Commit 0921dbc3 authored by Paul Eggert's avatar Paul Eggert
Browse files

Improve time stamp handling, and be more consistent about it.

This implements a suggestion made in:
http://lists.gnu.org/archive/html/emacs-devel/2014-10/msg00587.html
Among other things, this means timer.el no longer needs to
autoload the time-date module.
* doc/lispref/os.texi (Time of Day, Time Conversion, Time Parsing)
(Processor Run Time, Time Calculations):
Document the new behavior, plus be clearer about the old behavior.
(Idle Timers): Take advantage of new functionality.
* etc/NEWS: Document the changes.
* lisp/allout-widgets.el (allout-elapsed-time-seconds): Doc fix.
* lisp/arc-mode.el (archive-ar-summarize):
* lisp/calendar/time-date.el (seconds-to-time, days-to-time, time-since):
* lisp/emacs-lisp/timer.el (timer-relative-time, timer-event-handler)
(run-at-time, with-timeout-suspend, with-timeout-unsuspend):
* lisp/net/tramp.el (tramp-time-less-p, tramp-time-subtract):
* lisp/proced.el (proced-time-lessp):
* lisp/timezone.el (timezone-time-from-absolute):
* lisp/type-break.el (type-break-schedule, type-break-time-sum):
Simplify by using new functionality.
* lisp/calendar/cal-dst.el (calendar-next-time-zone-transition):
Do not return time values in obsolete and undocumented (HI . LO)
format; use (HI LO) instead.
* lisp/calendar/time-date.el (with-decoded-time-value):
Treat 'nil' as current time.  This is mostly for XEmacs.
(encode-time-value, with-decoded-time-value): Obsolete.
(time-add, time-subtract, time-less-p): Use no-op autoloads, for
XEmacs.  Define only if XEmacs, as they're now C builtins in Emacs.
* lisp/ldefs-boot.el: Update to match new time-date.el
* lisp/proced.el: Do not require time-date.
* src/editfns.c (invalid_time): New function.
Use it instead of 'error ("Invalid time specification")'.
(time_add, time_subtract, time_arith, Ftime_add, Ftime_less_p)
(decode_float_time, lisp_to_timespec, lisp_time_struct):
New functions.
(make_time_tail, make_time): Remove.  All uses changed to use
new functions or plain list4i.
(disassemble_lisp_time): Return effective length if successful.
Check that LOW is an integer, if it's combined with other components.
(decode_time_components): Decode into struct lisp_time, not
struct timespec, so that we can support a wide set of times
regardless of whether time_t is signed.  Decode plain numbers
as seconds since the Epoch, and nil as the current time.
(lisp_time_argument, lisp_seconds_argument, Ffloat_time):
Reimplement in terms of new functions.
(Fencode_time): Just use list2i.
(syms_of_editfns): Add time-add, time-subtract, time-less-p.
* src/keyboard.c (decode_timer): Don't allow the new formats (floating
point or nil) in timers.
* src/systime.h (LO_TIME_BITS): New constant.  Use it everywhere in
place of the magic number '16'.
(struct lisp_time): New type.
(decode_time_components): Use it.
(lisp_to_timespec): New decl.
parent 058f56d2
2014-11-17 Paul Eggert <eggert@cs.ucla.edu>
Improve time stamp handling, and be more consistent about it.
* os.texi (Time of Day, Time Conversion, Time Parsing)
(Processor Run Time, Time Calculations):
Document the new behavior, plus be clearer about the old behavior.
(Idle Timers): Take advantage of new functionality.
2014-11-16 Lars Magne Ingebrigtsen <larsi@gnus.org> 2014-11-16 Lars Magne Ingebrigtsen <larsi@gnus.org>
* text.texi (Special Properties): Mention `inhibit-read-only'. * text.texi (Special Properties): Mention `inhibit-read-only'.
......
...@@ -1213,37 +1213,34 @@ return value is @code{nil}. ...@@ -1213,37 +1213,34 @@ return value is @code{nil}.
zone. zone.
@cindex epoch @cindex epoch
Most of these functions represent time as a list of either four Most of these functions represent time as a list of four integers
integers, @code{(@var{sec-high} @var{sec-low} @var{microsec} @code{(@var{sec-high} @var{sec-low} @var{microsec} @var{picosec})}.
@var{picosec})}, or of three This represents the number of seconds from the @dfn{epoch} (January
integers, @code{(@var{sec-high} @var{sec-low} @var{microsec})}, or of 1, 1970 at 00:00 UTC), using the formula:
two integers, @code{(@var{sec-high} @var{sec-low})}. The integers
@var{sec-high} and @var{sec-low} give the high and low bits of an
integer number of seconds. This integer,
@ifnottex @ifnottex
@var{high} * 2**16 + @var{low}, @var{high} * 2**16 + @var{low} + @var{micro} * 10**@minus{}6 +
@var{pico} * 10**@minus{}12.
@end ifnottex @end ifnottex
@tex @tex
$high*2^{16}+low$, $high*2^{16} + low + micro*10^{-6} + pico*10^{-12}$.
@end tex @end tex
is the number of seconds from the @dfn{epoch} (January 1, 1970 at 00:00 The return value of @code{current-time} represents time using this
UTC) to the specified time. The third list element @var{microsec}, if form, as do the timestamps in the return values of other functions
present, gives the number of microseconds from the start of that such as @code{file-attributes} (@pxref{Definition of
second to the specified time. file-attributes}). In some cases, functions may return two- or
Similarly, the fourth list element @var{picosec}, if present, gives three-element lists, with omitted @var{microsec} and @var{picosec}
the number of picoseconds from the start of that microsecond to the components defaulting to zero.
specified time.
@cindex time value
The return value of @code{current-time} represents time using four Function arguments, e.g., the @var{time-value} argument to
integers, as do the timestamps in the return value of @code{current-time-string}, accept a more-general @dfn{time value}
@code{file-attributes} (@pxref{Definition of format, which can be a list of integers as above, or a single number
file-attributes}). In function arguments, e.g., the @var{time-value} for seconds since the epoch, or @code{nil} for the current time. You
argument to @code{current-time-string}, two-, three-, and four-integer can convert a time value into a human-readable string using
lists are accepted. You can convert times from the list @code{current-time-string} and @code{format-time-string}, into a list
representation into standard human-readable strings using of integers using @code{seconds-to-time}, and into other forms using
@code{current-time-string}, or to other forms using the @code{decode-time} and @code{float-time}. These functions are
@code{decode-time} and @code{format-time-string} functions documented described in the following sections.
in the following sections.
@defun current-time-string &optional time-value @defun current-time-string &optional time-value
This function returns the current time and date as a human-readable This function returns the current time and date as a human-readable
...@@ -1256,8 +1253,8 @@ characters from the beginning of the string rather than from the end, ...@@ -1256,8 +1253,8 @@ characters from the beginning of the string rather than from the end,
as the year might not have exactly four digits, and additional as the year might not have exactly four digits, and additional
information may some day be added at the end. information may some day be added at the end.
The argument @var{time-value}, if given, specifies a time to format The argument @var{time-value}, if given, specifies a time to format,
(represented as a list of integers), instead of the current time. instead of the current time.
@example @example
@group @group
...@@ -1279,11 +1276,19 @@ become available. ...@@ -1279,11 +1276,19 @@ become available.
@defun float-time &optional time-value @defun float-time &optional time-value
This function returns the current time as a floating-point number of This function returns the current time as a floating-point number of
seconds since the epoch. The optional argument @var{time-value}, if seconds since the epoch. The optional argument @var{time-value}, if
given, specifies a time (represented as a list of integers) to convert given, specifies a time to convert instead of the current time.
instead of the current time.
@emph{Warning}: Since the result is floating point, it may not be @emph{Warning}: Since the result is floating point, it may not be
exact. Do not use this function if precise time stamps are required. exact. Do not use this function if precise time stamps are required.
@code{time-to-seconds} is an alias for this function.
@end defun
@defun seconds-to-time time-value
This function converts a time value to list-of-integer form.
For example, if @var{time-value} is a number, @code{(time-to-seconds
(seconds-to-time @var{time-value}))} equals the number unless overflow
or rounding errors occur.
@end defun @end defun
@defun current-time-zone &optional time-value @defun current-time-zone &optional time-value
...@@ -1302,8 +1307,8 @@ adjustment, then the value is constant through time. ...@@ -1302,8 +1307,8 @@ adjustment, then the value is constant through time.
If the operating system doesn't supply all the information necessary to If the operating system doesn't supply all the information necessary to
compute the value, the unknown elements of the list are @code{nil}. compute the value, the unknown elements of the list are @code{nil}.
The argument @var{time-value}, if given, specifies a time (represented The argument @var{time-value}, if given, specifies a time value to
as a list of integers) to analyze instead of the current time. analyze instead of the current time.
@end defun @end defun
The current time zone is determined by the @env{TZ} environment The current time zone is determined by the @env{TZ} environment
...@@ -1316,15 +1321,15 @@ time zone. ...@@ -1316,15 +1321,15 @@ time zone.
@section Time Conversion @section Time Conversion
@cindex calendrical information @cindex calendrical information
These functions convert time values (lists of two to four integers, These functions convert time values (@pxref{Time of Day}) into
as explained in the previous section) into calendrical information and calendrical information and vice versa.
vice versa.
Many 32-bit operating systems are limited to time values containing Many 32-bit operating systems are limited to system times containing
32 bits of information; these systems typically handle only the times 32 bits of information in their seconds component; these systems
from 1901-12-13 20:45:52 UTC through 2038-01-19 03:14:07 UTC@. typically handle only the times from 1901-12-13 20:45:52 UTC through
However, 64-bit and some 32-bit operating systems have larger time 2038-01-19 03:14:07 UTC@. However, 64-bit and some 32-bit operating
values, and can represent times far in the past or future. systems have larger seconds components, and can represent times far in
the past or future.
Time conversion functions always use the Gregorian calendar, even Time conversion functions always use the Gregorian calendar, even
for dates before the Gregorian calendar was introduced. Year numbers for dates before the Gregorian calendar was introduced. Year numbers
...@@ -1332,9 +1337,9 @@ count the number of years since the year 1 B.C., and do not skip zero ...@@ -1332,9 +1337,9 @@ count the number of years since the year 1 B.C., and do not skip zero
as traditional Gregorian years do; for example, the year number as traditional Gregorian years do; for example, the year number
@minus{}37 represents the Gregorian year 38 B.C@. @minus{}37 represents the Gregorian year 38 B.C@.
@defun decode-time &optional time @defun decode-time &optional time-value
This function converts a time value into calendrical information. If This function converts a time value into calendrical information. If
you don't specify @var{time}, it decodes the current time. The return you don't specify @var{time-value}, it decodes the current time. The return
value is a list of nine elements, as follows: value is a list of nine elements, as follows:
@example @example
...@@ -1373,8 +1378,9 @@ Greenwich. ...@@ -1373,8 +1378,9 @@ Greenwich.
@defun encode-time seconds minutes hour day month year &optional zone @defun encode-time seconds minutes hour day month year &optional zone
This function is the inverse of @code{decode-time}. It converts seven This function is the inverse of @code{decode-time}. It converts seven
items of calendrical data into a time value. For the meanings of the items of calendrical data into a list-of-integer time value. For the
arguments, see the table above under @code{decode-time}. meanings of the arguments, see the table above under
@code{decode-time}.
Year numbers less than 100 are not treated specially. If you want them Year numbers less than 100 are not treated specially. If you want them
to stand for years above 1900, or years above 2000, you must alter them to stand for years above 1900, or years above 2000, you must alter them
...@@ -1418,9 +1424,11 @@ This function parses the time-string @var{string} and returns the ...@@ -1418,9 +1424,11 @@ This function parses the time-string @var{string} and returns the
corresponding time value. corresponding time value.
@end defun @end defun
@defun format-time-string format-string &optional time universal @defun format-time-string format-string &optional time-value universal
This function converts @var{time} (or the current time, if @var{time} is
omitted) to a string according to @var{format-string}. The argument This function converts @var{time-value} (or the current time, if
@var{time-value} is omitted) to a string according to
@var{format-string}. The argument
@var{format-string} may contain @samp{%}-sequences which say to @var{format-string} may contain @samp{%}-sequences which say to
substitute parts of the time. Here is a table of what the substitute parts of the time. Here is a table of what the
@samp{%}-sequences mean: @samp{%}-sequences mean:
...@@ -1540,12 +1548,6 @@ specified by @code{locale-coding-system} (@pxref{Locales}); after ...@@ -1540,12 +1548,6 @@ specified by @code{locale-coding-system} (@pxref{Locales}); after
system. system.
@end defun @end defun
@defun seconds-to-time seconds
This function converts @var{seconds}, the number of seconds since the
epoch, to a time value and returns that. To convert back, use
@code{float-time} (@pxref{Time of Day}).
@end defun
@defun format-seconds format-string seconds @defun format-seconds format-string seconds
This function converts its argument @var{seconds} into a string of This function converts its argument @var{seconds} into a string of
years, days, hours, etc., according to @var{format-string}. The years, days, hours, etc., according to @var{format-string}. The
...@@ -1619,7 +1621,7 @@ When called interactively, it prints the uptime in the echo area. ...@@ -1619,7 +1621,7 @@ When called interactively, it prints the uptime in the echo area.
@defun get-internal-run-time @defun get-internal-run-time
This function returns the processor run time used by Emacs as a list This function returns the processor run time used by Emacs as a list
of four integers: @code{(@var{high} @var{low} @var{microsec} of four integers: @code{(@var{sec-high} @var{sec-low} @var{microsec}
@var{picosec})}, using the same format as @code{current-time} @var{picosec})}, using the same format as @code{current-time}
(@pxref{Time of Day}). (@pxref{Time of Day}).
...@@ -1643,7 +1645,7 @@ interactively, it prints the duration in the echo area. ...@@ -1643,7 +1645,7 @@ interactively, it prints the duration in the echo area.
@section Time Calculations @section Time Calculations
These functions perform calendrical computations using time values These functions perform calendrical computations using time values
(the kind of list that @code{current-time} returns). (@pxref{Time of Day}).
@defun time-less-p t1 t2 @defun time-less-p t1 t2
This returns @code{t} if time value @var{t1} is less than time value This returns @code{t} if time value @var{t1} is less than time value
...@@ -1652,26 +1654,26 @@ This returns @code{t} if time value @var{t1} is less than time value ...@@ -1652,26 +1654,26 @@ This returns @code{t} if time value @var{t1} is less than time value
@defun time-subtract t1 t2 @defun time-subtract t1 t2
This returns the time difference @var{t1} @minus{} @var{t2} between This returns the time difference @var{t1} @minus{} @var{t2} between
two time values, in the same format as a time value. two time values, as a time value.
@end defun @end defun
@defun time-add t1 t2 @defun time-add t1 t2
This returns the sum of two time values, one of which ought to This returns the sum of two time values, as a time value.
represent a time difference rather than a point in time. One argument should represent a time difference rather than a point in time.
Here is how to add a number of seconds to a time value: Here is how to add a number of seconds to a time value:
@example @example
(time-add @var{time} (seconds-to-time @var{seconds})) (time-add @var{time} @var{seconds})
@end example @end example
@end defun @end defun
@defun time-to-days time @defun time-to-days time-value
This function returns the number of days between the beginning of year This function returns the number of days between the beginning of year
1 and @var{time}. 1 and @var{time-value}.
@end defun @end defun
@defun time-to-day-in-year time @defun time-to-day-in-year time-value
This returns the day number within the year corresponding to @var{time}. This returns the day number within the year corresponding to @var{time-value}.
@end defun @end defun
@defun date-leap-year-p year @defun date-leap-year-p year
...@@ -1915,8 +1917,7 @@ idleness. Here's an example: ...@@ -1915,8 +1917,7 @@ idleness. Here's an example:
(run-with-idle-timer (run-with-idle-timer
;; Compute an idle time @var{break-length} ;; Compute an idle time @var{break-length}
;; more than the current value. ;; more than the current value.
(time-add (current-idle-time) (time-add (current-idle-time) @var{break-length})
(seconds-to-time @var{break-length}))
nil nil
'my-timer-function)))) 'my-timer-function))))
@end example @end example
......
2014-11-17 Paul Eggert <eggert@cs.ucla.edu>
Improve time stamp handling, and be more consistent about it.
* NEWS: Document the changes.
2014-11-14 Lars Magne Ingebrigtsen <larsi@gnus.org> 2014-11-14 Lars Magne Ingebrigtsen <larsi@gnus.org>
* NEWS: Mention the new `M-s M-s' keystroke. * NEWS: Mention the new `M-s M-s' keystroke.
......
...@@ -377,6 +377,25 @@ optional repeat-count argument. ...@@ -377,6 +377,25 @@ optional repeat-count argument.
*** New macros `thread-first' and `thread-last' allow threading a form *** New macros `thread-first' and `thread-last' allow threading a form
as the first or last argument of subsequent forms. as the first or last argument of subsequent forms.
+++
** Time-related changes:
*** Time-related functions now consistently accept numbers
(representing seconds since the epoch) and nil (representing the
current time) as well as the usual list-of-integer representation.
Affected functions include `current-time-string', `current-time-zone',
`decode-time', `float-time', `format-time-string', `seconds-to-time',
`time-add', `time-less-p', `time-subtract', `time-to-day-in-year',
`time-to-days', and `time-to-seconds'.
*** The `encode-time-value' and `with-decoded-time-value' macros have
been obsoleted.
*** `calendar-next-time-zone-transition', `time-add', and
`time-subtract' no longer return time values in the obsolete and
undocumented integer-pair format. Instead, they return a list of two
integers.
* Changes in Frames and Windows Code in Emacs 25.1 * Changes in Frames and Windows Code in Emacs 25.1
......
2014-11-17 Paul Eggert <eggert@cs.ucla.edu>
Improve time stamp handling, and be more consistent about it.
This implements a suggestion made in:
http://lists.gnu.org/archive/html/emacs-devel/2014-10/msg00587.html
Among other things, this means timer.el no longer needs to
autoload the time-date module.
* allout-widgets.el (allout-elapsed-time-seconds): Doc fix.
* arc-mode.el (archive-ar-summarize):
* calendar/time-date.el (seconds-to-time, days-to-time, time-since):
* emacs-lisp/timer.el (timer-relative-time, timer-event-handler)
(run-at-time, with-timeout-suspend, with-timeout-unsuspend):
* net/tramp.el (tramp-time-less-p, tramp-time-subtract):
* proced.el (proced-time-lessp):
* timezone.el (timezone-time-from-absolute):
* type-break.el (type-break-schedule, type-break-time-sum):
Simplify by using new functionality.
* calendar/cal-dst.el (calendar-next-time-zone-transition):
Do not return time values in obsolete and undocumented (HI . LO)
format; use (HI LO) instead.
* calendar/time-date.el (with-decoded-time-value):
Treat 'nil' as current time. This is mostly for XEmacs.
(encode-time-value, with-decoded-time-value): Obsolete.
(time-add, time-subtract, time-less-p): Use no-op autoloads, for
XEmacs. Define only if XEmacs, as they're now C builtins in Emacs.
* ldefs-boot.el: Update to match new time-date.el
* proced.el: Do not require time-date.
2014-11-16 Lars Magne Ingebrigtsen <larsi@gnus.org> 2014-11-16 Lars Magne Ingebrigtsen <larsi@gnus.org>
   
* net/eww.el (eww-mode): Make the buffer read-only. * net/eww.el (eww-mode): Make the buffer read-only.
......
...@@ -2342,9 +2342,9 @@ We use a caching strategy, so the caller doesn't need to do so." ...@@ -2342,9 +2342,9 @@ We use a caching strategy, so the caller doesn't need to do so."
got))) got)))
;;;_ : Miscellaneous ;;;_ : Miscellaneous
;;;_ > allout-elapsed-time-seconds (triple) ;;;_ > allout-elapsed-time-seconds (time-value time-value)
(defun allout-elapsed-time-seconds (end start) (defun allout-elapsed-time-seconds (end start)
"Return seconds between `current-time' style time START/END triples." "Return seconds between START/END time values."
(let ((elapsed (time-subtract end start))) (let ((elapsed (time-subtract end start)))
(float-time elapsed))) (float-time elapsed)))
;;;_ > allout-frame-property (frame property) ;;;_ > allout-frame-property (frame property)
......
...@@ -2181,11 +2181,7 @@ This doesn't recover lost files, it just undoes changes in the buffer itself." ...@@ -2181,11 +2181,7 @@ This doesn't recover lost files, it just undoes changes in the buffer itself."
(size (string-to-number (match-string 6)))) (size (string-to-number (match-string 6))))
;; Move to the beginning of the data. ;; Move to the beginning of the data.
(goto-char (match-end 0)) (goto-char (match-end 0))
(setq time (setq time (format-time-string "%Y-%m-%d %H:%M" time))
(format-time-string
"%Y-%m-%d %H:%M"
(let ((high (truncate (/ time 65536))))
(list high (truncate (- time (* 65536.0 high)))))))
(setq extname (setq extname
(cond ((equal name "// ") (cond ((equal name "// ")
(propertize ".<ExtNamesTable>." 'face 'italic)) (propertize ".<ExtNamesTable>." 'face 'italic))
......
...@@ -179,6 +179,7 @@ Return nil if no such transition can be found." ...@@ -179,6 +179,7 @@ Return nil if no such transition can be found."
(if (eq (car (current-time-zone probe)) hi-utc-diff) (if (eq (car (current-time-zone probe)) hi-utc-diff)
(setq hi probe) (setq hi probe)
(setq lo probe))) (setq lo probe)))
(setcdr hi (list (cdr hi)))
hi)))) hi))))
(autoload 'calendar-persian-to-absolute "cal-persia") (autoload 'calendar-persian-to-absolute "cal-persia")
......
...@@ -30,10 +30,9 @@ ...@@ -30,10 +30,9 @@
;; value equal to HIGH * 2^16 + LOW + USEC * 10^-6 + PSEC * 10^-12 ;; value equal to HIGH * 2^16 + LOW + USEC * 10^-6 + PSEC * 10^-12
;; seconds, where missing components are treated as zero. HIGH can be ;; seconds, where missing components are treated as zero. HIGH can be
;; negative, either because the value is a time difference, or because ;; negative, either because the value is a time difference, or because
;; the machine supports negative time stamps that fall before the epoch. ;; it represents a time stamp before the epoch. Typically, there are
;; The macro `with-decoded-time-value' and the function ;; more time values than the underlying system time type supports,
;; `encode-time-value' make it easier to deal with these formats. ;; but the reverse can also be true.
;; See `time-subtract' for an example of how to use them.
;;; Code: ;;; Code:
...@@ -71,6 +70,7 @@ list (HIGH LOW MICRO PICO)." ...@@ -71,6 +70,7 @@ list (HIGH LOW MICRO PICO)."
,low ,micro) ,low ,micro)
(when pico `(,pico)) (when pico `(,pico))
(when type `(,type))) (when type `(,type)))
(or ,gensym (setq ,gensym (current-time)))
(if (consp ,gensym) (if (consp ,gensym)
(progn (progn
(setq ,low (pop ,gensym)) (setq ,low (pop ,gensym))
...@@ -108,6 +108,10 @@ it is assumed that PICO was omitted and should be treated as zero." ...@@ -108,6 +108,10 @@ it is assumed that PICO was omitted and should be treated as zero."
((eq type 3) (list high low micro pico)) ((eq type 3) (list high low micro pico))
((null type) (encode-time-value high low micro 0 pico)))) ((null type) (encode-time-value high low micro 0 pico))))
(when (featurep 'emacs)
(make-obsolete 'encode-time-value nil "25.1")
(make-obsolete 'with-decoded-time-value nil "25.1"))
(autoload 'parse-time-string "parse-time") (autoload 'parse-time-string "parse-time")
(autoload 'timezone-make-date-arpa-standard "timezone") (autoload 'timezone-make-date-arpa-standard "timezone")
...@@ -158,47 +162,17 @@ TIME defaults to the current time." ...@@ -158,47 +162,17 @@ TIME defaults to the current time."
;;;###autoload ;;;###autoload
(defun seconds-to-time (seconds) (defun seconds-to-time (seconds)
"Convert SECONDS (a floating point number) to a time value." "Convert SECONDS to a time value."
(let* ((usec (* 1000000 (mod seconds 1))) (time-add 0 seconds))
(ps (round (* 1000000 (mod usec 1))))
(us (floor usec))
(lo (floor (mod seconds 65536)))
(hi (floor seconds 65536)))
(if (eq ps 1000000)
(progn
(setq ps 0)
(setq us (1+ us))
(if (eq us 1000000)
(progn
(setq us 0)
(setq lo (1+ lo))
(if (eq lo 65536)
(progn
(setq lo 0)
(setq hi (1+ hi))))))))
(list hi lo us ps)))
;;;###autoload
(defun time-less-p (t1 t2)
"Return non-nil if time value T1 is earlier than time value T2."
(with-decoded-time-value ((high1 low1 micro1 pico1 type1 t1)
(high2 low2 micro2 pico2 type2 t2))
(or (< high1 high2)
(and (= high1 high2)
(or (< low1 low2)
(and (= low1 low2)
(or (< micro1 micro2)
(and (= micro1 micro2)
(< pico1 pico2)))))))))
;;;###autoload ;;;###autoload
(defun days-to-time (days) (defun days-to-time (days)
"Convert DAYS into a time value." "Convert DAYS into a time value."
(let* ((seconds (* 1.0 days 60 60 24)) (let ((time (condition-case nil (seconds-to-time (* 86400.0 days))
(high (condition-case nil (floor (/ seconds 65536)) (range-error (list most-positive-fixnum 65535)))))
(range-error most-positive-fixnum)))) (if (integerp days)
(list high (condition-case nil (floor (- seconds (* 1.0 high 65536))) (setcdr (cdr time) nil))
(range-error 65535))))) time))
;;;###autoload ;;;###autoload
(defun time-since (time) (defun time-since (time)
...@@ -207,53 +181,71 @@ TIME should be either a time value or a date-time string." ...@@ -207,53 +181,71 @@ TIME should be either a time value or a date-time string."
(when (stringp time) (when (stringp time)
;; Convert date strings to internal time. ;; Convert date strings to internal time.
(setq time (date-to-time time))) (setq time (date-to-time time)))
(time-subtract (current-time) time)) (time-subtract nil time))
;;;###autoload ;;;###autoload
(defalias 'subtract-time 'time-subtract) (defalias 'subtract-time 'time-subtract)
;;;###autoload ;; These autoloads do nothing in Emacs 25, where the functions are builtin.
(defun time-subtract (t1 t2) ;;;###autoload(autoload 'time-add "time-date")
"Subtract two time values, T1 minus T2. ;;;###autoload(autoload 'time-subtract "time-date")
Return the difference in the format of a time value." ;;;###autoload(autoload 'time-less-p "time-date")
(with-decoded-time-value ((high low micro pico type t1)
(high2 low2 micro2 pico2 type2 t2))
(setq high (- high high2)
low (- low low2)
micro (- micro micro2)
pico (- pico pico2)
type (max type type2))
(when (< pico 0)
(setq micro (1- micro)
pico (+ pico 1000000)))
(when (< micro 0)
(setq low (1- low)
micro (+ micro 1000000)))
(when (< low 0)
(setq high (1- high)