Philipp Stephani [Thu, 20 Apr 2017 13:41:15 +0000 (15:41 +0200)]
Add missing remappings for Ido mode
Among others, add a remapping for C-x 4 d, cf. Bug#26360.
* lisp/ido.el (ido-mode): Remap missing commands.
(ido-file-internal, ido-visit-buffer): Add support for new
methods.
(ido-display-buffer-other-frame)
(ido-find-alternate-file-other-window, ido-dired-other-window)
(ido-dired-other-frame): New commands.
* test/lisp/ido-tests.el (ido-tests--other-window-frame): Add unit
test for the bindings.
"Fix ert-tests when running compiled" 2016-12-06 accidentally
introduced a dependency on `bytecomp' into `ert'. As mentioned in
"Avoid ert test failures" 2017-04-18, the accidental dependency of ert
on bytecomp was masked by loading other libraries until recently.
* lisp/emacs-lisp/ert.el (ert--expand-should-1): Only use
`byte-compile-macro-environment' if it's bound.
* test/src/eval-tests.el: Add defvar for dynamic variable
`byte-compile-debug'.
Andrew G Cohen [Sun, 23 Apr 2017 00:25:02 +0000 (08:25 +0800)]
Allow limiting gnus summary buffers to a thread
* lisp/gnus/gnus-sum.el (gnus-summary-limit-include-thread): Include
an optional argument to allow limiting the summary buffer to just the
thread-related articles.
(gnus-refer-thread-limit-to-thread): Introduce customizable variable
to control whether thread-referral adds the thread to the summary
buffer or limits to just the thread.
(gnus-summary-refer-thread): Use the new variable.
Andrew G Cohen [Sun, 23 Apr 2017 00:13:09 +0000 (08:13 +0800)]
Correct gnus-newsgroup-limits in gnus when including thread
* lisp/gnus/gnus-sum.el (gnus-summary-limit-include-thread): Should
only add one list of thread-related articles to gnus-newsgroup-limits
rather than two.
Andrew G Cohen [Sun, 23 Apr 2017 00:07:22 +0000 (08:07 +0800)]
Improve gnus thread matching of similar subjects
* lisp/gnus/gnus-sum.el (gnus-summary-limit-include-thread):
Use the more liberal gnus-general-simplify-subject regexp to
find thread articles with similar subjects.
Noam Postavsky [Thu, 16 Mar 2017 02:27:27 +0000 (22:27 -0400)]
Add new `lisp-indent-region' that doesn't reparse the code.
Both `lisp-indent-region' and `lisp-indent-line' now use `syntax-ppss'
to get initial state, so they will no longer indent string literal
contents.
* lisp/emacs-lisp/lisp-mode.el (lisp-ppss): New function, like
`syntax-ppss', but with a more dependable item 2.
(lisp-indent-region): New function, like `indent-region-line-by-line'
but additionally keep a running parse state to avoid reparsing the
code repeatedly. Use `lisp-ppss' to get initial state.
(lisp-indent-line): Take optional PARSE-STATE argument, pass it to
`calculate-lisp-indent', use `lisp-ppss' if not given.
(lisp-mode-variables): Set `indent-region-function' to
`lisp-indent-region'.
Noam Postavsky [Thu, 16 Mar 2017 01:59:13 +0000 (21:59 -0400)]
Remove ignored argument from lisp-indent-line
* lisp/emacs-lisp/lisp-mode.el (lisp-indent-line): Remove WHOLE-EXP
argument, the behavior has long since been handled in
`indent-for-tab-command'. Also remove redundant `beg' and `shift-amt'
variables and use `indent-line-to'.
Noam Postavsky [Mon, 13 Mar 2017 03:59:19 +0000 (23:59 -0400)]
Don't reparse the sexp in indent-sexp (Bug#25122)
* lisp/emacs-lisp/lisp-mode.el (calculate-lisp-indent): Let
PARSE-START be a parse state that can be reused.
(indent-sexp): Pass the running parse state to calculate-lisp-indent
instead of the sexp beginning position. Saving the
CONTAINING-SEXP-START returned by `calculate-lisp-indent' is no longer
needed. Don't bother stopping if we don't descend below init-depth,
since we now alway scan the whole buffer (via syntax-ppss) anyway.
* test/lisp/emacs-lisp/lisp-mode-tests.el (indent-sexp): Add blank
line to test case.
b-c--cond-jump-table-info: Use correct body for singleton clauses
* lisp/emacs-lisp/bytecomp.el (byte-compile-cond-jump-table-info):
When a clause's body consists of a single constant expression, use
that expression as the body to be compiled. This fixes switch bytecode
evaluating to nil to such clauses.
When using ffap-other-window, don't change the window configuration
unless a new buffer has actually been created (Bug#25352).
* lisp/ffap.el (ffap-other-frame): Don't change the window
configuration if no new buffer has been created.
* test/lisp/ffap-tests.el (ffap-other-window--bug-25352): Add unit
test.
Alan Mackenzie [Sat, 22 Apr 2017 14:45:24 +0000 (14:45 +0000)]
Fix fontification of C++ declaration with type FOO::FOO.
* lisp/progmodes/cc-engine.el (c-find-decl-spots): Initialize
cfd-top-level properly.
(c-forward-decl-or-cast-1): On finding FOO::FOO, check it is followed by "("
before deciding it is a constructor.
* lisp/progmodes/cc-fonts.el (c-font-lock-complex-decl-prepare): Negate the
result of the c-bs-at-toplevel-p call passed to c-font-lock-declarators
(simple bug fix).
Alan Third [Thu, 20 Apr 2017 14:25:56 +0000 (15:25 +0100)]
Add no-accept-focus and frame-list-z-order to NS port
* lisp/frame.el (frame-list-z-order): Add NS.
* src/nsfns.m: Add x_set_no_accept_focus to handler struct.
(Fx_create_frame): Handle no-accept-focus parameter.
(ns_window_is_ancestor):
(Fns_frame_list_z_order): New functions.
* src/nsterm.m (x_set_no_accept_focus): New function.
(initFrameFromEmacs): Use EmacsWindow instead of EmacsFSWindow for
non-fullscreen windows.
(EmacsWindow:canBecomeKeyWindow): New function.
Stefan Monnier [Fri, 21 Apr 2017 16:12:42 +0000 (12:12 -0400)]
* lisp/emacs-lisp/cl-macs.el: Fix symbol-macrolet
Revert 0d112c00ba0ec14bd3014efcd3430b9ddcfe1fc1 (to fix bug#26325)
and use a different fix for bug#26068.
(cl--symbol-macro-key): New function.
(cl--sm-macroexpand, cl-symbol-macrolet): Use it instead of `symbol-name`.
* test/lisp/emacs-lisp/cl-lib-tests.el (cl-lib-symbol-macrolet):
Failure is not expected any more.
Eli Zaretskii [Fri, 21 Apr 2017 08:10:14 +0000 (11:10 +0300)]
Avoid infinite loop in redisplay when header-line-format is invalid
* src/xdisp.c (handle_invisible_prop): Avoid inflooping when the
string has an invalid %-construct in it and is displayed as part
of mode-line or header-line. (Bug#26586)
Treat non-erroring lisp call as successful eshell command (Bug#26161)
This lets a compound command like 'cd .. && echo ok' print 'ok',
similar to how most other shells behave.
* lisp/eshell/esh-cmd.el (eshell-exit-success-p): Only check if the
last exit code was zero, rather than first checking whether the last
command returned nil.
(eshell-exec-lisp): Set `eshell-last-command-status' to 1 on error.
Reuben Thomas [Thu, 2 Mar 2017 12:50:06 +0000 (12:50 +0000)]
Fix reading of tab settings in whitespace-mode
lisp/whitespace.el (whitespace-indent-tabs-mode)
whitespace-tab-width): Remove these variables. The underlying
variables `indent-tabs-mode' and `tab-width' are already buffer-local
when needed, and whitespace-mode never changes them.
(whitespace-ensure-local-variables): Remove this function, which only
existed to set the above variables.
(whitespace-cleanup-region, whitespace-regexp)
(whitespace-indentation-regexp, whitespace-report-region)
(whitespace-turn-on, whitespace-color-on): Adjust these functions to
use `indent-tabs-mode' and `tab-width' directly, and not call
`whitespace-ensure-local-variables'.
bytecomp: Don't inline functions that use byte-switch (Bug#26518)
* lisp/emacs-lisp/bytecomp.el (byte-compile-unfold-bcf): Don't inline
FORM if the bytecode uses the byte-switch instruction. It is
impossible to guess the correct stack depth while inlining such
bytecode, resulting in faulty code.
Stefan Monnier [Thu, 20 Apr 2017 00:43:41 +0000 (20:43 -0400)]
Use substring completion for Info menus and index
* lisp/info.el (Info-complete-menu-item): Add `category' metadata.
(Info-menu): Simplify now that we use the `default' arg of completing-read.
* lisp/minibuffer.el (completion-category-defaults): Use substring
completion for `info-menu`.
Pcase is macros, so these should have used eval-when-compile.
Anyway, pcase entry points are autoloaded, so the compiler handles it.
* lisp/profiler.el, lisp/emacs-lisp/eieio-core.el:
* lisp/emacs-lisp/generator.el, lisp/emacs-lisp/subr-x.el:
* lisp/progmodes/xref.el: No need to require pcase.
Alan Third [Fri, 14 Apr 2017 09:02:38 +0000 (10:02 +0100)]
Add new frame functionality to NS port
* lisp/frame.el (frame-restack): Call ns-frame-restack.
* src/keyboard.c (kbd_buffer_get_event) [HAVE_NS]: Enable
MOVE_FRAME_EVENT handling.
* src/frame.h:
* src/frame.c: Enable 'z-group', 'undecorated' and 'parent' frame
definitions.
* src/nsfns.m: Add x_set_z_group, x_set_parent_frame and
x_set_undecorated (Cocoa only) to handler struct.
(Fx_create_frame): Handle 'z-group', 'parent-frame' and 'undecorated'
frame parameter.
(Fns_frame_restack): New function.
* src/nsmenu.m (free_frame_tool_bar, update_frame_tool_bar):
FRAME_TOOLBAR_HEIGHT is no longer a variable.
* src/nsterm.h (NS_PARENT_WINDOW_LEFT_POS, NS_PARENT_WINDOW_TOP_POS):
Add #defines to find the screen position of the parent frame.
(NS_TOP_POS): Remove defun.
(EmacsView): Remove redundant toolbar variables and add createToolbar
method.
(FRAME_NS_TITLEBAR_HEIGHT, FRAME_TOOLBAR_HEIGHT): Always calculate the
values instead of storing them in a variable.
* src/nsterm.m (x_set_offset, windowDidMove): Take parent frame
position into account when positioning frames.
(initFrameFromEmacs): Remove toolbar creation code and handle new
frame parameters.
(x_set_window_size): Remove toolbar height calculation.
(x_set_z_group):
(x_set_parent_frame):
(x_set_undecorated) [NS_IMPL_COCOA]: New function.
(x_destroy_window): Detach parent if child closes.
(updateFrameSize): Change NSTRACE message to reflect new reality and
no longer reset frame size.
(windowWillResize): Don’t change NS window name when the titlebar
is invisible.
(createToolbar): Move toolbar creation code into it’s own method.
(toggleFullScreen): FRAME_TOOLBAR_HEIGHT and FRAME_NS_TITLEBAR_HEIGHT
are no longer variables.
(windowDidMove): Fire MOVE_FRAME_EVENT Emacs event.
* lisp/doc-view.el, lisp/filenotify.el, lisp/info-look.el:
* lisp/svg.el, lisp/emacs-lisp/byte-opt.el, lisp/net/shr.el:
* lisp/textmodes/sgml-mode.el, test/lisp/dom-tests.el:
No need to load subr-x at run-time.
* lisp/gnus/nnheader.el: No need to load subr-x.
; * lisp/emacs-lisp/subr-x.el, lisp/gnus/message.el, lisp/net/nsm.el:
; Comments.
* lisp/emacs-lisp/ert.el (ert--expand-should-1):
Avoid errors related to undefined byte-compile-macro-environment.
Somehow masked until very recently because loading seq (eg)
loads bytecomp. http://hydra.nixos.org/build/51730765
Noam Postavsky [Sun, 26 Mar 2017 02:41:34 +0000 (22:41 -0400)]
Fix find-library-name for load-history entries with nil FILE-NAME (Bug#26355)
* lisp/emacs-lisp/find-func.el (find-library--from-load-history):
Rename from find-library--from-load-path. Check for `load-history'
entries with nil FILE-NAMEs. Simplify by not double
checking for suffixes and making use of `locate-file'.
Paul Eggert [Mon, 17 Apr 2017 17:19:39 +0000 (10:19 -0700)]
Tighten recently-added UTF-8 check
* src/coding.c (encode_coding_utf_8): Now extern.
* src/terminal.c (terminal_glyph_code) [HAVE_STRUCT_UNIPAIR_UNICODE]:
Check for UTF-8, not just for multibyte.
David Engster [Mon, 17 Apr 2017 14:52:08 +0000 (16:52 +0200)]
xml: Properly handle symbol-qnames for attribute parsing
* lisp/xml.el (xml-parse-attlist): Do not strip 'symbol-qnames from
xml-ns argument (reverts aea67018) (Bug#26533).
(xml-maybe-do-ns): Properly handle default namespace by not
interning new symbol when 'special' flag is set.
* tests/lisp/xml-tests.el (xml-parse-test--namespace-attribute-qnames)
(xml-parse-namespace-attribute-qnames): Add test for Bug#26533.
* lisp/textmodes/css-mode.el (css--font-lock-keywords): Highlight
selectors where the part before a colon is only one character long,
such as `a:hover'.
Eli Zaretskii [Sun, 16 Apr 2017 07:43:22 +0000 (10:43 +0300)]
Fix redisplay performance problems with some fonts
* src/font.c (font_list_entities): Revert part of the changes
introduced on Apr 2, 2014 to fix bug#17125. It turns out having
zero_vector in the font-cache is an important indication that
cannot be removed. (Bug#21028)
Eli Zaretskii [Sun, 16 Apr 2017 07:27:23 +0000 (10:27 +0300)]
Add assertion to STRING_SET_CHARS
* src/lisp.h (STRING_SET_CHARS): Add an assertion and commentary
to prevent incorrect usage. For details, see this discussion:
http://lists.gnu.org/archive/html/emacs-devel/2017-04/msg00412.html.
Paul Eggert [Sun, 16 Apr 2017 05:22:49 +0000 (22:22 -0700)]
Merge from gnulib
This incorporates:
2017-04-14 intprops: try to avoid tickling similar bugs
2017-04-14 intprops: port to Oracle Studio 12.3 x86
* doc/misc/texinfo.tex, lib/intprops.h: Copy from gnulib.
Martin Rudalics [Sat, 15 Apr 2017 14:16:26 +0000 (16:16 +0200)]
Fix bugs in `with-displayed-buffer-window' and `fit-window-to-buffer'
* lisp/window.el (with-displayed-buffer-window): When a
'window-height' action alist entry specifies a function, call
`temp-buffer-window-show' with a '(window-height . t)' dummy
entry so `window--try-to-split-window' will bind
`window-combination-limit' to t and that function does not
resize any other window but the one we split this one off
(Bug#25055, Bug#25179).
(fit-window-to-buffer): Call `window-max-delta' with NOUP t so
we steal space only from windows in the same combination.
Stealing space from other windows would not allow us to return
that space later when this window is deleted (Bug#25055,
Bug#25179).
Create generated lisp files before main loaddefs.el
This should improve reproducibility of lisp/loaddefs.el.
* lisp/Makefile.in (gen-lisp): New phony target.
($(lisp)/loaddefs.el, compile-main): Depend on gen-lisp.
* src/Makefile.in ($(leimdir)/leim-list.el): Depend on all of ../leim.
* lisp/cedet/semantic.el (semantic-mode):
* lisp/cedet/semantic/fw.el (top-level):
* lisp/emacs-lisp/eieio-core.el (top-level):
Robustify to generated input files maybe not yet existing.
Martin Rudalics [Thu, 13 Apr 2017 15:45:12 +0000 (17:45 +0200)]
Describe recent frame and window changes in manuals
* doc/emacs/emacs.texi (Top):
* doc/emacs/cmdargs.texi (Borders X): Clearly separate the terms
"outer border" (for the X border which can be set from within
Emacs) and "external border" (for the border which is added by
the window manager).
* doc/lispref/display.texi (Tooltips): Clarify slightly.
* doc/lispref/elisp.texi (Top): Update node and section names.
* doc/lispref/frames.texi (Frames): Describe difference between
top-level and child frames.
(Frame Layout): Describe outer border. Add more details about
how Emacs obtains the outer size and position of a frame and
about menu bar/tool bar wrapping. Add references to new frame
parameters.
(Size and Position): Remove subsection.
(Frame Position): New subsection excerpted from the earlier Size
and Position subsection. Clarify positioning concepts and
some of their shortcomings. Describe `move-frame-functions'.
(Frame Size): New subsection excerpted from the earlier Size
and Position subsection. Describe how to track frame size
changes and the new function `frame-size-changed-p'.
(Position Parameters): Describe child frame positioning. Warn
about negative offsets. Describe 'z-group' parameter.
(Size Parameters): Describe 'text-pixels' specification
facility and new 'min-width' and 'min-height' parameters.
(Layout Parameters): Clarify description of 'tool-bar-lines' and
'menu-bar-lines' parameters.
(Frame Interaction Parameters): New subsubsection describing
'parent-frame', 'delete-before', 'mouse-wheel-frame' and
'no-other-frame' parameters.
(Management Parameters): Describe 'skip-taskbar',
'no-focus-on-map', 'no-accept-focus', 'undecorated' and
'override-redirect' parameters.
(Deleting Frames): Describe handling of 'delete-before'
parameter and child frames for `delete-frame' and
`delete-other-frames'.
(Finding All Frames): Describe `frame-list-z-order' and handling
of 'no-other-frame' parameter by `next-frame'.
(Minibuffers and Frames): Minor clarifications.
(Input Focus): Document `x-focus-frame'. Clarify descriptions
of `focus-in-hook', `focus-out-hook' and `focus-follows-mouse'.
(Visibility of Frames): Describe mapping and how the visibility
of a parent frame affects that of its child frames.
(Raising and Lowering): Describe restacking of frames and
z-groups.
(Child Frames): New section.
* doc/lispref/windows.texi (Selecting Windows): Describe
additional semantics of NORECORD argument of `select-window' and
how `buffer-list-update-hook' can emulate a "select window
hook".
(Mouse Window Auto-selection): New section.
Eli Zaretskii [Thu, 13 Apr 2017 07:03:09 +0000 (10:03 +0300)]
Avoid unnecessary regeneration of the entire loaddefs.el
* lisp/Makefile.in (autoloads .PHONY): Add commentary explaining
why $(lisp)/loaddefs.el is a dependency of '.PHONY'.
($(lisp)/loaddefs.el): Copy an existing loaddefs.el to
loaddefs.tmp before running 'batch-update-autoloads' on it, to
avoid slow regeneration of the full contents. (Bug#26459)
Use 'move-if-change' instead of 'mv', to avoid producing a new
Emacs binary when not necessary.
Martin Rudalics [Wed, 12 Apr 2017 16:22:44 +0000 (18:22 +0200)]
New internal-border face and args for select-window and x-focus-frame
Add `internal-border' face and handle it whenever clearing the
internal border. If NORECORD equals the symbol
'mark-for-redisplay', `select-window' will not record the window
but still mark it for redisplay. The new argument NOACTIVATE
for `x-focus-frame' tries to not activate FRAME when set.
* lisp/faces.el (internal-border): New face.
* lisp/mwheel.el (mwheel-scroll): Select window to scroll with
`mark-for-redisplay'.
* lisp/scroll-bar.el (scroll-bar-drag)
(scroll-bar-horizontal-drag, scroll-bar-scroll-down)
(scroll-bar-scroll-up, scroll-bar-toolkit-scroll)
(scroll-bar-toolkit-horizontal-scroll): Select window to scroll
with `mark-for-redisplay'.
* lisp/window.el (handle-select-window): When
`focus-follows-mouse' is not 'auto-raise' try to not activate
FRAME.
* src/dispextern.h (face_id): Add INTERNAL_BORDER_FACE_ID.
* src/frame.c (Fx_focus_frame): New argument NOACTIVATE.
* src/frame.h (x_focus_frame): Update extern declaration.
* src/gtkutil.c (xg_clear_under_internal_border): Remove
function.
(xg_frame_resized, xg_frame_set_char_size): Call
x_clear_under_internal_border.
(xg_tool_bar_callback): Adapt x_focus_frame call.
* src/gtkutil.h (xg_clear_under_internal_border): Remove
declaration.
* src/nsfns.m (x_focus_frame): Add argument NOACTIVATE.
* src/w32fns.c (x_clear_under_internal_border): Fill border
with internal-border background if specified.
* src/w32term.h (x_clear_under_internal_border): Add extern
declaration.
* src/w32term.c (x_after_update_window_line): Fill border
with internal-border background if specified.
(w32_set_vertical_scroll_bar, w32_set_horizontal_scroll_bar)
(x_scroll_bar_clear, w32_read_socket): Call
x_clear_under_internal_border.
(x_focus_frame): New argument NOACTIVATE.
* src/window.c (select_window): Mark WINDOW for redisplay when
NORECORD equals 'mark-for-redisplay'.
(Fselect_window): Update doc-string.
(syms_of_window): Define Qmark_for_redisplay.
* src/xdisp.c (clear_garbaged_frames, echo_area_display)
(redisplay_internal): Call x_clear_under_internal_border.
* src/xfaces.c (lookup_basic_face): Handle `window-divider'
and `internal-border' faces.
(realize_basic_faces): Realize `internal-border' face.
(syms_of_xfaces): Define Qinternal_border.
* src/xfns.c (x_set_internal_border_width): Remove call for
xg_clear_under_internal_border.
(x_focus_frame): New argument NOACTIVATE. When non-nil try to not
activate frame.
* src/xterm.c (x_fill_rectangle): No more static.
(x_clear_under_internal_border, x_after_update_window_line):
Fill border with internal-border background if specified.
(xt_horizontal_action_hook): Rewrite.
(handle_one_xevent): Call x_clear_under_internal_border.
* src/xterm.h (x_fill_rectangle): Add extern declaration.
Tak Kunihiro [Wed, 12 Apr 2017 13:29:35 +0000 (16:29 +0300)]
Scroll right and left using wheel-right and wheel-left.
These changes also make use of touchpad and trackpad (Bug#26347).
* doc/emacs/frames.texi (Mouse Commands): Document horizontal
scrolling using the mouse wheel.
* lisp/mwheel.el (mwheel-scroll): Respond to wheel-right and wheel-left.
(mwheel-tilt-scroll-p, mwheel-flip-direction)
(mwheel-scroll-left-function, mwheel-scroll-right-function): New
defcustoms.
(mouse-wheel-left-event, mouse-wheel-right-event): New variables,
events that calls wheel-left/right.
* etc/NEWS: Mention horizontal scrolling using the mouse wheel.
* lisp/emacs-lisp/package.el (package-import-keyring)
(package--check-signature-content, package-check-signature):
Use new variable package-gnupghome-dir to control which GnuPG
homedir to use.
* doc/emacs/package.texi: Mention package-gnupghome-dir.
* etc/NEWS: Mention package-gnupghome-dir.
Martin Rudalics [Tue, 11 Apr 2017 06:29:41 +0000 (08:29 +0200)]
Set x_gtk_use_window_move by default for fixing bug#25851 and bug#25943
This activates a change that was installed a few weeks ago but whose
ChangeLog was inadvertently dropped during its commit. The proper
ChangeLog is included below as part of the present commit.
* src/gtkutil.c (xg_set_geometry): When x_gtk_use_window_move
is set avoid calling x_gtk_parse_geometry (Bug#25851).
(x_wm_set_size_hint): When x_gtk_use_window_move is set, set
PPosition, USPosition and USSize flags if requested.
* src/xterm.c (x_set_offset): With GTK when
x_gtk_use_window_move is set, leave it entirely to
gtk_window_move to position the window and skip any
post-adjustments (Bug#25851 and Bug#25943).
(x_gtk_use_window_move): New variable.
Alan Mackenzie [Mon, 10 Apr 2017 21:01:38 +0000 (21:01 +0000)]
Fix a loop in C Mode caused by inadequate analysis of comments.
After M-;, and the insertion of the opening "/*", the CC Mode after-change
function got confused, since the new comment opener matched the end of a
subsequent comment, but moving back over that comment did not come back to the
starting point. Fix this.
* lisp/progmodes/cc-engine.el (c-end-of-macro): Add a limit parameter, wherer
point is left if no end-of-macro is found before it.
(c-forward-sws): Change the `safe-start' mechanism. Now `safe-start' is
non-nil except where we have an unclosed block comment at the end of a macro.
This enables us to populate the cache more fully, at the cost of some run
time.