirmed prompt?", ok);

C++ Sync

nsCOMPtr<nsIPromptService> promptSvc =
  do_GetService("@mozilla.org/prompter;1");

if(!promptSvc) {
  // Error handling
  return;
}

// Assuming you have the browsing context as a member.
// You might need to get the browsing context from somewhere else.
BrowsingContext* bc = mBrowsingContext;

bool ok;
bool checked = false;

nsresult rv = promptSvc->confirmCheck(mBrowsingContext,
                                      nsIPromptService::MODAL_TYPE_TAB,
                                      "My Title"_ns
                                      "Hello, World!"_ns,
                                      "Check this box if you agree"_ns,
                                      &checked, &ok);

// ok is the boolean indicating if the user clicked "ok" (true) or
// "cancel" (false).
// checked is a boolean indicating the final checkbox state

C++ Async

nsCOMPtr<nsIPromptService> promptSvc =
  do_GetService("@mozilla.org/prompter;1");

if(!promptSvc) {
  // Error handling
  return;
}

bool checked = false;
Promise* promise;

// Assuming you have the browsing context as a member.
// You might need to get the browsing context from somewhere else.
BrowsingContext* bc = mBrowsingContext;

// As opposed to the sync case, here we pass the checked flag by value
nsresult rv = promptSvc->confirmCheckAsync(mBrowsingContext,
                                           nsIPromptService::MODAL_TYPE_TAB, "My Title"_ns,
                                           "Hello, World!"_ns,
                                           "Check this box if you agree"_ns,
                                           checked, promise);

// Attach a promise handler
RefPtr<PromptHandler> handler = new PromptHandler(promise);
promise->AppendNativeHandler(handler);

Then, in your promise handler callback function:

void PromptHandler::ResolvedCallback(JSContext* aCx,
                                 JS::Handle<JS::Value> aValue) {
  JS::Rooted<JSObject*> detailObj(aCx, &aValue.toObject());

  // Convert the JSObject back to a property bag
  nsresult rv;
  nsCOMPtr<nsIPropertyBag2> propBag;
  rv = UnwrapArg<nsIPropertyBag2>(aCx, detailObj, getter_AddRefs(propBag));
  if (NS_FAILED(rv)) return;

  bool ok;
  bool checked;
  propBag->GetPropertyAsBool(u"ok"_ns, &ok);
  propBag->GetPropertyAsBool(u"checked"_ns, &checked);

  // ok is the boolean indicating if the user clicked "ok" (true) or
  // "cancel" (false).
  // checked is a boolean indicating the final checkbox state.
}

For a full list of prompt methods check nsIPromptService reference.