I've some problem related to menu alignment whenever I change the buffer "zoom" exploting text-scale-decrease

When popup was used in auto-complete, change color of common part of all candidates.

Just tried doing an el-get update on auto-complete, died trying to clone the popup submodule with the following error:

fatal: remote error: access denied or repository not exported: /c/nw/cc/ce/b2/1410546/1410546.git

Trying to do a git clone git://github.com/auto-complete/popup-el gives the same error

Popup is disrupted in the following test code.

(ert-deftest popup-test-scroll-bar-right ()
  (popup-test-with-temp-buffer
    (insert (make-string (- (window-width) 1) ? ))
    (let* ((popup-scroll-bar-foreground-char
            (propertize "f" 'face 'popup-scroll-bar-foreground-face))
           (popup-scroll-bar-background-char
            (propertize "b" 'face 'popup-scroll-bar-background-face))
           (popup (popup-tip "Foo\nBar\nBaz\nFez\nOz"
                             :nowait t :height 3 :scroll-bar t :margin t)))
      (with-current-buffer (popup-test-helper-get-overlays-buffer)
        (let ((points (popup-test-helper-match-points
                       '("Foof" "Barb" "Bazb"))))
          (should (equal (popup-test-helper-same-all-p
                          (popup-test-helper-points-to-columns points))
                         (- (window-width) 5)))
          (should (eq (line-number-at-pos (car points)) 2))
          )))))

Good day. Is it possible to put new release tag, so that new features like popup-menu* :initial-index deployed to melpa-stable?
Thanks in advance.

I'm using popup.el but I'm not able to close or open the popup help
I'm using the code

(require 'popup)
(define-key popup-menu-keymap [esc] 'popup-close)
(popup-menu* '("hello" "word"))

But using left or esc have no effect and the popup remain opened until I select an element

Create a predicate that allows user to see whether or not they are currently in a snippet. This would help with configuring other modes to disable themselves (or enable themselves) in the event that the user is currently inside of a snippet.

My primary motivation would be something like the following:

(defadvice ac-expand (before advice-for-ac-expand activate)
  (when (yas-in-snippet-p)
    (ac-stop))```

Which would allow you to disable autocompletion if inside a snippet.

Hi, here is the issue description:

(defvar issue
  '(:env '((os "Debian wheezy/sid")
           (:name "emacs"         :from "debian apt-get"     :version "GNU Emacs 23.4.1")
           (:name "popup"         :from "m2ym/popup-el"      :version "master")
           (:name "fuzzy"         :from "m2ym/fuzzy-el"      :version "master")
           (:name "auto-complete" :from "m2ym/auto-complete" :version "master"))

         :issue-detail
         '((:in "(nth-line -1)" :symptom "ac-quick-help keeps in dropdown list's last line")
           (:in "(nth-line -2)" :symptom "ac-quick-help keeps in dropdown list's last line")
           (:in "(nth-line -3)" :symptom
                "ac-quick-help keeps in dropdown list's first line, but:
               lines from 2 to 9 (= ac-menu-height 10) are all empty, and
               all the rest help string are in the last line."))

         :solve-it
         '(replace-string "(when (eobp)
        (popup-save-buffer-state
          (let ((begin (point)))
            (insert \" \")
            (setq padding (make-overlay begin (point)))
            (overlay-put padding 'evaporate t))))"
                          "(let ((newlines (max 0 (+ (- height (count-lines point (point-max))) (if around 1 0)))))
        (when (> newlines 0)
          (popup-save-buffer-state
            (goto-char (point-max))
            (let ((begin (point)))
              (insert (make-string newlines ?\n))
              (setq padding (make-overlay begin (point)))
              (overlay-put padding 'evaporate t)))))")))

This is based on conversation with @supermomonga.

In some cases, for example, using autocomplete.el, we want to show candidates without selecting one of them. By default, auto-complete.el force to select the first candidate and pressing Enter key expands the candidate.

What we really want is:

Enter and Tab: expand candidate if selected, otherwise self insert.
(Optional) Tab: always expand. If nothing selected, expand first one.

AFAIU, popup.el doesn't admit this. I think we'll easily implement the above (optional) feature to auto-complete.el if popup.el admits it. (We may need some additional APIs.) However the modification for it on popup.el is not straightforward, I guess.

What do you think of it?

popup-isearch blocks test execution in the following test code.

(ert-deftest popup-test-isearch ()
  (popup-test-with-common-setup
    (popup-set-list popup '("foo" "bar" "baz"))
    (popup-draw popup)
    (popup-test-helper-input ?a)
    (popup-isearch popup)
    (should (equal (popup-list popup) '("bar" "baz")))
    (should (every #'identity
                   (popup-test-helper-match-points '("bar" "baz"))))
    ))

Is it possible to change popup-isearch spec to use overriding-terminal-local-map as isearch-mode do?

(popup-cascade-menu '(("Foo" "a" "b" "c") "Bar"))

"c" seems not to be in "Foo".

For some reason flycheck compilation error popup can't be displayed over text so it chooses to add each line of compilation error popup at the end of existing line:

Disabling fci-mode fixes issue, would be great to make popup-el to work properly with fci-mode tho.

I was looking at enhancing git-messenger to copy the commit for a given commit to the kill-ring if a certain key was pressed. It seems hard to do this with the popup-tip and using the menu machinery seems excessive given I don't want to loose the info in the tip. Any tips?

Hi,

Is there a variable that I configure to turn on the word-wrap feature in popup-tip function?

There might be an advantage I'm not aware of, but wouldn't it be better to run the tests with --batch? That would it possible for non-interactive test/builds since we wouldn't have to close the buffer of test results.

My default font is `:family "Lucida Sans"' - variable width. Popup renders menus in variable-width-face-buffers unaligned:

Perhaps there are calls to string length, which should be changed to string width?

In any case I can live with it, it just doesn't look as good as it could.

Moved from auto-complete/auto-complete#91

Please add annotated tag in git repo for Version: 0.5

Feature request to be used in combo w/ac-complete ac-isearch
I think generic isearch filtering might be useful - for example filtering by docstring.

When using popup.el with whitespace-mode and linum-mode, unwanted artifacts are seen on the screen due to the fact that popup.el adds empty lines at the end of the buffer to render the popup menu.

Please see:
auto-complete/auto-complete#90
sebastiw/edts#35
bbatsov/prelude#19

I think that popup.el is the root cause for all these issues.

popup-tip ignore margin when beginning of line,and popup is disrupted when end of line.
Here is the test code.

(ert-deftest popup-test-margin-at-left ()
  (popup-test-with-temp-buffer
    (let ((popup (popup-tip "Margin?" :nowait t :margin t)))
      (with-current-buffer (popup-test-helper-get-overlays-buffer)
        (let ((points (popup-test-helper-match-points '(" Margin? "))))
          (should (every #'identity points))
          (should (equal (car (popup-test-helper-points-to-columns points))
                         0))
          )))))

(ert-deftest popup-test-margin-at-right ()
  (popup-test-with-temp-buffer
    (insert (make-string (- (window-width) 1) ? ))
    (let ((popup (popup-tip "Margin?" :nowait t :margin t)))
      (with-current-buffer (popup-test-helper-get-overlays-buffer)
        (let ((points (popup-test-helper-match-points '(" Margin? "))))
          (should (every #'identity points))
          (should (equal (car (popup-test-helper-points-to-columns points))
                         (- (window-width) 9)))
          )))))

I found a brilliantly useful function called popup-complete here http://www.emacswiki.org/emacs/PopUp which references your project.

But I can't find it here. Assuming you have something to do with it, where is popup-complete distributed? It would be really fantastic if it were available on MELPA.

Many third-party projects consider this as a de-facto popup library. It would be good to be able to use it everywhere, including Company, if only to spare the users from having to set up the same faces when different popups are used.

In order to do that, all non-trivial contributors would have to sign copyright assignments to FSF. Is that feasible?

If this goes through, I'll try to take care of the current display bugs. You have several long-standing ones, and I believe most (all?) of them are fixed in Company's popup now.

Reading open issues, it seems most of the problems are from overlays from other modules and non-uniform text size (e.g., proportional font). I guess these problems are difficult to solve, probably impossible for the later case. Alternatively I think we can use x-popup-menu as the "backend" and then these problems won't show up with it. I am just raising the idea and wondering if there is somebody already tried.

For those who are curious, you can try x-popup-menu easily by:

(x-popup-menu
 t
 '("Title"
   ("Title 1" ("Item 1-1" . 11) ("Item 1-2" . 12))
   ("Title 2" ("Item 2-1" . 21) ("Item 2-2" . 22))))

I can't seem to get popup to use the keymap I'm giving it. Here's a snippet of how I'm declaring the keymap and creating the actual popup:

(defvar om-menu-keymap
  (let ((map (make-sparse-keymap)))
    (define-key map "\M-n"    'om-menu-next)
    (define-key map "\M-p"    'om-menu-previous)
    map))

(setq om-menu (popup-create (point) 15 10 ;om-menu-width om-menu-height
                                :around t
                                :face 'om-candidate-face
                                :mouse-face 'om-candidate-mouse-face
                                :selection-face 'om-selection-face
                                :symbol t
                                :scroll-bar t
                                :margin-left 0
                                :keymap om-menu-map)))
  (popup-set-list om-menu om-autocomplete-candidates)
  (popup-draw om-menu)

Why is popup not accepting the keymap I'm passing to popup-create?

This is a feature request for the following use case:

The user has a popup, with a lot of elements in it. He/she would like
to match a part of them at a time, each time filtering the remaining
candidates with a new search.

Example:

The user has these words in his/her popup:

• schtroumpfed
• squirrelled
• broughammed
• schmaltzed
• squirreled
• scrootched
• scroonched
• scraunched
• strengthed

to match words starting with "sc" the user can isearch for "sc".
Then the user would like to match words with "led" in them, so the
user performs another search for "led".
The remaining match is "scrolled".

The words were taken from:

https://en.wikipedia.org/wiki/List_of_the_longest_English_words_with_one_syllable

In common cases (posn-at-point) returns structure like:
(#<window 12 on TAGS<idid>> 90 (0 . 100) 0 nil 90 (0 . 5) nil (0 . 0) (10 . 20))

In rare cases (posn-at-point) returns:
(#<frame : 0x1149ad8> nil (0 . 0) 0)

E.g. after ^L char in TAGS buffers.

This leads to error on popup-create call:

Debugger entered--Lisp error: (wrong-type-argument windowp #<frame : /home/kostafey/.emacs.d/elpa/popup-20150116.1223/popup.el 0x1149ad8>)
  window-buffer(#<frame : /home/kostafey/.emacs.d/elpa/popup-20150116.1223/popup.el 0x1149ad8>)
  posn-col-row((#<frame : /home/kostafey/.emacs.d/elpa/popup-20150116.1223/popup.el 0x1149ad8> nil (0 . 0) 0))
  popup-create(640 50 15 :max-width nil :around nil :face popup-menu-face :mouse-face popup-menu-mouse-face :selection-face popup-menu-selection-face :summary-face popup-menu-summary-face :margin-left 1 :margin-right 0 :scroll-bar t :symbol nil :parent nil :parent-offset nil)
  popup-menu*(("popup.el<popup-20150116.1223>" "TAGS<idid>" "*scratch*" "*Messages*" "*Compile-Log*" "*Quail Completions*" "popup.el<popup-el>" "profiles.clj" "keyboard.c" "subr.el.gz" "popup-switcher.el" "speed.clj" "report.html" "wservice.clj" "clojure-conf.el" "cider-interaction.el" "key-bindings.org" "clomacs.el" "core.clj" "plot.sh" "helpers.clj" "form-init916177696597684299.clj" "project.clj" "form-init5549897678859169329.clj" "handler.clj" "launch.txt" "form-init5785223491863655970.clj" "plotting_data1.dat" "org-loaddefs.el<org-plus-contrib-20150223>" "org-loaddefs.el<org-20150223>" "stat.h") :point 640 :height 15 :scroll-bar t :margin-left 1 :margin-right 1 :around nil :isearch t)
  psw-popup-menu(("popup.el<popup-20150116.1223>" "TAGS<idid>" "*scratch*" "*Messages*" "*Compile-Log*" "*Quail Completions*" "popup.el<popup-el>" "profiles.clj" "keyboard.c" "subr.el.gz" "popup-switcher.el" "speed.clj" "report.html" "wservice.clj" "clojure-conf.el" "cider-interaction.el" "key-bindings.org" "clomacs.el" "core.clj" "plot.sh" "helpers.clj" "form-init916177696597684299.clj" "project.clj" "form-init5549897678859169329.clj" "handler.clj" "launch.txt" "form-init5785223491863655970.clj" "plotting_data1.dat" "org-loaddefs.el<org-plus-contrib-20150223>" "org-loaddefs.el<org-20150223>" "stat.h"))
  psw-switcher(:items-list (#<buffer popup.el<popup-20150116.1223>> #<buffer TAGS<idid>> #<buffer TAGS<idid>> #<buffer *scratch*> #<buffer *Messages*> #<buffer *Compile-Log*> #<buffer *Quail Completions*> #<buffer popup.el<popup-el>> #<buffer profiles.clj> #<buffer keyboard.c> #<buffer subr.el.gz> #<buffer popup-switcher.el> #<buffer speed.clj> #<buffer report.html> #<buffer wservice.clj> #<buffer clojure-conf.el> #<buffer cider-interaction.el> #<buffer key-bindings.org> #<buffer clomacs.el> #<buffer core.clj> #<buffer plot.sh> #<buffer helpers.clj> #<buffer form-init916177696597684299.clj> #<buffer project.clj> #<buffer form-init5549897678859169329.clj> #<buffer handler.clj> #<buffer launch.txt> #<buffer form-init5785223491863655970.clj> #<buffer plotting_data1.dat> #<buffer org-loaddefs.el<org-plus-contrib-20150223>> #<buffer org-loaddefs.el<org-20150223>> #<buffer stat.h>) :item-name-getter #[257 "r\211q\210�\203�\301 \203�\302 \204�\303 \304P\202�\303 )\207" [psw-mark-modified-buffers buffer-modified-p psw-is-temp-buffer buffer-name " *"] 3 "\n\n(fn BUFFER)"] :switcher switch-to-buffer)
  psw-switch-buffer()
  call-interactively(psw-switch-buffer nil nil)
  command-execute(psw-switch-buffer)

The early used (popup-current-physical-column) (see: kenbeese@c02ec73#diff-99b97c0478ea73bf04732050d7160dc4L539)
don't have such problems.

Is it possible to return previous (popup-current-physical-column) usage or any other solutions to get current point position correctly in all cases?

Line feeds can be ignored in popup-tip because fill-region ignores them.
Here is an example:

ELISP> (popup-fill-string "line 1\nline 2" 12)
(11 "line 1 line" "2")

I wonder if this is expected behavior.
FYI, we can avoid it by setting use-hard-newlines to non-nil.

ELISP> (defadvice fill-region (before my-fill-region activate)
         (setq use-hard-newlines t))
fill-region
ELISP> (popup-fill-string (format "line 1%sline 2" (propertize "\n" 'hard t)) 12)
(6 "line 1" "line 2")

I'm having the strangest problem all of a sudden with alignment of popup menus in auto-complete when using php-mode. This is what it looks like:

To try to diagnose whether this was just an issue with my config, I tried:

1. Rolling back my .emacs by several weeks (it's in git, of course).
2. Deleting my entire ~/.emacs.d/elpa directory and re-downloading/re-compiling all packages.

Neither of these helped. This problem doesn't seem to affect text-mode or lisp-interaction-mode, though, as shown here:

I've also opened an issue with php-mode, but they haven't been able to reproduce it. I have attempted to debug what is going on in popup.el, but admittedly my Lisp isn't very strong and I can't figure out how the placement of the left edge of the popup is computed.

Any help or direction on how to figure this out would be greatly appreciated.

Are you plannig to add fuzzy matching to isearch on popup-menu?
Will you accept such idea to popup-el?

because referenced ert module removed from github

Please add a 0.5 tag, so that the new version is included into melpa-stable

Otherwise auto-complete is uninstallable there, since it depends on popup 0.5

Thanks for your great work

(Previously posted as Jedi-issue #174.)

Docstrings in auto-complete list all of a sudden shows some weird behavior in my emacs setup, as shown in the picture below.

To ensure that the docstrings will show when the dot is close to the right margin, I have the following in my .emacs file:

(setq jedi:complete-on-dot)
(add-hook 'python-mode-hook 'jedi:setup)
; preventing unwanted behavior of the doc strings
(defun my/jedi-mode-hook ()
  (setq-local ac-max-width 0.5))
(add-hook 'jedi-mode-hook 'my/jedi-mode-hook)

All help would be greatly appreciated.

Hi, I have created a popup-tip with :scroll-bar t, and now I don't know hot to scroll it down. I have seen the popup-menu-keymap in the source, but that doesn't seem to be tied to popup-tip.

Hey,

I've noticed that popups get corrupted by hidden text in modes like allout mode (and likely outline mode, etc).

For example, if we want an autocomplete popup near these hidden sections:

Corruption happens at exactly the lines that were hidden:

Here is a file with allout structure: http://git.savannah.gnu.org/cgit/emacs.git/plain/lisp/allout.el

When viewing this file, M-x allout-mode, M-x allout-hide-bodies, and try opening a popup (maybe through autocomplete) near hidden text.

Here;'s what I get when using auto-complete master + popup-el from either this repo or AC 's . For elisp it works fine , but the clangs' extension https://github.com/brianjcj/auto-complete-clang looks as it looks .

Also rised here :

brianjcj/auto-complete-clang#25

On OSX terminal, I tried to type Japanese letter while showing popup-tip but the letter went unreadable.
I’m not using Japanese input method on emacs but using system wide Japanese input method.
the gif shows what I did.

Hi, I am trying to display some text with popup-tip. The text may contain italics or bold, but I want it to display with the tooltip face as the base, rather than the buffer's default.

(popup-tip  #("foo bar" 0 3 (face bold) 3 4 nil 4 7 (face italic)) :nostrip t)

Can popup-tip do this conversion or do I have to do the conversion myself?

I think this issue might be related to issue #3. Here is the screenshot from the org-mode buffer. Couple headers are closed and their overlay breaks the popup

when using popup.el in a buffer with variable-pitch-mode the popup looks distorted.

Popup is disrupted in the following test code.

(ert-deftest popup-test-two-tip ()
  (popup-test-with-temp-buffer
   (popup-tip "Foo\nBar" :nowait t)
   (save-excursion (insert "\n"))
   (popup-tip "Baz\nQux" :nowait t)
   (with-current-buffer (popup-test-helper-get-overlays-buffer)
     (let ((points1 (popup-test-helper-match-points '("Foo" "Bar")))
           (points2 (popup-test-helper-match-points '("Baz" "Qux"))))
       (should (popup-test-helper-same-all-p
                (popup-test-helper-points-to-columns points1)))
       (should (popup-test-helper-same-all-p
                (popup-test-helper-points-to-columns points2)))
       (should (eq (line-number-at-pos (car points2)) 2))
       (should (eq (line-number-at-pos (car points1)) 3))
       ))))

I think it is mentioned somewhere in the AC issue tracker, but I couldn't find it. Anyway, this issue still exists in the latest master.

From time to time I get an error looking like the below. Most of the time everything works fine, and I haven't been able to determine exactly what triggers the error. The error arises when using auto-complete with nREPL 0.1.8-preview and ac-nrepl 20130420.929.

Debugger entered--Lisp error: (args-out-of-range 1 143)
count-lines(143 1)
popup-create(143 69 20 :min-height 0 :around t :margin-left nil :margin-right nil :scroll-bar nil :face popup-tip-face :parent nil :parent-offset nil)
popup-tip("clojure.core/take\n([n coll])\nReturns a lazy sequence of the first n items in coll, or all items if\nthere are fewer than n.\n" :point 143 :height 20 :min-height 0 :around t :parent nil :parent-offset nil :point 143 :height 20 :nowait t)
apply(popup-tip "clojure.core/take\n([n coll])\nReturns a lazy sequence of the first n items in coll, or all items if\nthere are fewer than n.\n" :point 143 :height 20 :min-height 0 :around t :parent nil :parent-offset nil (:point 143 :height 20 :nowait t))
popup-menu-show-quick-help([cl-struct-popup 143 8 0 10 10 nil 1 # # # # # # # # # # (mouse-4 . ac-mouse-4) (down-mouse-1 . ac-ignore) (mouse-1 . ac-mouse-1) (19 . ac-isearch) (16 . ac-previous) (14 . ac-next) keymap (C-up . ac-quick-help-scroll-up) (C-down . ac-quick-help-scroll-down) (67108927 . ac-help) (M-f1 . ac-persist-help) (f1 . ac-help) (up . ac-previous) (down . ac-next) (27 keymap (57 . ac-complete-9) (56 . ac-complete-8) (55 . ac-complete-7) (54 . ac-complete-6) (53 . ac-complete-5) (52 . ac-complete-4) (51 . ac-complete-3) (50 . ac-complete-2) (49 . ac-complete-1) (16 . ac-quick-help-scroll-up) (14 . ac-quick-help-scroll-down) (67108927 . ac-persist-help) (112 . ac-previous) (110 . ac-next) (9 . auto-complete)) (return) (13) (tab . ac-expand) (9 . ac-expand)) nil 0 ac-candidate-face ac-candidate-mouse-face ac-selection-face popup-summary-face 1 0 nil t t 0 0 0 0 (#("take" 0 4 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v")) #("take-nth" 0 8 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v")) #("take-last" 0 9 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v")) #("take-while" 0 10 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v"))) 9 nil (#("take" 0 4 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v")) #("take-nth" 0 8 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v")) #("take-last" 0 9 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v")) #("take-while" 0 10 (selection-face ac-nrepl-selection-face popup-face ac-nrepl-candidate-face document ac-nrepl-documentation symbol "v")))] nil :point 143 :height 20 :nowait t)
ac-quick-help()
apply(ac-quick-help nil)
byte-code("r\301\302H\303H\"\210)\301\207" [timer apply 5 6] 4)
timer-event-handler([t 0 0 500000 0.5 ac-quick-help nil idle 0])
accept-process-output(nil 0.005)
nrepl-send-request-sync(("ns" "com.jayway.rps.core" "op" "eval" "session" "c6135125-2940-4693-bdaa-741f1e075b90" "code" "(require 'complete.core) (filter #(.startsWith % \"ta\")(complete.core/namespaces ns))"))
nrepl-send-string-sync("(require 'complete.core) (filter #(.startsWith % \"ta\")(complete.core/namespaces ns))" "com.jayway.rps.core" "c6135125-2940-4693-bdaa-741f1e075b90")
ac-nrepl-sync-eval("(require 'complete.core) (filter #(.startsWith % \"ta\")(complete.core/namespaces ns))")
ac-nrepl-candidates*("(filter #(.startsWith % \"ta\")(complete.core/namespaces ns))")
ac-nrepl-candidates-ns()
ac-candidates-1(((candidates . ac-nrepl-candidates-ns) (symbol . "n") (available . ac-nrepl-available-p) (candidate-face . ac-nrepl-candidate-face) (selection-face . ac-nrepl-selection-face) (prefix . ac-nrepl-symbol-start-pos) (document . ac-nrepl-documentation)))
ac-candidates()
ac-update(t)
ac-show-menu()
apply(ac-show-menu nil)
byte-code("r\301\302H\303H\"\210)\301\207" [timer apply 5 6] 4)
timer-event-handler([t 0 0 400000 0.4 ac-show-menu nil idle 0])

Is it possible?

Thanks.

While I was debugging jorgenschaefer/elpy#182, I dived deep into auto-complete.el and popup.el to find the cause for the "permanently" added empty lines I was seeing w/ elpy.

In the case of elpy, auto-complete queried for a documentation string and got an empty string which it used to create a quick tip (ac-quick-help -> popup-menu-show-quick-help -> popup-tip -> popup-create).

Most likely, a popup w/ width 0 is created, which will cause trouble in some conditional blocks. I haven't had the time to debug this even further, I am afraid. But if a helping hand is needed, I'll gladly lend one or two. :)

Besides that, the bug could be seen as twofold: Neither should auto-complete.el try to create a popup w/ no contents, nor should popup.el get in an undefined state if someone tries something like that.

It works incorrectly in elisp mode. Different lines in the function list (the left dialogue) seem have different length, so do the associated dialogue (the right one) for more details. The problem is illustrated as

Maybe due to the upgrade of packages one month ago. It showed correctly before.
However, it still works well in cc-mode or python-mode, shown as

One more question:
Could we stop the blank in the left dialogue due to the line wrapping in the right dialogue by adding some configurations, as shown in the above figures or the figure below in python mode? (when I use "wrap at window edge" or "word wrap"; appearing in any modes) .

When I use "truncate long line", it doesn't have this problem, but not satisfactory to be a solution due to the large length of most functions.

version:
emacs: 24.4.1
ac-complete: 20150408.1132
popup: 20150315.612

• Clean up API