Method handleAction

agent/x/agentdesktop/api.go:131–531

(rw http.ResponseWriter, r *http.Request)

}

func (a *API) handleAction(rw http.ResponseWriter, r *http.Request) {
	ctx := r.Context()
	logger := a.logger.With(agentchat.Fields(ctx)...)
	handlerStart := a.clock.Now()

	// Update last desktop action timestamp for idle recording monitor.
	a.desktop.RecordActivity()

	// Ensure the desktop is running and grab native dimensions.
	cfg, err := a.desktop.Start(ctx)
	if err != nil {
		logger.Warn(ctx, "handleAction: desktop.Start failed",
			slog.Error(err),
			slog.F("elapsed_ms", a.clock.Since(handlerStart).Milliseconds()),
		)
		httpapi.Write(ctx, rw, http.StatusInternalServerError, codersdk.Response{
			Message: "Failed to start desktop session.",
			Detail:  err.Error(),
		})
		return
	}

	var action DesktopAction
	if err := json.NewDecoder(r.Body).Decode(&action); err != nil {
		httpapi.Write(ctx, rw, http.StatusBadRequest, codersdk.Response{
			Message: "Failed to decode request body.",
			Detail:  err.Error(),
		})
		return
	}

	logger.Info(ctx, "handleAction: started",
		slog.F("action", action.Action),
		slog.F("elapsed_ms", a.clock.Since(handlerStart).Milliseconds()),
	)

	geometry := desktopGeometryForAction(cfg, action)
	scaleXY := geometry.DeclaredPointToNative

	var resp DesktopActionResponse

	switch action.Action {
	case "key":
		if action.Text == nil {
			httpapi.Write(ctx, rw, http.StatusBadRequest, codersdk.Response{
				Message: "Missing \"text\" for key action.",
			})
			return
		}
		if err := a.desktop.KeyPress(ctx, *action.Text); err != nil {
			httpapi.Write(ctx, rw, http.StatusInternalServerError, codersdk.Response{
				Message: "Key press failed.",
				Detail:  err.Error(),
			})
			return
		}
		resp.Output = "key action performed"