GNU bug report logs - #27706
MacOS: decode-time hang

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 27706 in the body.
You can then email your comments to 27706 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: "Charles A. Roelli" <charles <at> aurox.ch>
To: bug-gnu-emacs <at> gnu.org
Subject: MacOS: decode-time hang
Date: Sat, 15 Jul 2017 16:19:39 +0200

Evaluating the following form on my system (macOS 10.6) results in a
hang from Emacs 23 to master:

(decode-time '(-1034058203136 0))

[other values around -1034058203136 don't cause an issue]

The stack trace (stuck in macOS' libc, I think):

(gdb) bt full
#0  0x00007fff83860aef in timesub () from /usr/lib/libSystem.B.dylib
No symbol table info available.
#1  0x00007fff8386081c in _st_localsub () from /usr/lib/libSystem.B.dylib
No symbol table info available.
#2  0x00007fff83869d71 in localtime_r () from /usr/lib/libSystem.B.dylib
No symbol table info available.
#3  0x00000001004586c7 in localtime_rz (tz=0x10210bba0, 
t=0x7fff5fbf8d20, tm=0x7fff5fbf8ce8) at time_rz.c:296
        abbr_saved = false
        old_tz = 0x1
#4  0x00000001002ba08b in emacs_localtime_rz (tz=0x10210bba0, 
t=0x7fff5fbf8d20, tm=0x7fff5fbf8ce8) at editfns.c:103
No locals.
#5  0x00000001002c2680 in Fdecode_time (specified_time=..., zone=...) at 
editfns.c:2163
        time_spec = -67768038400720896
        gmt_tm = {
          tm_sec = 1606388960,
          tm_min = 32767,
          tm_hour = 1766166,
          tm_mday = 1,
          tm_mon = 0,
          tm_year = 0,
          tm_wday = 0,
          tm_yday = 0,
          tm_isdst = 0,
          tm_gmtoff = 10261984,
          tm_zone = 0x7fff5fbf8d30 "`\215\277_\377\177"
        }
        tz = 0x10210bba0
        local_tm = {
          tm_sec = 24,
          tm_min = 38,
          tm_hour = 8,
          tm_mday = 32767,
          tm_mon = 0,
          tm_year = 0,
          tm_wday = 0,
          tm_yday = 0,
          tm_isdst = 0,
          tm_gmtoff = 0,
          tm_zone = 0x0
        }
        tm = 0x0
        tm_year_base = 0

The form is called when compiling the new org-timer.el, which requires
org-clock.el, which itself contains this call that causes the issue:

(defconst org-clock--oldest-date
  (let* ((dichotomy
	  (lambda (min max pred)
	    (if (funcall pred min) min
	      (cl-incf min)
	      (while (> (- max min) 1)
		(let ((mean (+ (ash min -1) (ash max -1) (logand min max 1))))
		  (if (funcall pred mean) (setq max mean) (setq min mean)))))
	    max))
	 (high
	  (funcall dichotomy
		   most-negative-fixnum
		   0
		   (lambda (m) (ignore-errors (decode-time (list m 0))))))
	 (low
	  (funcall dichotomy
		   most-negative-fixnum
		   0
		   (lambda (m) (ignore-errors (decode-time (list high m)))))))
    (list high low))
  "Internal time for oldest date representable on the system.")

As a result, "make" hangs for me, so I delete org-timer.el locally to
get around the problem for now.  But I'd like to find a more permanent
solution to make sure this doesn't happen elsewhere.

Can anyone reproduce this under macOS, and if so, under which macOS
version?  I'd like to know which versions might need a fix for this
problem.

Message #8 received at 27706 <at> debbugs.gnu.org (full text, mbox):

From: Alan Third <alan <at> idiocy.org>
To: "Charles A. Roelli" <charles <at> aurox.ch>
Cc: 27706 <at> debbugs.gnu.org
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sat, 15 Jul 2017 22:32:03 +0100

On Sat, Jul 15, 2017 at 04:19:39PM +0200, Charles A. Roelli wrote:
> Evaluating the following form on my system (macOS 10.6) results in a
> hang from Emacs 23 to master:
> 
> (decode-time '(-1034058203136 0))
> 
> [other values around -1034058203136 don't cause an issue]

I can’t reproduce in 10.12.
-- 
Alan Third

Message #11 received at 27706 <at> debbugs.gnu.org (full text, mbox):

From: "Charles A. Roelli" <charles <at> aurox.ch>
To: Alan Third <alan <at> idiocy.org>
Cc: 27706 <at> debbugs.gnu.org
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sun, 16 Jul 2017 18:11:23 +0200

Thanks for testing it.


> On 15 Jul 2017, at 23:32, Alan Third <alan <at> idiocy.org> wrote:
> 
>> On Sat, Jul 15, 2017 at 04:19:39PM +0200, Charles A. Roelli wrote:
>> Evaluating the following form on my system (macOS 10.6) results in a
>> hang from Emacs 23 to master:
>> 
>> (decode-time '(-1034058203136 0))
>> 
>> [other values around -1034058203136 don't cause an issue]
> 
> I can’t reproduce in 10.12.
> -- 
> Alan Third

Message #16 received at 27706 <at> debbugs.gnu.org (full text, mbox):

From: charles <at> aurox.ch (Charles A. Roelli)
To: 27706 <at> debbugs.gnu.org
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sat, 30 Sep 2017 21:09:08 +0200

> From: "Charles A. Roelli" <charles <at> aurox.ch>
> Date: Sat, 15 Jul 2017 16:19:39 +0200
> 
> Evaluating the following form on my system (macOS 10.6) results in a
> hang from Emacs 23 to master:
> 
> (decode-time '(-1034058203136 0))
> 
> [other values around -1034058203136 don't cause an issue]
> 
> The stack trace (stuck in macOS' libc, I think):
> 
> [...]
> 
> The form is called when compiling the new org-timer.el, which requires
> org-clock.el, which itself contains this call that causes the issue:
> 
> (defconst org-clock--oldest-date
>    (let* ((dichotomy
> 	  (lambda (min max pred)
> 	    (if (funcall pred min) min
> 	      (cl-incf min)
> 	      (while (> (- max min) 1)
> 		(let ((mean (+ (ash min -1) (ash max -1) (logand min max 1))))
> 		  (if (funcall pred mean) (setq max mean) (setq min mean)))))
> 	    max))
> 	 (high
> 	  (funcall dichotomy
> 		   most-negative-fixnum
> 		   0
> 		   (lambda (m) (ignore-errors (decode-time (list m 0))))))
> 	 (low
> 	  (funcall dichotomy
> 		   most-negative-fixnum
> 		   0
> 		   (lambda (m) (ignore-errors (decode-time (list high m)))))))
>      (list high low))
>    "Internal time for oldest date representable on the system.")
> 
> As a result, "make" hangs for me, so I delete org-timer.el locally to
> get around the problem for now.  But I'd like to find a more permanent
> solution to make sure this doesn't happen elsewhere.
> 
> Can anyone reproduce this under macOS, and if so, under which macOS
> version?  I'd like to know which versions might need a fix for this
> problem.

I'd like to fix this for emacs-26.  Is this fix okay to apply?

From b7a871e126b6e2036dbbe81b5c8b36bfbf6da419 Mon Sep 17 00:00:00 2001
From: "Charles A. Roelli" <charles <at> aurox.ch>
Date: Sat, 30 Sep 2017 20:42:03 +0200
Subject: [PATCH] Workaround for faulty localtime() under macOS 10.6

* lisp/org/org-clock.el (org-clock--oldest-date): Only execute
'decode-time' on times later than year -2**31 under macOS 10.6.
See Bug#27706.
---
 lisp/org/org-clock.el | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

diff --git a/lisp/org/org-clock.el b/lisp/org/org-clock.el
index 8df185d..97d9612 100644
--- a/lisp/org/org-clock.el
+++ b/lisp/org/org-clock.el
@@ -478,7 +478,18 @@ org-clock--oldest-date
 	  (funcall dichotomy
 		   most-negative-fixnum
 		   0
-		   (lambda (m) (ignore-errors (decode-time (list m 0))))))
+		   (lambda (m)
+                     ;; libc in macOS 10.6 hangs when decoding times
+                     ;; around year -2**31.  Limit `high' not to go
+                     ;; any earlier than that.
+                     (if (and (featurep 'ns)
+                              (string-match-p
+                               "10\\.6\\.[[:digit:]]"
+                               (shell-command-to-string
+                                "sw_vers -productVersion")))
+                         (if (> m -1034058203136)
+                             (ignore-errors (decode-time (list m 0))))
+                       (ignore-errors (decode-time (list m 0)))))))
 	 (low
 	  (funcall dichotomy
 		   most-negative-fixnum
-- 
2.9.4

Message #19 received at 27706 <at> debbugs.gnu.org (full text, mbox):

From: Alan Third <alan <at> idiocy.org>
To: "Charles A. Roelli" <charles <at> aurox.ch>
Cc: 27706 <at> debbugs.gnu.org
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sat, 30 Sep 2017 21:21:02 +0100

On Sat, Sep 30, 2017 at 09:09:08PM +0200, Charles A. Roelli wrote:
> +                     (if (and (featurep 'ns)
> +                              (string-match-p
> +                               "10\\.6\\.[[:digit:]]"
> +                               (shell-command-to-string
> +                                "sw_vers -productVersion")))

I can’t speak for the issue generally as I don’t know what this code
does, but this particular check looks incorrect to me.

sw_vers doesn’t exist on GNU/Linux, and probably other platforms that
GNUstep runs on, so you can’t just use a check for NS. Probably you
need to check for Darwin, which should work on terminal only versions
of Emacs too.

    (string-equal system-type "darwin")

(There may be other ways to check, this is the only one I know.)
-- 
Alan Third

Message #22 received at 27706 <at> debbugs.gnu.org (full text, mbox):

From: Noam Postavsky <npostavs <at> users.sourceforge.net>
To: Alan Third <alan <at> idiocy.org>
Cc: "Charles A. Roelli" <charles <at> aurox.ch>, 27706 <at> debbugs.gnu.org
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sat, 30 Sep 2017 17:21:49 -0400

Alan Third <alan <at> idiocy.org> writes:

>     (string-equal system-type "darwin")

`system-type' contains a symbol, so generally the comparison should look
like this (unless there is some strong reason to do string comparison)

    (eq system-type 'darwin)

Message #25 received at 27706 <at> debbugs.gnu.org (full text, mbox):

From: charles <at> aurox.ch (Charles A. Roelli)
To: Noam Postavsky <npostavs <at> users.sourceforge.net>
Cc: alan <at> idiocy.org, 27706 <at> debbugs.gnu.org
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sun, 01 Oct 2017 11:27:38 +0200

> From: Noam Postavsky <npostavs <at> users.sourceforge.net>
> Date: Sat, 30 Sep 2017 17:21:49 -0400
> 
> Alan Third <alan <at> idiocy.org> writes:
> 
> >     (string-equal system-type "darwin")
> 
> `system-type' contains a symbol, so generally the comparison should look
> like this (unless there is some strong reason to do string comparison)
> 
>     (eq system-type 'darwin)

Thanks to both of you for your help.  Updated patch follows.  I've
also made the check a bit easier to read, and verified that it also
works on the Darwin terminal.

From 7ddfdae2359c07bdddeeeafd69517e02ee3ee2f5 Mon Sep 17 00:00:00 2001
From: "Charles A. Roelli" <charles <at> aurox.ch>
Date: Sat, 30 Sep 2017 20:42:03 +0200
Subject: [PATCH] Workaround for faulty localtime() under macOS 10.6

* lisp/org/org-clock.el (org-clock--oldest-date): Only execute
'decode-time' on times later than year -2**31 under macOS 10.6.
See Bug#27706.
---
 lisp/org/org-clock.el | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/lisp/org/org-clock.el b/lisp/org/org-clock.el
index 8df185d..2eec817 100644
--- a/lisp/org/org-clock.el
+++ b/lisp/org/org-clock.el
@@ -478,7 +478,17 @@ org-clock--oldest-date
 	  (funcall dichotomy
 		   most-negative-fixnum
 		   0
-		   (lambda (m) (ignore-errors (decode-time (list m 0))))))
+		   (lambda (m)
+                     ;; libc in macOS 10.6 hangs when decoding times
+                     ;; around year -2**31.  Limit `high' not to go
+                     ;; any earlier than that.
+                     (unless (and (eq system-type 'darwin)
+                                  (string-match-p
+                                   "10\\.6\\.[[:digit:]]"
+                                   (shell-command-to-string
+                                    "sw_vers -productVersion"))
+                                  (<= m -1034058203136))
+                       (ignore-errors (decode-time (list m 0)))))))
 	 (low
 	  (funcall dichotomy
 		   most-negative-fixnum
-- 
2.9.4

Message #28 received at 27706 <at> debbugs.gnu.org (full text, mbox):

From: charles <at> aurox.ch (Charles A. Roelli)
To: 27706 <at> debbugs.gnu.org
Cc: alan <at> idiocy.org, npostavs <at> users.sourceforge.net
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sun, 01 Oct 2017 19:11:01 +0200

I pushed the fix to emacs-26:

  commit bd49b6f1b39cffeaf6098bc7b0182552683b1c07

  Workaround for faulty localtime() under macOS 10.6

  * lisp/org/org-clock.el (org-clock--oldest-date): Only execute
  'decode-time' on times later than year -2**31 under macOS 10.6.
  See Bug#27706.

I'll close this bug in the next few days.

Message #33 received at 27706-done <at> debbugs.gnu.org (full text, mbox):

From: charles <at> aurox.ch (Charles A. Roelli)
To: charles <at> aurox.ch (Charles A. Roelli)
Cc: alan <at> idiocy.org, npostavs <at> users.sourceforge.net, 27706-done <at> debbugs.gnu.org
Subject: Re: bug#27706: MacOS: decode-time hang
Date: Sat, 07 Oct 2017 17:33:24 +0200

> Date: Sun, 01 Oct 2017 19:11:01 +0200
> From: charles <at> aurox.ch (Charles A. Roelli)
> CC: npostavs <at> users.sourceforge.net, alan <at> idiocy.org
> 
> I pushed the fix to emacs-26:
> 
>   commit bd49b6f1b39cffeaf6098bc7b0182552683b1c07
> 
>   Workaround for faulty localtime() under macOS 10.6
> 
>   * lisp/org/org-clock.el (org-clock--oldest-date): Only execute
>   'decode-time' on times later than year -2**31 under macOS 10.6.
>   See Bug#27706.
> 
> I'll close this bug in the next few days.

Works, closing.

Previous Next

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