GNU bug report logs - #21963
Spurious byte compile message: "Warning: the function `bar' might not be defined at runtime."

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 21963 in the body.
You can then email your comments to 21963 AT debbugs.gnu.org in the normal way.

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

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

From: Alan Mackenzie <acm <at> muc.de>
To: bug-gnu-emacs <at> gnu.org
Subject: Spurious byte compile message: "Warning: the function `bar' might
 not be defined at runtime."
Date: Fri, 20 Nov 2015 10:44:19 +0000

Hello Emacs.

Using the latest from the emacs-25 branch:

This bug is a variant of bug #11918, which hasn't yet been resolved.

1. Create two Emacs Lisp source files, ~/warnings-bar.el and
~/warnings-foo.el with the following contents:

;;;;;;;; warnings-bar.el ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(defun bar ()
  "Doc string"
  "bar")

(provide 'warning-bar)
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;


;;;;;;;; warnings-foo.el ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
(eval-when-compile (require 'warning-bar))
(require 'warning-bar)

(defun foo ()
  (bar))
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;



2. (Optional) byte compile warnings-bar.el.
3. Byte compile warnings-foo.el with:
    emacs -Q -batch -L ~ -f batch-byte-compile warning-foo.el

.  This produces the spurious warning message:

    warning-foo.el:6:1:Warning: the function `bar' might not be defined
    at runtime.

.

N.B. if the `eval-when-compile' form is removed from warnings-foo.el, the
warning is not produced.

-- 
Alan Mackenzie (Nuremberg, Germany).

Message #8 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Alan Mackenzie <acm <at> muc.de>
Cc: 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the
 function	`bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 13:19:19 +0200

> Date: Fri, 20 Nov 2015 10:44:19 +0000
> From: Alan Mackenzie <acm <at> muc.de>
> 
> This bug is a variant of bug #11918, which hasn't yet been resolved.

I'm not sure this is the same issue.

> 1. Create two Emacs Lisp source files, ~/warnings-bar.el and
> ~/warnings-foo.el with the following contents:
> 
> ;;;;;;;; warnings-bar.el ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
> (defun bar ()
>   "Doc string"
>   "bar")
> 
> (provide 'warning-bar)
> ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
> 
> 
> ;;;;;;;; warnings-foo.el ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
> (eval-when-compile (require 'warning-bar))
> (require 'warning-bar)
> 
> (defun foo ()
>   (bar))
> ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
> 
> 
> 
> 2. (Optional) byte compile warnings-bar.el.
> 3. Byte compile warnings-foo.el with:
>     emacs -Q -batch -L ~ -f batch-byte-compile warning-foo.el
> 
> .  This produces the spurious warning message:
> 
>     warning-foo.el:6:1:Warning: the function `bar' might not be defined
>     at runtime.
> 
> .

eval-when-compile has no effect at run time, and the warning is
explicitly about what might happen at runtime.

IOW, I don't think this is a bug.

Message #11 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Alan Mackenzie <acm <at> muc.de>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 11:40:50 +0000

Hello, Eli.

On Fri, Nov 20, 2015 at 01:19:19PM +0200, Eli Zaretskii wrote:
> > Date: Fri, 20 Nov 2015 10:44:19 +0000
> > From: Alan Mackenzie <acm <at> muc.de>

> > This bug is a variant of bug #11918, which hasn't yet been resolved.

> I'm not sure this is the same issue.

Possibly not.

> > 1. Create two Emacs Lisp source files, ~/warnings-bar.el and
> > ~/warnings-foo.el with the following contents:

> > ;;;;;;;; warnings-bar.el ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
> > (defun bar ()
> >   "Doc string"
> >   "bar")

> > (provide 'warning-bar)
> > ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;


> > ;;;;;;;; warnings-foo.el ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
> > (eval-when-compile (require 'warning-bar))
> > (require 'warning-bar)

> > (defun foo ()
> >   (bar))
> > ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;



> > 2. (Optional) byte compile warnings-bar.el.
> > 3. Byte compile warnings-foo.el with:
> >     emacs -Q -batch -L ~ -f batch-byte-compile warning-foo.el

> > .  This produces the spurious warning message:

> >     warning-foo.el:6:1:Warning: the function `bar' might not be defined
> >     at runtime.

> > .

> eval-when-compile has no effect at run time, and the warning is
> explicitly about what might happen at runtime.

> IOW, I don't think this is a bug.

It is, at the very least, puzzling behaviour.  The warning is issued for
a function that most definitely has been defined at runtime.

What seems to be happening (I'm speculating a little) is that the byte
compiler maintains a list of not-yet-defined functions in
byte-compile-unresolved-functions.  Functions defined at compile time
are pushed onto this list.  The problem is that when they are later
defined for run time, they aren't taken off the list.

Given that we've all been asked to get rid of warnings for 25.1, I have
to disagree with you about whether this is a bug.

-- 
Alan Mackenzie (Nuremberg, Germany).

Message #14 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Alan Mackenzie <acm <at> muc.de>
Cc: 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the
 function	`bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 16:48:13 +0200

> Date: Fri, 20 Nov 2015 11:40:50 +0000
> Cc: 21963 <at> debbugs.gnu.org
> From: Alan Mackenzie <acm <at> muc.de>
> 
> > eval-when-compile has no effect at run time, and the warning is
> > explicitly about what might happen at runtime.
> 
> > IOW, I don't think this is a bug.
> 
> It is, at the very least, puzzling behaviour.  The warning is issued for
> a function that most definitely has been defined at runtime.

No, it was not (or, more accurately, will not be).  'eval-when-compile'
does nothing to make sure it will be defined at runtime; you need
either 'require' or 'eval-and-compile'.

Message #17 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Alan Mackenzie <acm <at> muc.de>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 15:26:38 +0000

Hello, Eli.

On Fri, Nov 20, 2015 at 04:48:13PM +0200, Eli Zaretskii wrote:
> > Date: Fri, 20 Nov 2015 11:40:50 +0000
> > Cc: 21963 <at> debbugs.gnu.org
> > From: Alan Mackenzie <acm <at> muc.de>

> > > eval-when-compile has no effect at run time, and the warning is
> > > explicitly about what might happen at runtime.

> > > IOW, I don't think this is a bug.

> > It is, at the very least, puzzling behaviour.  The warning is issued for
> > a function that most definitely has been defined at runtime.

> No, it was not (or, more accurately, will not be).  'eval-when-compile'
> does nothing to make sure it will be defined at runtime; you need
> either 'require' or 'eval-and-compile'.

The first two lines of the file are:

   (eval-when-compile (require 'warning-bar))
   (require 'warning-bar)
 
The second line will define the function `bar' at run time.

The byte-compiler has already decided after the first line that `bar' is
not defined at run time (fair enough), but seems unwilling to change its
mind after the second line.

This isn't just an academic problem.  CC Mode uses such sequences (in
effect), producing warnings in the Emacs build.

-- 
Alan Mackenzie (Nuremberg, Germany).

Message #20 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Alan Mackenzie <acm <at> muc.de>
Cc: 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the
 function	`bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 17:48:49 +0200

> Date: Fri, 20 Nov 2015 15:26:38 +0000
> Cc: 21963 <at> debbugs.gnu.org
> From: Alan Mackenzie <acm <at> muc.de>
> 
> The first two lines of the file are:
> 
>    (eval-when-compile (require 'warning-bar))
>    (require 'warning-bar)

I think if you leave only the second one, Bob's your uncle.

Message #23 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: John Wiegley <jwiegley <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Alan Mackenzie <acm <at> muc.de>, 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 08:01:31 -0800

>>>>> Eli Zaretskii <eliz <at> gnu.org> writes:

>> (eval-when-compile (require 'warning-bar))
>> (require 'warning-bar)

> I think if you leave only the second one, Bob's your uncle.

I think the question is: Why does haven't the first one cause the definedness
introduced by the second to not be recognized?

John

Message #26 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Alan Mackenzie <acm <at> muc.de>
To: John Wiegley <jwiegley <at> gmail.com>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 16:15:35 +0000

Hello, John.

On Fri, Nov 20, 2015 at 08:01:31AM -0800, John Wiegley wrote:
> >>>>> Eli Zaretskii <eliz <at> gnu.org> writes:

> >> (eval-when-compile (require 'warning-bar))
> >> (require 'warning-bar)

> > I think if you leave only the second one, Bob's your uncle.

> I think the question is: Why does haven't the first one cause the definedness
> introduced by the second to not be recognized?

While eval-when-compile'ing the first line, the byte compiler collects a
list of symbols defined by it (in variable
byte-compile-noruntime-functions).  More or less.

There is no mechanism for taking symbols off of that list when they
subsequently become defined.

> John

-- 
Alan Mackenzie (Nuremberg, Germany).

Message #29 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: John Wiegley <jwiegley <at> gmail.com>
To: Alan Mackenzie <acm <at> muc.de>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 08:44:56 -0800

>>>>> Alan Mackenzie <acm <at> muc.de> writes:

> While eval-when-compile'ing the first line, the byte compiler collects a
> list of symbols defined by it (in variable
> byte-compile-noruntime-functions). More or less.

> There is no mechanism for taking symbols off of that list when they
> subsequently become defined.

Not optimal, but I must ask: Should it be fixed? Or, for now, simply noted in
the elisp manual when discussing the merits of this warning?

John

Message #32 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Alan Mackenzie <acm <at> muc.de>
To: John Wiegley <jwiegley <at> gmail.com>
Cc: Eli Zaretskii <eliz <at> gnu.org>, 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Fri, 20 Nov 2015 17:50:34 +0000

Hello, John.

On Fri, Nov 20, 2015 at 08:44:56AM -0800, John Wiegley wrote:
> >>>>> Alan Mackenzie <acm <at> muc.de> writes:

> > While eval-when-compile'ing the first line, the byte compiler collects a
> > list of symbols defined by it (in variable
> > byte-compile-noruntime-functions). More or less.

> > There is no mechanism for taking symbols off of that list when they
> > subsequently become defined.

> Not optimal, but I must ask: Should it be fixed? Or, for now, simply noted in
> the elisp manual when discussing the merits of this warning?

I think it should be fixed.  It will eliminate warnings from CC Mode (at
the very least).

I am working on it as we speak.  I expect to produce a patch for review
soon.

> John

-- 
Alan Mackenzie (Nuremberg, Germany).

Message #35 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefan <at> marxist.se>
To: Alan Mackenzie <acm <at> muc.de>
Cc: John Wiegley <jwiegley <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>,
 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Sun, 29 Sep 2019 23:24:29 +0200

Alan Mackenzie <acm <at> muc.de> writes:

> On Fri, Nov 20, 2015 at 08:44:56AM -0800, John Wiegley wrote:
>> >>>>> Alan Mackenzie <acm <at> muc.de> writes:
>
>> > While eval-when-compile'ing the first line, the byte compiler collects a
>> > list of symbols defined by it (in variable
>> > byte-compile-noruntime-functions). More or less.
>
>> > There is no mechanism for taking symbols off of that list when they
>> > subsequently become defined.
>
>> Not optimal, but I must ask: Should it be fixed? Or, for now, simply noted in
>> the elisp manual when discussing the merits of this warning?
>
> I think it should be fixed.  It will eliminate warnings from CC Mode (at
> the very least).
>
> I am working on it as we speak.  I expect to produce a patch for review
> soon.

Hi Alan,

That was four years ago.  Did you finish the patch which would resolve
this issue?

Best regards,
Stefan Kangas

Message #38 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Alan Mackenzie <acm <at> muc.de>
To: Stefan Kangas <stefan <at> marxist.se>
Cc: John Wiegley <jwiegley <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>,
 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Mon, 30 Sep 2019 18:22:27 +0000

Hello, Stefan.

On Sun, Sep 29, 2019 at 23:24:29 +0200, Stefan Kangas wrote:
> Alan Mackenzie <acm <at> muc.de> writes:

> > On Fri, Nov 20, 2015 at 08:44:56AM -0800, John Wiegley wrote:
> >> >>>>> Alan Mackenzie <acm <at> muc.de> writes:

> >> > While eval-when-compile'ing the first line, the byte compiler collects a
> >> > list of symbols defined by it (in variable
> >> > byte-compile-noruntime-functions). More or less.

> >> > There is no mechanism for taking symbols off of that list when they
> >> > subsequently become defined.

> >> Not optimal, but I must ask: Should it be fixed? Or, for now, simply noted in
> >> the elisp manual when discussing the merits of this warning?

> > I think it should be fixed.  It will eliminate warnings from CC Mode (at
> > the very least).

> > I am working on it as we speak.  I expect to produce a patch for review
> > soon.

> Hi Alan,

> That was four years ago.  Did you finish the patch which would resolve
> this issue?

Yes, I did indeed; with this patch:

acb96f2718ccb0d36af514ce63b5decf0f24a697
Squashed commit of the following: .....
Tue Nov 24 17:37:49 2015 +0000

Sorry I forgot to close the bug.

> Best regards,
> Stefan Kangas

-- 
Alan Mackenzie (Nuremberg, Germany).

Message #41 received at 21963 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefan <at> marxist.se>
To: Alan Mackenzie <acm <at> muc.de>
Cc: John Wiegley <jwiegley <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>,
 21963 <at> debbugs.gnu.org
Subject: Re: bug#21963: Spurious byte compile message: "Warning: the function
 `bar' might not be defined at runtime."
Date: Mon, 30 Sep 2019 23:24:06 +0200

close 21963 25.1
quit

> > That was four years ago.  Did you finish the patch which would resolve
> > this issue?
>
> Yes, I did indeed; with this patch:
>
> acb96f2718ccb0d36af514ce63b5decf0f24a697
> Squashed commit of the following: .....
> Tue Nov 24 17:37:49 2015 +0000
>
> Sorry I forgot to close the bug.

Thanks; closed now.

Best regards,
Stefan Kangas

Previous Next

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