GNU bug report logs - #77063
29.1; ido-read-buffer

Previous Next

To reply to this bug, email your comments to 77063 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Keith David Bershatsky <esq <at> lawlist.com>
To: Emacs Bug Reports <bug-gnu-emacs <at> gnu.org>
Subject: 29.1; ido-read-buffer
Date: Sun, 16 Mar 2025 20:16:45 -0700

This is a two part bug report:

PART 1 of 2:

STEP 0:  Start-up Emacs 29 without any user configuration.

STEP 1:  (require 'ido)

STEP 2:  (ido-read-buffer "PROMPT:  " nil 'require-match nil)

STEP 3:  Observe that there are no visible ido rotational buffer choices.

STEP 4:  (ido-common-initialization)

STEP 5:  Repeat STEP 2 and observe that it now works; i.e., there are now visible ido rotational buffer choices.

================

PART 2 of 2

STEP 0:  Start-up Emacs 29 without any user configuration.

STEP 1:  (require 'ido)

STEP 2:  (ido-read-buffer "PROMPT:  " nil 'require-match nil)

STEP 3:  Press the TAB key to open the *Ido Completions* buffer.

STEP 4:  Navigate to the *Ido Completions* buffer and select the *Messages* buffer, and press the enter key.

STEP 5:  Now that the word *Messages* is in the minibuffer following the word PROMPT:, press the enter key.  The function erroneously returns the "*scratch*" buffer instead of the *Messages* buffer.

Message #8 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Keith David Bershatsky <esq <at> lawlist.com>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 15:40:07 +0200

> Date: Sun, 16 Mar 2025 20:16:45 -0700
> From: Keith David Bershatsky <esq <at> lawlist.com>
> 
> This is a two part bug report:
> 
> PART 1 of 2:
> 
> STEP 0:  Start-up Emacs 29 without any user configuration.
> 
> STEP 1:  (require 'ido)
> 
> STEP 2:  (ido-read-buffer "PROMPT:  " nil 'require-match nil)
> 
> STEP 3:  Observe that there are no visible ido rotational buffer choices.
> 
> STEP 4:  (ido-common-initialization)
> 
> STEP 5:  Repeat STEP 2 and observe that it now works; i.e., there are now visible ido rotational buffer choices.
> 
> ================
> 
> PART 2 of 2
> 
> STEP 0:  Start-up Emacs 29 without any user configuration.
> 
> STEP 1:  (require 'ido)
> 
> STEP 2:  (ido-read-buffer "PROMPT:  " nil 'require-match nil)
> 
> STEP 3:  Press the TAB key to open the *Ido Completions* buffer.
> 
> STEP 4:  Navigate to the *Ido Completions* buffer and select the *Messages* buffer, and press the enter key.
> 
> STEP 5:  Now that the word *Messages* is in the minibuffer following the word PROMPT:, press the enter key.  The function erroneously returns the "*scratch*" buffer instead of the *Messages* buffer.

This appears to be fixed already in Emacs 30 and later.

Message #11 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Keith David Bershatsky <esq <at> lawlist.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 08:16:31 -0700

[Message part 1 (text/plain, inline)]

Thank you, Eli, for looking into this particular bug report.  I just installed Emacs 30.1 on Debian 8.11 (Jessie) and was able to repeat the steps to reproduce the reported behavior.  Attached hereto are four (4) screenshots depicting both of the issues reported:

SCREENSHOT:  001__part_1_of_2__step_2.png

This screenshot depicts the minibuffer PROMPT _without_ any ido completion choices.

PROMPT:  [no ido choices]

In order to display ido completion choices following the PROMPT, it is necessary to first run (ido-common-initialization).  Once that is done, repeating Step 2 yields the expected result:

PROMPT:  {*scratch* | *Messages* | *GNU Emacs*}

SCREENSHOT:  002__part_2_of_2__step_4__select_messages_buffer.png

After pressing the tab key, navigate to the ido choices buffer and select the *Messages* buffer.

SCREENSHOT:  002__part_2_of_2__step_4__messages_buffer_selected.png

We have chosen the *Messages* buffer from the ido choices buffer, and inserted that choice into the minibuffer -- following the PROMPT:.

SCREENSHOT:  002__part_2_of_2__step_4__returns_scratch_buffer.png

We pressed the enter button to select the *Messages* buffer and exit the ido-read-buffer function.  The erroneous result returned is the *scratch* buffer.

One idea to fix the first issue is to add (ido-common-initialization) inside ido-read-buffer:

(defun ido-read-buffer (prompt &optional default require-match predicate)
  "..."
  (let* (...
         (buf (progn (ido-common-initialization)
                     (ido-read-internal 'buffer prompt 'ido-buffer-history default require-match))))
    ...

I do not have a solution for the second problem described in part 2 of 2.

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

> Date: [03-17-2025 06:40:07] <17 Mar 2025 15:40:07 +0200>
> From: Eli Zaretskii <eliz <at> gnu.org>
> To: Keith David Bershatsky <esq <at> lawlist.com>
> Cc: 77063 <at> debbugs.gnu.org
> Subject: Re: bug#77063: 29.1; ido-read-buffer
> 
> > Date: Sun, 16 Mar 2025 20:16:45 -0700
> > From: Keith David Bershatsky <esq <at> lawlist.com>
> >
> > This is a two part bug report:
> >
> > PART 1 of 2:
> >
> > STEP 0:  Start-up Emacs 29 without any user configuration.
> >
> > STEP 1:  (require 'ido)
> >
> > STEP 2:  (ido-read-buffer "PROMPT:  " nil 'require-match nil)
> >
> > STEP 3:  Observe that there are no visible ido rotational buffer choices.
> >
> > STEP 4:  (ido-common-initialization)
> >
> > STEP 5:  Repeat STEP 2 and observe that it now works; i.e., there are now visible ido rotational buffer choices.
> >
> > ================
> >
> > PART 2 of 2
> >
> > STEP 0:  Start-up Emacs 29 without any user configuration.
> >
> > STEP 1:  (require 'ido)
> >
> > STEP 2:  (ido-read-buffer "PROMPT:  " nil 'require-match nil)
> >
> > STEP 3:  Press the TAB key to open the *Ido Completions* buffer.
> >
> > STEP 4:  Navigate to the *Ido Completions* buffer and select the *Messages* buffer, and press the enter key.
> >
> > STEP 5:  Now that the word *Messages* is in the minibuffer following the word PROMPT:, press the enter key.  The function erroneously returns the "*scratch*" buffer instead of the *Messages* buffer.
> 
> This appears to be fixed already in Emacs 30 and later.

[002__part_2_of_2__step_4__returns_scratch_buffer.png (application/png, inline)]

[002__part_2_of_2__step_4__messages_buffer_selected.png (application/png, inline)]

[002__part_2_of_2__step_4__select_messages_buffer.png (application/png, inline)]

[001__part_1_of_2__step_2.png (application/png, inline)]

Message #14 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Keith David Bershatsky <esq <at> lawlist.com>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 18:33:35 +0200

> Date:  Mon, 17 Mar 2025 08:16:31 -0700
> From:  Keith David Bershatsky <esq <at> lawlist.com>
> Cc:  77063 <at> debbugs.gnu.org
> 
> Thank you, Eli, for looking into this particular bug report.  I just installed Emacs 30.1 on Debian 8.11 (Jessie) and was able to repeat the steps to reproduce the reported behavior.  Attached hereto are four (4) screenshots depicting both of the issues reported:
> 
> SCREENSHOT:  001__part_1_of_2__step_2.png
> 
> This screenshot depicts the minibuffer PROMPT _without_ any ido completion choices.
> 
> PROMPT:  [no ido choices]
> 
> In order to display ido completion choices following the PROMPT, it is necessary to first run (ido-common-initialization).  Once that is done, repeating Step 2 yields the expected result:
> 
> PROMPT:  {*scratch* | *Messages* | *GNU Emacs*}
> 
> SCREENSHOT:  002__part_2_of_2__step_4__select_messages_buffer.png
> 
> After pressing the tab key, navigate to the ido choices buffer and select the *Messages* buffer.
> 
> SCREENSHOT:  002__part_2_of_2__step_4__messages_buffer_selected.png
> 
> We have chosen the *Messages* buffer from the ido choices buffer, and inserted that choice into the minibuffer -- following the PROMPT:.
> 
> SCREENSHOT:  002__part_2_of_2__step_4__returns_scratch_buffer.png
> 
> We pressed the enter button to select the *Messages* buffer and exit the ido-read-buffer function.  The erroneous result returned is the *scratch* buffer.
> 
> One idea to fix the first issue is to add (ido-common-initialization) inside ido-read-buffer:
> 
> (defun ido-read-buffer (prompt &optional default require-match predicate)
>   "..."
>   (let* (...
>          (buf (progn (ido-common-initialization)
>                      (ido-read-internal 'buffer prompt 'ido-buffer-history default require-match))))
>     ...
> 
> I do not have a solution for the second problem described in part 2 of 2.

Thanks, but the screenshots don't help.  I need a way to reproduce
this behavior.  Maybe your recipe was not detailed enough?

Message #17 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Keith David Bershatsky <esq <at> lawlist.com>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 18:44:26 +0200

> Date:  Mon, 17 Mar 2025 08:16:31 -0700
> From:  Keith David Bershatsky <esq <at> lawlist.com>
> Cc:  77063 <at> debbugs.gnu.org
> 
> I do not have a solution for the second problem described in part 2 of 2.

Oh, so there are two problems in this report?  I thought only the
second part was the bug, and it's that part which I cannot reproduce
in Emacs 30.

Regarding the first part: why it is considered a problem?  Your recipe
shows some effects of some actions, but please explain what is it you
consider to be a problematic behavior so it is clear even to those who
don't use Ido.

Message #20 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Keith David Bershatsky <esq <at> lawlist.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 09:50:37 -0700

As to part 1 of 2, the goal is to be able to start Emacs 30.1, (require 'ido), evaluate (ido-read-buffer ...) and be greeted with a prompt displaying the ido rotating choices that looks like this:

PROMPT:  {*scratch* | *Messages* | *GNU Emacs*}

Starting from Emacs 30.1 without any user configuration, if we only evaluate (require 'ido) and then evaluate (ido-read-buffer ....), we do _not_ see any choices following the prompt.  This is because there needs to be call to ido-common-initialization _before_ calling ido-read-internal within ido-read-buffer.  Once ido-common-initialization has been run one time, all subsequent calls to ido-read-buffer work as expected.

As to part 2 of 2, we just press the TAB key after evaluating (ido-read-buffer ...), and C-x o a couple of times to place focus in the choices buffer and select the *Messages* buffer.  Once the *Messages* buffer has been selected and inserted into the minibuffer, we press the RET key and exit ido-read-buffer.  The result that is thrown is *scratch*, which we did not expect.

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

> Date: [03-17-2025 09:33:35] <17 Mar 2025 18:33:35 +0200>
> From: Eli Zaretskii <eliz <at> gnu.org>
> 
> * * *
> 
> Thanks, but the screenshots don't help.  I need a way to reproduce
> this behavior.  Maybe your recipe was not detailed enough?

Message #23 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Keith David Bershatsky <esq <at> lawlist.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 10:15:20 -0700

I sent out the last email of 09:50:37 without the benefit of reading your email of 09:44:26.  My email of 09:50:37 perhaps addresses what you were seeking in terms of a better recipe.  Yes, there are two (2) distinct problems:

# 1:  The first problem is that no buffer choices are displayed in the minibuffer because there needs to be a call to ido-common-initialization _before_ calling ido-read-internal.  In at least one other section of the ido library, we can observe a call to ido-common-initialization before calling ido-read-internal.  I do not know why we don't just call ido-common-initialization at the bottom of the ido library and remove all other calls from inside functions ... maybe that function is supposed to do something special each time before calling ido-read-internal.

# 2:  The second problem relates to ido-read-buffer throwing the wrong buffer as a result when we use the separate ido completions buffer to choose a buffer instead of using the rotating choices inside the minibuffer.  The correct choice is inserted into the minibuffer, but when pressing RET and exiting ido-read-buffer, the wrong buffer is thrown as a result.

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

> Date: [03-17-2025 09:44:26] <17 Mar 2025 18:44:26 +0200>
> From: Eli Zaretskii <eliz <at> gnu.org>
> To: Keith David Bershatsky <esq <at> lawlist.com>
> Cc: 77063 <at> debbugs.gnu.org
> Subject: Re: bug#77063: 29.1; ido-read-buffer
> 
> > Date:  Mon, 17 Mar 2025 08:16:31 -0700
> > From:  Keith David Bershatsky <esq <at> lawlist.com>
> > Cc:  77063 <at> debbugs.gnu.org
> >
> > I do not have a solution for the second problem described in part 2 of 2.
> 
> Oh, so there are two problems in this report?  I thought only the
> second part was the bug, and it's that part which I cannot reproduce
> in Emacs 30.
> 
> Regarding the first part: why it is considered a problem?  Your recipe
> shows some effects of some actions, but please explain what is it you
> consider to be a problematic behavior so it is clear even to those who
> don't use Ido.

Message #26 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Keith David Bershatsky <esq <at> lawlist.com>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 21:25:22 +0200

> Date:  Mon, 17 Mar 2025 09:50:37 -0700
> From:  Keith David Bershatsky <esq <at> lawlist.com>
> Cc:  77063 <at> debbugs.gnu.org
> 
> As to part 1 of 2, the goal is to be able to start Emacs 30.1, (require 'ido), evaluate (ido-read-buffer ...) and be greeted with a prompt displaying the ido rotating choices that looks like this:
> 
> PROMPT:  {*scratch* | *Messages* | *GNU Emacs*}
> 
> Starting from Emacs 30.1 without any user configuration, if we only evaluate (require 'ido) and then evaluate (ido-read-buffer ....), we do _not_ see any choices following the prompt.  This is because there needs to be call to ido-common-initialization _before_ calling ido-read-internal within ido-read-buffer.  Once ido-common-initialization has been run one time, all subsequent calls to ido-read-buffer work as expected.

Aren't you supposed to turn on ido-mode?  Just loading ido.el doesn't
have to be enough to activate the mode.

> As to part 2 of 2, we just press the TAB key after evaluating (ido-read-buffer ...), and C-x o a couple of times to place focus in the choices buffer and select the *Messages* buffer.  Once the *Messages* buffer has been selected and inserted into the minibuffer, we press the RET key and exit ido-read-buffer.  The result that is thrown is *scratch*, which we did not expect.

That's what I did, and the result was *Messages*.

Message #29 received at 77063 <at> debbugs.gnu.org (full text, mbox):

From: Keith David Bershatsky <esq <at> lawlist.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 77063 <at> debbugs.gnu.org
Subject: Re: bug#77063: 29.1; ido-read-buffer
Date: Mon, 17 Mar 2025 13:20:15 -0700

Ah ... yes, enabling ido-mode would also do the trick.  I saw that ido-completing-read calls ido-common-initialization before calling ido-read-internal, and that is why I concluded that ido-common-initialization should likewise precede ido-read-internal inside ido-read-buffer.  I use ido-completing-read quite a bit, but do not enable ido-mode.

Perhaps once upon a time someone somewhere tried to use ido-completing-read without enabling ido-mode, and he/she decided to add ido-common-initialization before ido-read-internal.  E.g., the usage of gnus-ido-completing-read without turning on ido-mode.

As to part 2 of 2, I cannot explain the difference in behavior.

;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;

> Date: [03-17-2025 12:25:22] <17 Mar 2025 21:25:22 +0200>
> From: Eli Zaretskii <eliz <at> gnu.org>
> To: Keith David Bershatsky <esq <at> lawlist.com>
> Cc: 77063 <at> debbugs.gnu.org
> Subject: Re: bug#77063: 29.1; ido-read-buffer
> 
> > Date:  Mon, 17 Mar 2025 09:50:37 -0700
> > From:  Keith David Bershatsky <esq <at> lawlist.com>
> > Cc:  77063 <at> debbugs.gnu.org
> >
> > As to part 1 of 2, the goal is to be able to start Emacs 30.1, (require 'ido), evaluate (ido-read-buffer ...) and be greeted with a prompt displaying the ido rotating choices that looks like this:
> >
> > PROMPT:  {*scratch* | *Messages* | *GNU Emacs*}
> >
> > Starting from Emacs 30.1 without any user configuration, if we only evaluate (require 'ido) and then evaluate (ido-read-buffer ....), we do _not_ see any choices following the prompt.  This is because there needs to be call to ido-common-initialization _before_ calling ido-read-internal within ido-read-buffer.  Once ido-common-initialization has been run one time, all subsequent calls to ido-read-buffer work as expected.
> 
> Aren't you supposed to turn on ido-mode?  Just loading ido.el doesn't
> have to be enough to activate the mode.
> 
> > As to part 2 of 2, we just press the TAB key after evaluating (ido-read-buffer ...), and C-x o a couple of times to place focus in the choices buffer and select the *Messages* buffer.  Once the *Messages* buffer has been selected and inserted into the minibuffer, we press the RET key and exit ido-read-buffer.  The result that is thrown is *scratch*, which we did not expect.
> 
> That's what I did, and the result was *Messages*.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.