bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Ihor Radchenko]

I just got the following error trying to run native-compiled time-to-days:

Debugger entered--Lisp error: (wrong-number-of-arguments (lambda (&optional time) (funcall #<subr decode-time> (or time G167))) 3)
  (decode-time (22352 22528) nil nil)
  (time-to-days (22352 22528))

Re-evaluating defun of time-to-days fixes the error.

Best,
Ihor


In GNU Emacs 28.0.50 (build 1, x86_64-pc-linux-gnu, cairo version 1.16.0)
 of 2021-04-30 built on localhost
Repository revision: fa65c044f2ebe666467166075c1507a8d0e1347f
Repository branch: feature/native-comp
Windowing system distributor 'The X.Org Foundation', version 11.0.12011000
System Description: Gentoo/Linux

Configured using:
 'configure --prefix=/usr --build=x86_64-pc-linux-gnu
 --host=x86_64-pc-linux-gnu --mandir=/usr/share/man
 --infodir=/usr/share/info --datadir=/usr/share --sysconfdir=/etc
 --localstatedir=/var/lib --disable-silent-rules
 --docdir=/usr/share/doc/emacs-28.0.9999
 --htmldir=/usr/share/doc/emacs-28.0.9999/html --libdir=/usr/lib64
 --program-suffix=-emacs-28-vcs --includedir=/usr/include/emacs-28-vcs
 --infodir=/usr/share/info/emacs-28-vcs --localstatedir=/var
 --enable-locallisppath=/etc/emacs:/usr/share/emacs/site-lisp
 --without-compress-install --without-hesiod --without-pop
 --with-file-notification=inotify --with-pdumper --enable-acl
 --with-dbus --with-modules --without-gameuser --with-libgmp
 --without-gpm --with-json --without-kerberos --without-kerberos5
 --without-lcms2 --with-xml2 --without-mailutils
 --with-native-compilation --with-selinux --with-gnutls
 --without-libsystemd --with-threads --with-wide-int --with-zlib
 --with-sound=oss --with-x --without-ns --without-gconf
 --without-gsettings --without-toolkit-scroll-bars --with-gif
 --with-jpeg --with-png --with-rsvg --with-tiff --with-xpm
 --with-imagemagick --with-xft --with-cairo --with-harfbuzz
 --without-libotf --without-m17n-flt --with-x-toolkit=no
 --with-dumping=pdumper 'CFLAGS=-march=native -pipe -O2' CPPFLAGS=
 'LDFLAGS=-Wl,-O1 -Wl,--as-needed''

Configured features:
ACL CAIRO DBUS FREETYPE GIF GLIB GMP GNUTLS HARFBUZZ IMAGEMAGICK JPEG
JSON LIBSELINUX LIBXML2 MODULES NATIVE_COMP NOTIFY INOTIFY OLDXMENU
PDUMPER PNG RSVG SECCOMP SOUND THREADS TIFF X11 XDBE XIM XPM ZLIB

Important settings:
  value of $LC_COLLATE: C
  value of $LANG: en_US.utf8
  locale-coding-system: utf-8-unix

Major mode: ELisp/l

Memory information:
((conses 16 2803364 652253)
 (symbols 48 82982 8)
 (strings 32 394883 248801)
 (string-bytes 1 13653046)
 (vectors 16 187919)
 (vector-slots 8 4341859 573177)
 (floats 8 17548 3388)
 (intervals 56 124000 2648)
 (buffers 992 103))

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Eli Zaretskii]

> From: Ihor Radchenko <yantar92@gmail.com>
> Date: Sat, 01 May 2021 09:47:35 +0800
> 
> I just got the following error trying to run native-compiled time-to-days:
> 
> Debugger entered--Lisp error: (wrong-number-of-arguments (lambda (&optional time) (funcall #<subr decode-time> (or time G167))) 3)
>   (decode-time (22352 22528) nil nil)
>   (time-to-days (22352 22528))

I cannot reproduce this.  When was your time-date.el
natively-compiled?  What happens if you delete the .eln file and
recompile it?

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Ihor Radchenko]

Eli Zaretskii <eliz@gnu.org> writes:

> I cannot reproduce this.  When was your time-date.el
> natively-compiled?

It was compiled during emacs installation (I am on Gentoo):

/usr/lib64/emacs/28.0.50/native-lisp/28.0.50-7ff4cc51:
  -rw-r--r--. 1  46K May  1 11:19 time-date-40951a48-0eafe94e.eln

I am not sure if it is important, but the problem appears in one of
org-mode tests. I just ran =make test= on current Org mode master.

> What happens if you delete the .eln file and
> recompile it?

I deleted the file and restarted emacs -Q. The new eln file is in 

/home/yantar92/.emacs.d/eln-cache/28.0.50-7ff4cc51:
-rwxr-xr-x. 1  59K May  1 14:58 time-date-40951a48-0eafe94e.eln

The problem persist. I also tried to delete all the org eln files in my
local .emacs.d as well as system-wide eln files generated during Emacs
installation. It did not help.

The full backtrace:

decode-time((22352 22528) nil nil)
  time-to-days((22352 22528))
  org-time-stamp-to-now("2016-06-03 Fri")
  org-deadline-close-p("2016-06-03 Fri" 0)
  apply(org-deadline-close-p ("2016-06-03 Fri" 0))
  (setq value-13032 (apply fn-13030 args-13031))
  (unwind-protect (setq value-13032 (apply fn-13030 args-13031)) (setq
  (if (unwind-protect (setq value-13032 (apply fn-13030 args-13031)) (
  (let (form-description-13034) (if (unwind-protect (setq value-13032 
  (let ((value-13032 'ert-form-evaluation-aborted-13033)) (let (form-d
  (let* ((fn-13030 #'org-deadline-close-p) (args-13031 (condition-case
  (progn (org-mode) (let ((point (string-match "<point>" inside-text))
  (unwind-protect (progn (org-mode) (let ((point (string-match "<point
  (save-current-buffer (set-buffer temp-buffer) (unwind-protect (progn
  (let ((temp-buffer (generate-new-buffer " *temp*" t))) (save-current
  (let ((inside-text (if (stringp "* Heading") "* Heading" (eval "* He
  (progn (fset 'time-subtract vnew) (fset 'time-less-p vnew) (fset 'ti
  (unwind-protect (progn (fset 'time-subtract vnew) (fset 'time-less-p
  (let* ((vnew #'(lambda nil G89)) (vnew #'(lambda (&optional time &re
  (let* ((G88 "2016-06-03 Fri 01:43") (G89 (if (stringp G88) (apply #'
  (lambda nil (let* ((G88 "2016-06-03 Fri 01:43") (G89 (if (stringp G8
  ert--run-test-internal(#s(ert--test-execution-info :test #s(ert-test
  ert-run-test(#s(ert-test :name test-org/deadline-close-p :documentat
  ert-run-or-rerun-test(#s(ert--stats :selector "\\(org\\|ob\\)" :test
  ert-run-tests("\\(org\\|ob\\)" #f(compiled-function (event-type &res
  ert-run-tests-batch("\\(org\\|ob\\)")
  ert-run-tests-batch-and-exit("\\(org\\|ob\\)")
  (let ((org-id-track-globally t) (org-test-selector (if org-test-sele
  org-test-run-batch-tests("\\(org\\|ob\\)")
  command-line-1(("--eval" "(setq vc-handled-backends nil org-startup-
  command-line()
  normal-top-level()

Best,
Ihor

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Eli Zaretskii]

> From: Ihor Radchenko <yantar92@gmail.com>
> CC: 48133@debbugs.gnu.org
> Date: Sat, 01 May 2021 15:11:36 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > I cannot reproduce this.  When was your time-date.el
> > natively-compiled?
> 
> It was compiled during emacs installation (I am on Gentoo):
> 
> /usr/lib64/emacs/28.0.50/native-lisp/28.0.50-7ff4cc51:
>   -rw-r--r--. 1  46K May  1 11:19 time-date-40951a48-0eafe94e.eln
> 
> I am not sure if it is important, but the problem appears in one of
> org-mode tests. I just ran =make test= on current Org mode master.

So maybe the problem is not in time-date.el, maybe it's elsewhere?
Can you show the exact code of the test that fails and how to run it?

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Ihor Radchenko]

Eli Zaretskii <eliz@gnu.org> writes:

> So maybe the problem is not in time-date.el, maybe it's elsewhere?
> Can you show the exact code of the test that fails and how to run it?

I run it by cloning Org mode repo from
https://code.orgmode.org/bzg/org-mode.git
and running =make test=

The failing test is test-org/deadline-close-p failing at the first
should assertion:

(ert-deftest test-org/deadline-close-p ()
  "Test `org-deadline-close-p' specifications."
  (org-test-at-time "2016-06-03 Fri 01:43"
    ;; Timestamps are close if they are within `ndays' of lead time.
    (org-test-with-temp-text "* Heading"
      (should (org-deadline-close-p "2016-06-03 Fri" 0))
      (should (org-deadline-close-p "2016-06-02 Thu" 0))
      (should-not (org-deadline-close-p "2016-06-04 Sat" 0))
      (should (org-deadline-close-p "2016-06-04 Sat" 1))
      (should (org-deadline-close-p "2016-06-03 Fri 12:00" 0)))
    ;; Read `ndays' from timestamp if argument not given.
    (org-test-with-temp-text "* H"
      (should (org-deadline-close-p "2016-06-04 Sat -1d"))
      (should-not (org-deadline-close-p "2016-06-04 Sat -0d"))
      (should (org-deadline-close-p "2016-06-10 Fri -1w"))
      (should-not (org-deadline-close-p "2016-06-11 Sat -1w")))
    ;; Prefer `ndays' argument over lead time in timestamp.
    (org-test-with-temp-text "* H"
      (should (org-deadline-close-p "2016-06-04 Sat -0d" 1))
      (should-not (org-deadline-close-p "2016-06-04 Sat -0d" 0)))
    ;; Completed tasks are never close.
    (let ((org-todo-keywords '(("TODO" "|" "DONE"))))
      (org-test-with-temp-text "* TODO Heading"
	(should (org-deadline-close-p "2016-06-03")))
      (org-test-with-temp-text "* DONE Heading"
	(should-not (org-deadline-close-p "2016-06-03"))))))

org-deadline-close-p code:

(defun org-deadline-close-p (timestamp-string &optional ndays)
  "Is the time in TIMESTAMP-STRING close to the current date?"
  (setq ndays (or ndays (org-get-wdays timestamp-string)))
  (and (<= (org-time-stamp-to-now timestamp-string) ndays)
       (not (org-entry-is-done-p))))

org-time-stamp-to-now:

(defun org-time-stamp-to-now (timestamp-string &optional seconds)
  "Difference between TIMESTAMP-STRING and now in days.
If SECONDS is non-nil, return the difference in seconds."
  (let ((fdiff (if seconds #'float-time #'time-to-days)))
    (- (funcall fdiff (org-time-string-to-time timestamp-string))
       (funcall fdiff nil))))

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Eli Zaretskii]

> From: Ihor Radchenko <yantar92@gmail.com>
> CC: 48133@debbugs.gnu.org
> Date: Sat, 01 May 2021 16:22:32 +0800
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > So maybe the problem is not in time-date.el, maybe it's elsewhere?
> > Can you show the exact code of the test that fails and how to run it?
> 
> I run it by cloning Org mode repo from
> https://code.orgmode.org/bzg/org-mode.git
> and running =make test=
> 
> The failing test is test-org/deadline-close-p failing at the first
> should assertion:

Thanks, I hope Andrea will be able to look into this soon.

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Andrea Corallo via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Ihor Radchenko <yantar92@gmail.com> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>> So maybe the problem is not in time-date.el, maybe it's elsewhere?
>> Can you show the exact code of the test that fails and how to run it?
>
> I run it by cloning Org mode repo from
> https://code.orgmode.org/bzg/org-mode.git
> and running =make test=

Just to confirm I can reproduce this.  Will look into it.

  Andrea

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Lars Ingebrigtsen]

Ihor Radchenko <yantar92@gmail.com> writes:

> I just got the following error trying to run native-compiled time-to-days:
>
> Debugger entered--Lisp error: (wrong-number-of-arguments (lambda (&optional time) (funcall #<subr decode-time> (or time G167))) 3)
>   (decode-time (22352 22528) nil nil)
>   (time-to-days (22352 22528))
>
> Re-evaluating defun of time-to-days fixes the error.

I'm just taking a stab in the dark here -- but it's really odd that it
seems to say that the signature of `decode-time' is

(&optional time)

Is it possible that the test code is mocking/redefining `decode-time'
with the wrong signature, or something along those lines?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Ihor Radchenko]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> I'm just taking a stab in the dark here -- but it's really odd that it
> seems to say that the signature of `decode-time' is
>
> (&optional time)
>
> Is it possible that the test code is mocking/redefining `decode-time'
> with the wrong signature, or something along those lines?

Yes, it is the case. `decode-time' is redefined via `cl-letf' form in
`org-test-at-time' macro in org-test.el:

((symbol-function 'decode-time)
	     (lambda (&optional time) (funcall ,(symbol-function 'decode-time)
					       (or time ,at))))

Hope it helps.

Best,
Ihor

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Lars Ingebrigtsen]

Ihor Radchenko <yantar92@gmail.com> writes:

> Yes, it is the case. `decode-time' is redefined via `cl-letf' form in
> `org-test-at-time' macro in org-test.el:
>
> ((symbol-function 'decode-time)
> 	     (lambda (&optional time) (funcall ,(symbol-function 'decode-time)
> 					       (or time ,at))))
>
> Hope it helps.

Yes, indeed.  :-)  So this is a bug in the org test harness, I think?
I've added Amin to the CCs.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Ihor Radchenko]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> Yes, indeed.  :-)  So this is a bug in the org test harness, I think?
> I've added Amin to the CCs.

While I agree that it is a bug on Org mode side, I am also wondering how
to handle backward-compatibility with older Emacs versions in such cases.
Is there any way to know in which Emacs version the function argument
list was changed?

Best,
Ihor

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Lars Ingebrigtsen]

Ihor Radchenko <yantar92@gmail.com> writes:

> While I agree that it is a bug on Org mode side, I am also wondering how
> to handle backward-compatibility with older Emacs versions in such cases.

In this case, adding the additional optional arguments won't have an
adverse effect in any Emacs version, I think?

> Is there any way to know in which Emacs version the function argument
> list was changed?

You'd have to ask git.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Ihor Radchenko]

Lars Ingebrigtsen <larsi@gnus.org> writes:

> In this case, adding the additional optional arguments won't have an
> adverse effect in any Emacs version, I think?

Sure. I was thinking about more dramatic changes in argument list. Think
swapped args.

>> Is there any way to know in which Emacs version the function argument
>> list was changed?
>
> You'd have to ask git.

Fair enough. Though I was hoping for something like "Probably introduced
at or before Emacs version 19.29.", but for argument list. Or even
incorporating git changelog for given function into the help buffer.

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Lars Ingebrigtsen]

Ihor Radchenko <yantar92@gmail.com> writes:

> Sure. I was thinking about more dramatic changes in argument list. Think
> swapped args.

I don't think we ever do things like that a lot.  (If ever.)

> Fair enough. Though I was hoping for something like "Probably introduced
> at or before Emacs version 19.29.", but for argument list. Or even
> incorporating git changelog for given function into the help buffer.

No, it's just too much work.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Ihor Radchenko]

> Yes, indeed.  :-)  So this is a bug in the org test harness, I think?

Just letting know that it is now fixed on Org mode master. This bug
report can be closed.

bug#48133: 28.0.50; [native-comp] time-to-days throws wrong-number-of-arguments error

[Lars Ingebrigtsen]

Ihor Radchenko <yantar92@gmail.com> writes:

>> Yes, indeed.  :-)  So this is a bug in the org test harness, I think?
>
> Just letting know that it is now fixed on Org mode master. This bug
> report can be closed.

OK; closing.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no