1 files changed, 88 insertions, 0 deletions
diff --git a/itches/emacs/tracker.org b/itches/emacs/tracker.org
index 0f23f3a..a97e788 100644
--- a/itches/emacs/tracker.org
+++ b/itches/emacs/tracker.org
@@ -130,6 +130,94 @@ Cf. [[https://bugs.archlinux.org/task/58886][Arch Linux]], [[https://bugs.debian
 :Role:           author
 :PatchApplied:   t
 :END:
+*** TODO Add command to toggle password visibility
+**** cover letter
+Hello Emacs,
+
+On occasion, e.g. when saving .authinfo.gpg, I'd like to briefly peek at
+the password I'm typing.  Entering the password twice is an appreciated
+safeguard, but I know I'm perfectly able to make the same typo twice in
+a row; I'd love for read-passwd to offer a command to toggle the
+password visibility.
+
+(I'm dismissing the "write it in the clear, then cut it and paste it"
+solution, since I don't want to add the kill ring nor my desktop
+environment's clipboard manager to the equation)
+
+I'm attaching a proof-of-concept.  Additional considerations, in no
+particular order:
+
+- For the binding, I wanted something easy to mash quickly, but
+  something more mnemonic might be in order?
+
+- Some applications out there show the password while the keyboard
+  shortcut (or the mouse button) remains pressed, and hide it back as
+  soon as it is released; I like the idea, but I have no idea whether
+  that's feasible with Emacs.
+
+- The (member …) check could be replaced with e.g. (get-char-property
+  (minibuffer-prompt-end) 'display); that's how I implemented it at
+  first.  It would fail to turn hiding off when the user has not entered
+  anything though.
+
+- I don't know how much documentation this new command would warrant:
+
+    - Should it be mentioned in read-passwd's docstring?  C-u isn't now,
+      but it was way back when, before 2012-04-11 "* lisp/subr.el
+      (read-passwd): Use read-string." (088be6fbd2).
+
+    - Should it be mentioned in the manual?  delete-minibuffer-contents
+      is documented of course, but not under "(elisp) Reading a
+      Password".
+
+    - a NEWS entry, maybe?  AFAICT C-u was never announced there.
+
+(On a side-note, which might warrant another bug report, I wonder if
+anything could be done to advertise read-passwd-map bindings more.
+Sure, C-h b shows these bindings, but users have no reason to assume
+that read-passwd has extra bindings in addition to regular minibuffer
+ones, so why would they reach for C-h b?)
+
+Let me know if this sounds like something useful.  If so, and if the
+implementation is acceptable as-is, I'll work on documentation as soon
+as I have a clearer idea of where I should add some (and whether some
+documentation should be added for C-u as well).
+
+Thank you for your time.
+
+TODO:attach
+**** "serving suggestion" patch
+#+begin_src diff
+diff --git a/lisp/subr.el b/lisp/subr.el
+index 06ea503da6..da9253cedd 100644
+--- a/lisp/subr.el
++++ b/lisp/subr.el
+@@ -2820,6 +2820,7 @@ read-passwd-map
+   ;; minibuffer-local-map along the way!
+   (let ((map (make-sparse-keymap)))
+     (set-keymap-parent map minibuffer-local-map)
++    (define-key map "\C-c\C-c" #'read-passwd-toggle-visibility)
+     (define-key map "\C-u" #'delete-minibuffer-contents) ;bug#12570
+     map)
+   "Keymap used while reading passwords.")
+@@ -2830,6 +2831,14 @@ read-password--hide-password
+       (put-text-property (+ i beg) (+ 1 i beg)
+                          'display (string (or read-hide-char ?*))))))
+ 
++(defun read-passwd-toggle-visibility ()
++  (interactive)
++  (if (member #'read-password--hide-password post-command-hook)
++      (progn
++        (remove-hook 'post-command-hook #'read-password--hide-password t)
++        (remove-text-properties (minibuffer-prompt-end) (point-max) '(display)))
++    (add-hook 'post-command-hook #'read-password--hide-password nil t)))
++
+ (defun read-passwd (prompt &optional confirm default)
+   "Read a password, prompting with PROMPT, and return it.
+ If optional CONFIRM is non-nil, read the password twice to make sure.
+
+#+end_src
+
 *** imenu
 **** TODO Add grouping/sorting predicates to ~imenu~ completion
 So that ~imenu~ + ~icomplete-vertical-mode~ becomes a built-in,