diff --git a/emacs/flan-dev.el b/emacs/flan-dev.el
index 19c4b69..51e1289 100644
--- a/emacs/flan-dev.el
+++ b/emacs/flan-dev.el
@@ -179,6 +179,12 @@ that quietly did nothing, which is why it is on."
                  ;; A restarted daemon is a rebuilt program: everything known
                  ;; about its names was about the last one.
                  (flan-dev--forget-defs)
+                 ;; And asked again straight away.  An empty cache is honest
+                 ;; but silent: eldoc would go quiet and M-. would fall through
+                 ;; to some other backend until the next install happened to
+                 ;; refill it.  Safe to call from here — the connection is live
+                 ;; by now, so it does not come back through this function.
+                 (ignore-errors (flan-dev-refresh-defs))
                  (message "flan dev: reconnected to %s"
                           (abbreviate-file-name flan-dev--socket)))
         (error
@@ -254,8 +260,11 @@ mistake, so it is distinguished from never having connected."
       (_ (propertize " flan:off" 'face 'shadow
                      'help-echo "Not connected (C-c C-z)")))))
 
-;; Appended rather than prepended: this is the least urgent thing in the line.
-(add-to-list 'mode-line-misc-info '(:eval (flan-dev-mode-line)) t)
+;; Installed buffer-locally by `flan-dev-setup', not globally.  A global entry
+;; would evaluate on every redisplay of every buffer in the session — dired,
+;; eshell, everything — to return nil, for someone who may never open a .flan
+;; file at all.  The `derived-mode-p' guard above stays anyway: cheap, and it
+;; keeps the function honest wherever it is called from.
 
 ;;;###autoload
 (defun flan-show-output ()
@@ -533,8 +542,15 @@ Installed from here rather than from `flan-mode', which must keep working for
 someone editing Flan with no program running and this file never loaded."
   (add-hook 'completion-at-point-functions #'flan-dev-completion-at-point nil t)
   (add-hook 'xref-backend-functions #'flan-dev-xref-backend nil t)
+  ;; Registered, not switched on.  Contributing a source is this file's
+  ;; business; whether eldoc runs at all is the user's, and turning it on for
+  ;; someone who has `global-eldoc-mode' off is overruling a decision they made
+  ;; on purpose.  It is on by default, so this is what almost everyone gets.
   (add-hook 'eldoc-documentation-functions #'flan-dev-eldoc-function nil t)
-  (eldoc-mode 1))
+  (unless (member '(:eval (flan-dev-mode-line)) mode-line-misc-info)
+    ;; Appended rather than prepended: the least urgent thing in the line.
+    (setq-local mode-line-misc-info
+                (append mode-line-misc-info '((:eval (flan-dev-mode-line)))))))
 
 (add-hook 'flan-mode-hook #'flan-dev-setup)
 
diff --git a/emacs/test-flan-dev.el b/emacs/test-flan-dev.el
index cf2692c..116f208 100644
--- a/emacs/test-flan-dev.el
+++ b/emacs/test-flan-dev.el
@@ -54,6 +54,14 @@ is written instead — the real `message' call the real command makes."
                          (string-match-p "live" (flan-dev-mode-line))))
   (test-flan--check "and says nothing in a buffer that is not Flan's"
                     (with-temp-buffer (null (flan-dev-mode-line))))
+  ;; Buffer-locally, so that someone who loads this and never opens a .flan
+  ;; file is not evaluating it on every redisplay of every buffer they have.
+  (test-flan--check "the indicator is in this buffer's modeline"
+                    (member '(:eval (flan-dev-mode-line)) mode-line-misc-info))
+  (test-flan--check "and not in everyone else's"
+                    (with-temp-buffer
+                      (not (member '(:eval (flan-dev-mode-line))
+                                   mode-line-misc-info))))
 
   ;; A daemon restarted while Emacs was not looking is the ordinary case.  The
   ;; socket outlives this connection, so dropping the process and asking again
@@ -67,6 +75,11 @@ is written instead — the real `message' call the real command makes."
     (test-flan--check "the next request reconnects on its own"
                       (and (process-live-p flan-dev--connection)
                            (member "step" (plist-get r :fns)))))
+  ;; ...and knows the program's names again.  An empty cache after a reconnect
+  ;; is honest but silent: eldoc goes quiet and M-. falls through to another
+  ;; backend, with nothing said about why.
+  (test-flan--check "and knows the program's names again"
+                    (assoc "step" flan-dev--defs))
 
   ;; But a socket nobody is listening on is refused by name, rather than
   ;; retried forever or reported as some other failure.