xref: /vim-8.2.3635/runtime/doc/pattern.txt (revision 2286304c)
1*pattern.txt*   For Vim version 8.2.  Last change: 2021 Jul 16
2
3
4		  VIM REFERENCE MANUAL    by Bram Moolenaar
5
6
7Patterns and search commands				*pattern-searches*
8
9The very basics can be found in section |03.9| of the user manual.  A few more
10explanations are in chapter 27 |usr_27.txt|.
11
121. Search commands		|search-commands|
132. The definition of a pattern	|search-pattern|
143. Magic			|/magic|
154. Overview of pattern items	|pattern-overview|
165. Multi items			|pattern-multi-items|
176. Ordinary atoms		|pattern-atoms|
187. Ignoring case in a pattern	|/ignorecase|
198. Composing characters		|patterns-composing|
209. Compare with Perl patterns	|perl-patterns|
2110. Highlighting matches	|match-highlight|
2211. Fuzzy matching		|fuzzy-match|
23
24==============================================================================
251. Search commands				*search-commands*
26
27							*/*
28/{pattern}[/]<CR>	Search forward for the [count]'th occurrence of
29			{pattern} |exclusive|.
30
31/{pattern}/{offset}<CR>	Search forward for the [count]'th occurrence of
32			{pattern} and go |{offset}| lines up or down.
33			|linewise|.
34
35							*/<CR>*
36/<CR>			Search forward for the [count]'th occurrence of the
37			latest used pattern |last-pattern| with latest used
38			|{offset}|.
39
40//{offset}<CR>		Search forward for the [count]'th occurrence of the
41			latest used pattern |last-pattern| with new
42			|{offset}|.  If {offset} is empty no offset is used.
43
44							*?*
45?{pattern}[?]<CR>	Search backward for the [count]'th previous
46			occurrence of {pattern} |exclusive|.
47
48?{pattern}?{offset}<CR>	Search backward for the [count]'th previous
49			occurrence of {pattern} and go |{offset}| lines up or
50			down |linewise|.
51
52							*?<CR>*
53?<CR>			Search backward for the [count]'th occurrence of the
54			latest used pattern |last-pattern| with latest used
55			|{offset}|.
56
57??{offset}<CR>		Search backward for the [count]'th occurrence of the
58			latest used pattern |last-pattern| with new
59			|{offset}|.  If {offset} is empty no offset is used.
60
61							*n*
62n			Repeat the latest "/" or "?" [count] times.
63			If the cursor doesn't move the search is repeated with
64			count + 1.
65			|last-pattern|
66
67							*N*
68N			Repeat the latest "/" or "?" [count] times in
69			opposite direction. |last-pattern|
70
71							*star* *E348* *E349*
72*			Search forward for the [count]'th occurrence of the
73			word nearest to the cursor.  The word used for the
74			search is the first of:
75				1. the keyword under the cursor |'iskeyword'|
76				2. the first keyword after the cursor, in the
77				   current line
78				3. the non-blank word under the cursor
79				4. the first non-blank word after the cursor,
80				   in the current line
81			Only whole keywords are searched for, like with the
82			command "/\<keyword\>".  |exclusive|
83			'ignorecase' is used, 'smartcase' is not.
84
85							*#*
86#			Same as "*", but search backward.  The pound sign
87			(character 163) also works.  If the "#" key works as
88			backspace, try using "stty erase <BS>" before starting
89			Vim (<BS> is CTRL-H or a real backspace).
90
91							*gstar*
92g*			Like "*", but don't put "\<" and "\>" around the word.
93			This makes the search also find matches that are not a
94			whole word.
95
96							*g#*
97g#			Like "#", but don't put "\<" and "\>" around the word.
98			This makes the search also find matches that are not a
99			whole word.
100
101							*gd*
102gd			Goto local Declaration.  When the cursor is on a local
103			variable, this command will jump to its declaration.
104			First Vim searches for the start of the current
105			function, just like "[[".  If it is not found the
106			search stops in line 1.  If it is found, Vim goes back
107			until a blank line is found.  From this position Vim
108			searches for the keyword under the cursor, like with
109			"*", but lines that look like a comment are ignored
110			(see 'comments' option).
111			Note that this is not guaranteed to work, Vim does not
112			really check the syntax, it only searches for a match
113			with the keyword.  If included files also need to be
114			searched use the commands listed in |include-search|.
115			After this command |n| searches forward for the next
116			match (not backward).
117
118							*gD*
119gD			Goto global Declaration.  When the cursor is on a
120			global variable that is defined in the file, this
121			command will jump to its declaration.  This works just
122			like "gd", except that the search for the keyword
123			always starts in line 1.
124
125							*1gd*
1261gd			Like "gd", but ignore matches inside a {} block that
127			ends before the cursor position.
128
129							*1gD*
1301gD			Like "gD", but ignore matches inside a {} block that
131			ends before the cursor position.
132
133							*CTRL-C*
134CTRL-C			Interrupt current (search) command.  Use CTRL-Break on
135			MS-Windows |dos-CTRL-Break|.
136			In Normal mode, any pending command is aborted.
137
138							*:noh* *:nohlsearch*
139:noh[lsearch]		Stop the highlighting for the 'hlsearch' option.  It
140			is automatically turned back on when using a search
141			command, or setting the 'hlsearch' option.
142			This command doesn't work in an autocommand, because
143			the highlighting state is saved and restored when
144			executing autocommands |autocmd-searchpat|.
145			Same thing for when invoking a user function.
146
147While typing the search pattern the current match will be shown if the
148'incsearch' option is on.  Remember that you still have to finish the search
149command with <CR> to actually position the cursor at the displayed match.  Or
150use <Esc> to abandon the search.
151
152All matches for the last used search pattern will be highlighted if you set
153the 'hlsearch' option.  This can be suspended with the |:nohlsearch| command.
154
155When 'shortmess' does not include the "S" flag, Vim will automatically show an
156index, on which the cursor is. This can look like this: >
157
158  [1/5]		Cursor is on first of 5 matches.
159  [1/>99]	Cursor is on first of more than 99 matches.
160  [>99/>99]	Cursor is after 99 match of more than 99 matches.
161  [?/??]	Unknown how many matches exists, generating the
162		statistics was aborted because of search timeout.
163
164Note: the count does not take offset into account.
165
166When no match is found you get the error: *E486* Pattern not found
167Note that for the |:global| command this behaves like a normal message, for Vi
168compatibility.  For the |:s| command the "e" flag can be used to avoid the
169error message |:s_flags|.
170
171					*search-offset* *{offset}*
172These commands search for the specified pattern.  With "/" and "?" an
173additional offset may be given.  There are two types of offsets: line offsets
174and character offsets.
175
176The offset gives the cursor position relative to the found match:
177    [num]	[num] lines downwards, in column 1
178    +[num]	[num] lines downwards, in column 1
179    -[num]	[num] lines upwards, in column 1
180    e[+num]	[num] characters to the right of the end of the match
181    e[-num]	[num] characters to the left of the end of the match
182    s[+num]	[num] characters to the right of the start of the match
183    s[-num]	[num] characters to the left of the start of the match
184    b[+num]	[num] identical to s[+num] above (mnemonic: begin)
185    b[-num]	[num] identical to s[-num] above (mnemonic: begin)
186    ;{pattern}  perform another search, see |//;|
187
188If a '-' or '+' is given but [num] is omitted, a count of one will be used.
189When including an offset with 'e', the search becomes inclusive (the
190character the cursor lands on is included in operations).
191
192Examples:
193
194pattern			cursor position	~
195/test/+1		one line below "test", in column 1
196/test/e			on the last t of "test"
197/test/s+2		on the 's' of "test"
198/test/b-3		three characters before "test"
199
200If one of these commands is used after an operator, the characters between
201the cursor position before and after the search is affected.  However, if a
202line offset is given, the whole lines between the two cursor positions are
203affected.
204
205An example of how to search for matches with a pattern and change the match
206with another word: >
207	/foo<CR>	find "foo"
208	c//e<CR>	change until end of match
209	bar<Esc>	type replacement
210	//<CR>		go to start of next match
211	c//e<CR>	change until end of match
212	beep<Esc>	type another replacement
213			etc.
214<
215							*//;* *E386*
216A very special offset is ';' followed by another search command.  For example: >
217
218   /test 1/;/test
219   /test.*/+1;?ing?
220
221The first one first finds the next occurrence of "test 1", and then the first
222occurrence of "test" after that.
223
224This is like executing two search commands after each other, except that:
225- It can be used as a single motion command after an operator.
226- The direction for a following "n" or "N" command comes from the first
227  search command.
228- When an error occurs the cursor is not moved at all.
229
230							*last-pattern*
231The last used pattern and offset are remembered.  They can be used to repeat
232the search, possibly in another direction or with another count.  Note that
233two patterns are remembered: One for "normal" search commands and one for the
234substitute command ":s".  Each time an empty pattern is given, the previously
235used pattern is used.  However, if there is no previous search command, a
236previous substitute pattern is used, if possible.
237
238The 'magic' option sticks with the last used pattern.  If you change 'magic',
239this will not change how the last used pattern will be interpreted.
240The 'ignorecase' option does not do this.  When 'ignorecase' is changed, it
241will result in the pattern to match other text.
242
243All matches for the last used search pattern will be highlighted if you set
244the 'hlsearch' option.
245
246To clear the last used search pattern: >
247	:let @/ = ""
248This will not set the pattern to an empty string, because that would match
249everywhere.  The pattern is really cleared, like when starting Vim.
250
251The search usually skips matches that don't move the cursor.  Whether the next
252match is found at the next character or after the skipped match depends on the
253'c' flag in 'cpoptions'.  See |cpo-c|.
254	   with 'c' flag:   "/..." advances 1 to 3 characters
255	without 'c' flag:   "/..." advances 1 character
256The unpredictability with the 'c' flag is caused by starting the search in the
257first column, skipping matches until one is found past the cursor position.
258
259When searching backwards, searching starts at the start of the line, using the
260'c' flag in 'cpoptions' as described above.  Then the last match before the
261cursor position is used.
262
263In Vi the ":tag" command sets the last search pattern when the tag is searched
264for.  In Vim this is not done, the previous search pattern is still remembered,
265unless the 't' flag is present in 'cpoptions'.  The search pattern is always
266put in the search history.
267
268If the 'wrapscan' option is on (which is the default), searches wrap around
269the end of the buffer.  If 'wrapscan' is not set, the backward search stops
270at the beginning and the forward search stops at the end of the buffer.  If
271'wrapscan' is set and the pattern was not found the error message "pattern
272not found" is given, and the cursor will not be moved.  If 'wrapscan' is not
273set the message becomes "search hit BOTTOM without match" when searching
274forward, or "search hit TOP without match" when searching backward.  If
275wrapscan is set and the search wraps around the end of the file the message
276"search hit TOP, continuing at BOTTOM" or "search hit BOTTOM, continuing at
277TOP" is given when searching backwards or forwards respectively.  This can be
278switched off by setting the 's' flag in the 'shortmess' option.  The highlight
279method 'w' is used for this message (default: standout).
280
281							*search-range*
282You can limit the search command "/" to a certain range of lines by including
283\%>l items.  For example, to match the word "limit" below line 199 and above
284line 300: >
285	/\%>199l\%<300llimit
286Also see |/\%>l|.
287
288Another way is to use the ":substitute" command with the 'c' flag.  Example: >
289   :.,300s/Pattern//gc
290This command will search from the cursor position until line 300 for
291"Pattern".  At the match, you will be asked to type a character.  Type 'q' to
292stop at this match, type 'n' to find the next match.
293
294The "*", "#", "g*" and "g#" commands look for a word near the cursor in this
295order, the first one that is found is used:
296- The keyword currently under the cursor.
297- The first keyword to the right of the cursor, in the same line.
298- The WORD currently under the cursor.
299- The first WORD to the right of the cursor, in the same line.
300The keyword may only contain letters and characters in 'iskeyword'.
301The WORD may contain any non-blanks (<Tab>s and/or <Space>s).
302Note that if you type with ten fingers, the characters are easy to remember:
303the "#" is under your left hand middle finger (search to the left and up) and
304the "*" is under your right hand middle finger (search to the right and down).
305(this depends on your keyboard layout though).
306
307								*E956*
308In very rare cases a regular expression is used recursively.  This can happen
309when executing a pattern takes a long time and when checking for messages on
310channels a callback is invoked that also uses a pattern or an autocommand is
311triggered.  In most cases this should be fine, but if a pattern is in use when
312it's used again it fails.  Usually this means there is something wrong with
313the pattern.
314
315==============================================================================
3162. The definition of a pattern		*search-pattern* *pattern* *[pattern]*
317					*regular-expression* *regexp* *Pattern*
318					*E76* *E383* *E476*
319
320For starters, read chapter 27 of the user manual |usr_27.txt|.
321
322						*/bar* */\bar* */pattern*
3231. A pattern is one or more branches, separated by "\|".  It matches anything
324   that matches one of the branches.  Example: "foo\|beep" matches "foo" and
325   matches "beep".  If more than one branch matches, the first one is used.
326
327   pattern ::=	    branch
328		or  branch \| branch
329		or  branch \| branch \| branch
330		etc.
331
332						*/branch* */\&*
3332. A branch is one or more concats, separated by "\&".  It matches the last
334   concat, but only if all the preceding concats also match at the same
335   position.  Examples:
336	"foobeep\&..." matches "foo" in "foobeep".
337	".*Peter\&.*Bob" matches in a line containing both "Peter" and "Bob"
338
339   branch ::=	    concat
340		or  concat \& concat
341		or  concat \& concat \& concat
342		etc.
343
344						*/concat*
3453. A concat is one or more pieces, concatenated.  It matches a match for the
346   first piece, followed by a match for the second piece, etc.  Example:
347   "f[0-9]b", first matches "f", then a digit and then "b".
348
349   concat  ::=	    piece
350		or  piece piece
351		or  piece piece piece
352		etc.
353
354						*/piece*
3554. A piece is an atom, possibly followed by a multi, an indication of how many
356   times the atom can be matched.  Example: "a*" matches any sequence of "a"
357   characters: "", "a", "aa", etc.  See |/multi|.
358
359   piece   ::=	    atom
360		or  atom  multi
361
362						*/atom*
3635. An atom can be one of a long list of items.  Many atoms match one character
364   in the text.  It is often an ordinary character or a character class.
365   Parentheses can be used to make a pattern into an atom.  The "\z(\)"
366   construct is only for syntax highlighting.
367
368   atom    ::=	    ordinary-atom		|/ordinary-atom|
369		or  \( pattern \)		|/\(|
370		or  \%( pattern \)		|/\%(|
371		or  \z( pattern \)		|/\z(|
372
373
374				*/\%#=* *two-engines* *NFA*
375Vim includes two regexp engines:
3761. An old, backtracking engine that supports everything.
3772. A new, NFA engine that works much faster on some patterns, possibly slower
378   on some patterns.
379
380Vim will automatically select the right engine for you.  However, if you run
381into a problem or want to specifically select one engine or the other, you can
382prepend one of the following to the pattern:
383
384	\%#=0	Force automatic selection.  Only has an effect when
385	        'regexpengine' has been set to a non-zero value.
386	\%#=1	Force using the old engine.
387	\%#=2	Force using the NFA engine.
388
389You can also use the 'regexpengine' option to change the default.
390
391			 *E864* *E868* *E874* *E875* *E876* *E877* *E878*
392If selecting the NFA engine and it runs into something that is not implemented
393the pattern will not match.  This is only useful when debugging Vim.
394
395==============================================================================
3963. Magic							*/magic*
397
398Some characters in the pattern, such as letters, are taken literally.  They
399match exactly the same character in the text.  When preceded with a backslash
400however, these characters may get a special meaning.  For example, "a" matches
401the letter "a", while "\a" matches any alphabetic character.
402
403Other characters have a special meaning without a backslash.  They need to be
404preceded with a backslash to match literally.  For example "." matches any
405character while "\." matches a dot.
406
407If a character is taken literally or not depends on the 'magic' option and the
408items in the pattern mentioned next.  The 'magic' option should always be set,
409but it can be switched off for Vi compatibility.  We mention the effect of
410'nomagic' here for completeness, but we recommend against using that.
411							*/\m* */\M*
412Use of "\m" makes the pattern after it be interpreted as if 'magic' is set,
413ignoring the actual value of the 'magic' option.
414Use of "\M" makes the pattern after it be interpreted as if 'nomagic' is used.
415							*/\v* */\V*
416Use of "\v" means that after it, all ASCII characters except '0'-'9', 'a'-'z',
417'A'-'Z' and '_' have special meaning: "very magic"
418
419Use of "\V" means that after it, only a backslash and the terminating
420character (usually / or ?) have special meaning: "very nomagic"
421
422Examples:
423after:	  \v	   \m	    \M	     \V		matches ~
424		'magic' 'nomagic'
425	  a	   a	    a	     a		literal 'a'
426	  \a	   \a	    \a	     \a		any alphabetic character
427	  .	   .	    \.	     \.		any character
428	  \.	   \.	    .	     .		literal dot
429	  $	   $	    $	     \$		end-of-line
430	  *	   *	    \*	     \*		any number of the previous atom
431	  ~	   ~	    \~	     \~		latest substitute string
432	  ()	   \(\)     \(\)     \(\)	group as an atom
433	  |	   \|	    \|	     \|		nothing: separates alternatives
434	  \\	   \\	    \\	     \\		literal backslash
435	  \{	   {	    {	     {		literal curly brace
436
437{only Vim supports \m, \M, \v and \V}
438
439If you want to you can make a pattern immune to the 'magic' option being set
440or not by putting "\m" or "\M" at the start of the pattern.
441
442==============================================================================
4434. Overview of pattern items				*pattern-overview*
444						*E865* *E866* *E867* *E869*
445
446Overview of multi items.				*/multi* *E61* *E62*
447More explanation and examples below, follow the links.		*E64* *E871*
448
449	  multi ~
450     'magic' 'nomagic'	matches of the preceding atom ~
451|/star|	*	\*	0 or more	as many as possible
452|/\+|	\+	\+	1 or more	as many as possible
453|/\=|	\=	\=	0 or 1		as many as possible
454|/\?|	\?	\?	0 or 1		as many as possible
455
456|/\{|	\{n,m}	\{n,m}	n to m		as many as possible
457	\{n}	\{n}	n		exactly
458	\{n,}	\{n,}	at least n	as many as possible
459	\{,m}	\{,m}	0 to m		as many as possible
460	\{}	\{}	0 or more	as many as possible (same as *)
461
462|/\{-|	\{-n,m}	\{-n,m}	n to m		as few as possible
463	\{-n}	\{-n}	n		exactly
464	\{-n,}	\{-n,}	at least n	as few as possible
465	\{-,m}	\{-,m}	0 to m		as few as possible
466	\{-}	\{-}	0 or more	as few as possible
467
468							*E59*
469|/\@>|	\@>	\@>	1, like matching a whole pattern
470|/\@=|	\@=	\@=	nothing, requires a match |/zero-width|
471|/\@!|	\@!	\@!	nothing, requires NO match |/zero-width|
472|/\@<=|	\@<=	\@<=	nothing, requires a match behind |/zero-width|
473|/\@<!|	\@<!	\@<!	nothing, requires NO match behind |/zero-width|
474
475
476Overview of ordinary atoms.				*/ordinary-atom*
477More explanation and examples below, follow the links.
478
479      ordinary atom ~
480      magic   nomagic	matches ~
481|/^|	^	^	start-of-line (at start of pattern) |/zero-width|
482|/\^|	\^	\^	literal '^'
483|/\_^|	\_^	\_^	start-of-line (used anywhere) |/zero-width|
484|/$|	$	$	end-of-line (at end of pattern) |/zero-width|
485|/\$|	\$	\$	literal '$'
486|/\_$|	\_$	\_$	end-of-line (used anywhere) |/zero-width|
487|/.|	.	\.	any single character (not an end-of-line)
488|/\_.|	\_.	\_.	any single character or end-of-line
489|/\<|	\<	\<	beginning of a word |/zero-width|
490|/\>|	\>	\>	end of a word |/zero-width|
491|/\zs|	\zs	\zs	anything, sets start of match
492|/\ze|	\ze	\ze	anything, sets end of match
493|/\%^|	\%^	\%^	beginning of file |/zero-width|		*E71*
494|/\%$|	\%$	\%$	end of file |/zero-width|
495|/\%V|	\%V	\%V	inside Visual area |/zero-width|
496|/\%#|	\%#	\%#	cursor position |/zero-width|
497|/\%'m|	\%'m	\%'m	mark m position |/zero-width|
498|/\%l|	\%23l	\%23l	in line 23 |/zero-width|
499|/\%c|	\%23c	\%23c	in column 23 |/zero-width|
500|/\%v|	\%23v	\%23v	in virtual column 23 |/zero-width|
501
502Character classes:					*/character-classes*
503      magic   nomagic	matches ~
504|/\i|	\i	\i	identifier character (see 'isident' option)
505|/\I|	\I	\I	like "\i", but excluding digits
506|/\k|	\k	\k	keyword character (see 'iskeyword' option)
507|/\K|	\K	\K	like "\k", but excluding digits
508|/\f|	\f	\f	file name character (see 'isfname' option)
509|/\F|	\F	\F	like "\f", but excluding digits
510|/\p|	\p	\p	printable character (see 'isprint' option)
511|/\P|	\P	\P	like "\p", but excluding digits
512|/\s|	\s	\s	whitespace character: <Space> and <Tab>
513|/\S|	\S	\S	non-whitespace character; opposite of \s
514|/\d|	\d	\d	digit:				[0-9]
515|/\D|	\D	\D	non-digit:			[^0-9]
516|/\x|	\x	\x	hex digit:			[0-9A-Fa-f]
517|/\X|	\X	\X	non-hex digit:			[^0-9A-Fa-f]
518|/\o|	\o	\o	octal digit:			[0-7]
519|/\O|	\O	\O	non-octal digit:		[^0-7]
520|/\w|	\w	\w	word character:			[0-9A-Za-z_]
521|/\W|	\W	\W	non-word character:		[^0-9A-Za-z_]
522|/\h|	\h	\h	head of word character:		[A-Za-z_]
523|/\H|	\H	\H	non-head of word character:	[^A-Za-z_]
524|/\a|	\a	\a	alphabetic character:		[A-Za-z]
525|/\A|	\A	\A	non-alphabetic character:	[^A-Za-z]
526|/\l|	\l	\l	lowercase character:		[a-z]
527|/\L|	\L	\L	non-lowercase character:	[^a-z]
528|/\u|	\u	\u	uppercase character:		[A-Z]
529|/\U|	\U	\U	non-uppercase character		[^A-Z]
530|/\_|	\_x	\_x	where x is any of the characters above: character
531			class with end-of-line included
532(end of character classes)
533
534      magic   nomagic	matches ~
535|/\e|	\e	\e	<Esc>
536|/\t|	\t	\t	<Tab>
537|/\r|	\r	\r	<CR>
538|/\b|	\b	\b	<BS>
539|/\n|	\n	\n	end-of-line
540|/~|	~	\~	last given substitute string
541|/\1|	\1	\1	same string as matched by first \(\)
542|/\2|	\2	\2	Like "\1", but uses second \(\)
543	   ...
544|/\9|	\9	\9	Like "\1", but uses ninth \(\)
545								*E68*
546|/\z1|	\z1	\z1	only for syntax highlighting, see |:syn-ext-match|
547	   ...
548|/\z1|	\z9	\z9	only for syntax highlighting, see |:syn-ext-match|
549
550	x	x	a character with no special meaning matches itself
551
552|/[]|	[]	\[]	any character specified inside the []
553|/\%[]|	\%[]	\%[]	a sequence of optionally matched atoms
554
555|/\c|	\c	\c	ignore case, do not use the 'ignorecase' option
556|/\C|	\C	\C	match case, do not use the 'ignorecase' option
557|/\Z|	\Z	\Z	ignore differences in Unicode "combining characters".
558			Useful when searching voweled Hebrew or Arabic text.
559
560      magic   nomagic	matches ~
561|/\m|	\m	\m	'magic' on for the following chars in the pattern
562|/\M|	\M	\M	'magic' off for the following chars in the pattern
563|/\v|	\v	\v	the following chars in the pattern are "very magic"
564|/\V|	\V	\V	the following chars in the pattern are "very nomagic"
565|/\%#=|   \%#=1   \%#=1   select regexp engine |/zero-width|
566
567|/\%d|	\%d	\%d	match specified decimal character (eg \%d123)
568|/\%x|	\%x	\%x	match specified hex character (eg \%x2a)
569|/\%o|	\%o	\%o	match specified octal character (eg \%o040)
570|/\%u|	\%u	\%u	match specified multibyte character (eg \%u20ac)
571|/\%U|	\%U	\%U	match specified large multibyte character (eg
572			\%U12345678)
573|/\%C|	\%C	\%C	match any composing characters
574
575Example			matches ~
576\<\I\i*		or
577\<\h\w*
578\<[a-zA-Z_][a-zA-Z0-9_]*
579			An identifier (e.g., in a C program).
580
581\(\.$\|\. \)		A period followed by <EOL> or a space.
582
583[.!?][])"']*\($\|[ ]\)	A search pattern that finds the end of a sentence,
584			with almost the same definition as the ")" command.
585
586cat\Z			Both "cat" and "càt" ("a" followed by 0x0300)
587			Does not match "càt" (character 0x00e0), even
588			though it may look the same.
589
590
591==============================================================================
5925. Multi items						*pattern-multi-items*
593
594An atom can be followed by an indication of how many times the atom can be
595matched and in what way.  This is called a multi.  See |/multi| for an
596overview.
597
598							*/star* */\star*
599*	(use \* when 'magic' is not set)
600	Matches 0 or more of the preceding atom, as many as possible.
601	Example  'nomagic'	matches ~
602	a*	   a\*		"", "a", "aa", "aaa", etc.
603	.*	   \.\*		anything, also an empty string, no end-of-line
604	\_.*	   \_.\*	everything up to the end of the buffer
605	\_.*END	   \_.\*END	everything up to and including the last "END"
606				in the buffer
607
608	Exception: When "*" is used at the start of the pattern or just after
609	"^" it matches the star character.
610
611	Be aware that repeating "\_." can match a lot of text and take a long
612	time.  For example, "\_.*END" matches all text from the current
613	position to the last occurrence of "END" in the file.  Since the "*"
614	will match as many as possible, this first skips over all lines until
615	the end of the file and then tries matching "END", backing up one
616	character at a time.
617
618							*/\+*
619\+	Matches 1 or more of the preceding atom, as many as possible.
620	Example		matches ~
621	^.\+$		any non-empty line
622	\s\+		white space of at least one character
623
624							*/\=*
625\=	Matches 0 or 1 of the preceding atom, as many as possible.
626	Example		matches ~
627	foo\=		"fo" and "foo"
628
629							*/\?*
630\?	Just like \=.  Cannot be used when searching backwards with the "?"
631	command.
632
633					*/\{* *E60* *E554* *E870*
634\{n,m}	Matches n to m of the preceding atom, as many as possible
635\{n}	Matches n of the preceding atom
636\{n,}	Matches at least n of the preceding atom, as many as possible
637\{,m}	Matches 0 to m of the preceding atom, as many as possible
638\{}	Matches 0 or more of the preceding atom, as many as possible (like *)
639							*/\{-*
640\{-n,m}	matches n to m of the preceding atom, as few as possible
641\{-n}	matches n of the preceding atom
642\{-n,}	matches at least n of the preceding atom, as few as possible
643\{-,m}	matches 0 to m of the preceding atom, as few as possible
644\{-}	matches 0 or more of the preceding atom, as few as possible
645
646	n and m are positive decimal numbers or zero
647								*non-greedy*
648	If a "-" appears immediately after the "{", then a shortest match
649	first algorithm is used (see example below).  In particular, "\{-}" is
650	the same as "*" but uses the shortest match first algorithm.  BUT: A
651	match that starts earlier is preferred over a shorter match: "a\{-}b"
652	matches "aaab" in "xaaab".
653
654	Example			matches ~
655	ab\{2,3}c		"abbc" or "abbbc"
656	a\{5}			"aaaaa"
657	ab\{2,}c		"abbc", "abbbc", "abbbbc", etc.
658	ab\{,3}c		"ac", "abc", "abbc" or "abbbc"
659	a[bc]\{3}d		"abbbd", "abbcd", "acbcd", "acccd", etc.
660	a\(bc\)\{1,2}d		"abcd" or "abcbcd"
661	a[bc]\{-}[cd]		"abc" in "abcd"
662	a[bc]*[cd]		"abcd" in "abcd"
663
664	The } may optionally be preceded with a backslash: \{n,m\}.
665
666							*/\@=*
667\@=	Matches the preceding atom with zero width.
668	Like "(?=pattern)" in Perl.
669	Example			matches ~
670	foo\(bar\)\@=		"foo" in "foobar"
671	foo\(bar\)\@=foo	nothing
672							*/zero-width*
673	When using "\@=" (or "^", "$", "\<", "\>") no characters are included
674	in the match.  These items are only used to check if a match can be
675	made.  This can be tricky, because a match with following items will
676	be done in the same position.  The last example above will not match
677	"foobarfoo", because it tries match "foo" in the same position where
678	"bar" matched.
679
680	Note that using "\&" works the same as using "\@=": "foo\&.." is the
681	same as "\(foo\)\@=..".  But using "\&" is easier, you don't need the
682	parentheses.
683
684
685							*/\@!*
686\@!	Matches with zero width if the preceding atom does NOT match at the
687	current position. |/zero-width|
688	Like "(?!pattern)" in Perl.
689	Example			matches ~
690	foo\(bar\)\@!		any "foo" not followed by "bar"
691	a.\{-}p\@!		"a", "ap", "app", "appp", etc. not immediately
692				followed by a "p"
693	if \(\(then\)\@!.\)*$	"if " not followed by "then"
694
695	Using "\@!" is tricky, because there are many places where a pattern
696	does not match.  "a.*p\@!" will match from an "a" to the end of the
697	line, because ".*" can match all characters in the line and the "p"
698	doesn't match at the end of the line.  "a.\{-}p\@!" will match any
699	"a", "ap", "app", etc. that isn't followed by a "p", because the "."
700	can match a "p" and "p\@!" doesn't match after that.
701
702	You can't use "\@!" to look for a non-match before the matching
703	position: "\(foo\)\@!bar" will match "bar" in "foobar", because at the
704	position where "bar" matches, "foo" does not match.  To avoid matching
705	"foobar" you could use "\(foo\)\@!...bar", but that doesn't match a
706	bar at the start of a line.  Use "\(foo\)\@<!bar".
707
708	Useful example: to find "foo" in a line that does not contain "bar": >
709		/^\%(.*bar\)\@!.*\zsfoo
710<	This pattern first checks that there is not a single position in the
711	line where "bar" matches.  If ".*bar" matches somewhere the \@! will
712	reject the pattern.  When there is no match any "foo" will be found.
713	The "\zs" is to have the match start just before "foo".
714
715							*/\@<=*
716\@<=	Matches with zero width if the preceding atom matches just before what
717	follows. |/zero-width|
718	Like "(?<=pattern)" in Perl, but Vim allows non-fixed-width patterns.
719	Example			matches ~
720	\(an\_s\+\)\@<=file	"file" after "an" and white space or an
721				end-of-line
722	For speed it's often much better to avoid this multi.  Try using "\zs"
723	instead |/\zs|.  To match the same as the above example:
724		an\_s\+\zsfile
725	At least set a limit for the look-behind, see below.
726
727	"\@<=" and "\@<!" check for matches just before what follows.
728	Theoretically these matches could start anywhere before this position.
729	But to limit the time needed, only the line where what follows matches
730	is searched, and one line before that (if there is one).  This should
731	be sufficient to match most things and not be too slow.
732
733	In the old regexp engine the part of the pattern after "\@<=" and
734	"\@<!" are checked for a match first, thus things like "\1" don't work
735	to reference \(\) inside the preceding atom.  It does work the other
736	way around:
737	Bad example			matches ~
738	\%#=1\1\@<=,\([a-z]\+\)		",abc" in "abc,abc"
739
740	However, the new regexp engine works differently, it is better to not
741	rely on this behavior, do not use \@<= if it can be avoided:
742	Example				matches ~
743	\([a-z]\+\)\zs,\1		",abc" in "abc,abc"
744
745\@123<=
746	Like "\@<=" but only look back 123 bytes. This avoids trying lots
747	of matches that are known to fail and make executing the pattern very
748	slow.  Example, check if there is a "<" just before "span":
749		/<\@1<=span
750	This will try matching "<" only one byte before "span", which is the
751	only place that works anyway.
752	After crossing a line boundary, the limit is relative to the end of
753	the line.  Thus the characters at the start of the line with the match
754	are not counted (this is just to keep it simple).
755	The number zero is the same as no limit.
756
757							*/\@<!*
758\@<!	Matches with zero width if the preceding atom does NOT match just
759	before what follows.  Thus this matches if there is no position in the
760	current or previous line where the atom matches such that it ends just
761	before what follows.  |/zero-width|
762	Like "(?<!pattern)" in Perl, but Vim allows non-fixed-width patterns.
763	The match with the preceding atom is made to end just before the match
764	with what follows, thus an atom that ends in ".*" will work.
765	Warning: This can be slow (because many positions need to be checked
766	for a match).  Use a limit if you can, see below.
767	Example			matches ~
768	\(foo\)\@<!bar		any "bar" that's not in "foobar"
769	\(\/\/.*\)\@<!in	"in" which is not after "//"
770
771\@123<!
772	Like "\@<!" but only look back 123 bytes. This avoids trying lots of
773	matches that are known to fail and make executing the pattern very
774	slow.
775
776							*/\@>*
777\@>	Matches the preceding atom like matching a whole pattern.
778	Like "(?>pattern)" in Perl.
779	Example		matches ~
780	\(a*\)\@>a	nothing (the "a*" takes all the "a"'s, there can't be
781			another one following)
782
783	This matches the preceding atom as if it was a pattern by itself.  If
784	it doesn't match, there is no retry with shorter sub-matches or
785	anything.  Observe this difference: "a*b" and "a*ab" both match
786	"aaab", but in the second case the "a*" matches only the first two
787	"a"s.  "\(a*\)\@>ab" will not match "aaab", because the "a*" matches
788	the "aaa" (as many "a"s as possible), thus the "ab" can't match.
789
790
791==============================================================================
7926.  Ordinary atoms					*pattern-atoms*
793
794An ordinary atom can be:
795
796							*/^*
797^	At beginning of pattern or after "\|", "\(", "\%(" or "\n": matches
798	start-of-line; at other positions, matches literal '^'. |/zero-width|
799	Example		matches ~
800	^beep(		the start of the C function "beep" (probably).
801
802							*/\^*
803\^	Matches literal '^'.  Can be used at any position in the pattern, but
804	not inside [].
805
806							*/\_^*
807\_^	Matches start-of-line. |/zero-width|  Can be used at any position in
808	the pattern, but not inside [].
809	Example		matches ~
810	\_s*\_^foo	white space and blank lines and then "foo" at
811			start-of-line
812
813							*/$*
814$	At end of pattern or in front of "\|", "\)" or "\n" ('magic' on):
815	matches end-of-line <EOL>; at other positions, matches literal '$'.
816	|/zero-width|
817
818							*/\$*
819\$	Matches literal '$'.  Can be used at any position in the pattern, but
820	not inside [].
821
822							*/\_$*
823\_$	Matches end-of-line. |/zero-width|  Can be used at any position in the
824	pattern, but not inside [].  Note that "a\_$b" never matches, since
825	"b" cannot match an end-of-line.  Use "a\nb" instead |/\n|.
826	Example		matches ~
827	foo\_$\_s*	"foo" at end-of-line and following white space and
828			blank lines
829
830.	(with 'nomagic': \.)				*/.* */\.*
831	Matches any single character, but not an end-of-line.
832
833							*/\_.*
834\_.	Matches any single character or end-of-line.
835	Careful: "\_.*" matches all text to the end of the buffer!
836
837							*/\<*
838\<	Matches the beginning of a word: The next char is the first char of a
839	word.  The 'iskeyword' option specifies what is a word character.
840	|/zero-width|
841
842							*/\>*
843\>	Matches the end of a word: The previous char is the last char of a
844	word.  The 'iskeyword' option specifies what is a word character.
845	|/zero-width|
846
847							*/\zs*
848\zs	Matches at any position, but not inside [], and sets the start of the
849	match there: The next char is the first char of the whole match.
850	|/zero-width|
851	Example: >
852		/^\s*\zsif
853<	matches an "if" at the start of a line, ignoring white space.
854	Can be used multiple times, the last one encountered in a matching
855	branch is used.  Example: >
856		/\(.\{-}\zsFab\)\{3}
857<	Finds the third occurrence of "Fab".
858	This cannot be followed by a multi. *E888*
859	{not available when compiled without the |+syntax| feature}
860							*/\ze*
861\ze	Matches at any position, but not inside [], and sets the end of the
862	match there: The previous char is the last char of the whole match.
863	|/zero-width|
864	Can be used multiple times, the last one encountered in a matching
865	branch is used.
866	Example: "end\ze\(if\|for\)" matches the "end" in "endif" and
867	"endfor".
868	This cannot be followed by a multi. |E888|
869	{not available when compiled without the |+syntax| feature}
870
871						*/\%^* *start-of-file*
872\%^	Matches start of the file.  When matching with a string, matches the
873	start of the string.
874	For example, to find the first "VIM" in a file: >
875		/\%^\_.\{-}\zsVIM
876<
877						*/\%$* *end-of-file*
878\%$	Matches end of the file.  When matching with a string, matches the
879	end of the string.
880	Note that this does NOT find the last "VIM" in a file: >
881		/VIM\_.\{-}\%$
882<	It will find the next VIM, because the part after it will always
883	match.  This one will find the last "VIM" in the file: >
884		/VIM\ze\(\(VIM\)\@!\_.\)*\%$
885<	This uses |/\@!| to ascertain that "VIM" does NOT match in any
886	position after the first "VIM".
887	Searching from the end of the file backwards is easier!
888
889						*/\%V*
890\%V	Match inside the Visual area.  When Visual mode has already been
891	stopped match in the area that |gv| would reselect.
892	This is a |/zero-width| match.  To make sure the whole pattern is
893	inside the Visual area put it at the start and just before the end of
894	the pattern, e.g.: >
895		/\%Vfoo.*ba\%Vr
896<	This also works if only "foo bar" was Visually selected. This: >
897		/\%Vfoo.*bar\%V
898<	would match "foo bar" if the Visual selection continues after the "r".
899	Only works for the current buffer.
900
901						*/\%#* *cursor-position*
902\%#	Matches with the cursor position.  Only works when matching in a
903	buffer displayed in a window.
904	WARNING: When the cursor is moved after the pattern was used, the
905	result becomes invalid.  Vim doesn't automatically update the matches.
906	This is especially relevant for syntax highlighting and 'hlsearch'.
907	In other words: When the cursor moves the display isn't updated for
908	this change.  An update is done for lines which are changed (the whole
909	line is updated) or when using the |CTRL-L| command (the whole screen
910	is updated).  Example, to highlight the word under the cursor: >
911		/\k*\%#\k*
912<	When 'hlsearch' is set and you move the cursor around and make changes
913	this will clearly show when the match is updated or not.
914
915						*/\%'m* */\%<'m* */\%>'m*
916\%'m	Matches with the position of mark m.
917\%<'m	Matches before the position of mark m.
918\%>'m	Matches after the position of mark m.
919	Example, to highlight the text from mark 's to 'e: >
920		/.\%>'s.*\%<'e..
921<	Note that two dots are required to include mark 'e in the match.  That
922	is because "\%<'e" matches at the character before the 'e mark, and
923	since it's a |/zero-width| match it doesn't include that character.
924	WARNING: When the mark is moved after the pattern was used, the result
925	becomes invalid.  Vim doesn't automatically update the matches.
926	Similar to moving the cursor for "\%#" |/\%#|.
927
928						*/\%l* */\%>l* */\%<l* *E951*
929\%23l	Matches in a specific line.
930\%<23l	Matches above a specific line (lower line number).
931\%>23l	Matches below a specific line (higher line number).
932\%.l    Matches at the cursor line.
933\%<.l   Matches above the cursor line.
934\%>.l   Matches below the cursor line.
935	These six can be used to match specific lines in a buffer.  The "23"
936	can be any line number.  The first line is 1.
937	WARNING: When inserting or deleting lines Vim does not automatically
938	update the matches.  This means Syntax highlighting quickly becomes
939	wrong.  Also when referring to the cursor position (".") and
940	the cursor moves the display isn't updated for this change.  An update
941	is done when using the |CTRL-L| command (the whole screen is updated).
942	Example, to highlight the line where the cursor currently is: >
943		:exe '/\%' . line(".") . 'l'
944<	Alternatively use: >
945		/\%.l
946<	When 'hlsearch' is set and you move the cursor around and make changes
947	this will clearly show when the match is updated or not.
948
949						*/\%c* */\%>c* */\%<c*
950\%23c	Matches in a specific column.
951\%<23c	Matches before a specific column.
952\%>23c	Matches after a specific column.
953\%.c    Matches at the cursor column.
954\%<.c   Matches before the cursor column.
955\%>.c   Matches after the cursor column.
956	These six can be used to match specific columns in a buffer or string.
957	The "23" can be any column number.  The first column is 1.  Actually,
958	the column is the byte number (thus it's not exactly right for
959	multibyte characters).
960	WARNING: When inserting or deleting text Vim does not automatically
961	update the matches.  This means Syntax highlighting quickly becomes
962	wrong.  Also when referring to the cursor position (".") and
963	the cursor moves the display isn't updated for this change.  An update
964	is done when using the |CTRL-L| command (the whole screen is updated).
965	Example, to highlight the column where the cursor currently is: >
966		:exe '/\%' . col(".") . 'c'
967<	Alternatively use: >
968		/\%.c
969<	When 'hlsearch' is set and you move the cursor around and make changes
970	this will clearly show when the match is updated or not.
971	Example for matching a single byte in column 44: >
972		/\%>43c.\%<46c
973<	Note that "\%<46c" matches in column 45 when the "." matches a byte in
974	column 44.
975						*/\%v* */\%>v* */\%<v*
976\%23v	Matches in a specific virtual column.
977\%<23v	Matches before a specific virtual column.
978\%>23v	Matches after a specific virtual column.
979\%.v    Matches at the current virtual column.
980\%<.v   Matches before the current virtual column.
981\%>.v   Matches after the current virtual column.
982	These six can be used to match specific virtual columns in a buffer or
983	string.  When not matching with a buffer in a window, the option
984	values of the current window are used (e.g., 'tabstop').
985	The "23" can be any column number.  The first column is 1.
986	Note that some virtual column positions will never match, because they
987	are halfway through a tab or other character that occupies more than
988	one screen character.
989	WARNING: When inserting or deleting text Vim does not automatically
990	update highlighted matches.  This means Syntax highlighting quickly
991	becomes wrong.  Also when referring to the cursor position (".") and
992	the cursor moves the display isn't updated for this change.  An update
993	is done when using the |CTRL-L| command (the whole screen is updated).
994	Example, to highlight all the characters after virtual column 72: >
995		/\%>72v.*
996<	When 'hlsearch' is set and you move the cursor around and make changes
997	this will clearly show when the match is updated or not.
998	To match the text up to column 17: >
999		/^.*\%17v
1000<	To match all characters after the current virtual column (where the
1001	cursor is): >
1002		/\%>.v.*
1003<	Column 17 is not included, because this is a |/zero-width| match. To
1004	include the column use: >
1005		/^.*\%17v.
1006<	This command does the same thing, but also matches when there is no
1007	character in column 17: >
1008		/^.*\%<18v.
1009<	Note that without the "^" to anchor the match in the first column,
1010	this will also highlight column 17: >
1011		/.*\%17v
1012<	Column 17 is highlighted by 'hlsearch' because there is another match
1013	where ".*" matches zero characters.
1014
1015
1016Character classes:
1017\i	identifier character (see 'isident' option)	*/\i*
1018\I	like "\i", but excluding digits			*/\I*
1019\k	keyword character (see 'iskeyword' option)	*/\k*
1020\K	like "\k", but excluding digits			*/\K*
1021\f	file name character (see 'isfname' option)	*/\f*
1022\F	like "\f", but excluding digits			*/\F*
1023\p	printable character (see 'isprint' option)	*/\p*
1024\P	like "\p", but excluding digits			*/\P*
1025
1026NOTE: the above also work for multibyte characters.  The ones below only
1027match ASCII characters, as indicated by the range.
1028
1029						*whitespace* *white-space*
1030\s	whitespace character: <Space> and <Tab>		*/\s*
1031\S	non-whitespace character; opposite of \s	*/\S*
1032\d	digit:				[0-9]		*/\d*
1033\D	non-digit:			[^0-9]		*/\D*
1034\x	hex digit:			[0-9A-Fa-f]	*/\x*
1035\X	non-hex digit:			[^0-9A-Fa-f]	*/\X*
1036\o	octal digit:			[0-7]		*/\o*
1037\O	non-octal digit:		[^0-7]		*/\O*
1038\w	word character:			[0-9A-Za-z_]	*/\w*
1039\W	non-word character:		[^0-9A-Za-z_]	*/\W*
1040\h	head of word character:		[A-Za-z_]	*/\h*
1041\H	non-head of word character:	[^A-Za-z_]	*/\H*
1042\a	alphabetic character:		[A-Za-z]	*/\a*
1043\A	non-alphabetic character:	[^A-Za-z]	*/\A*
1044\l	lowercase character:		[a-z]		*/\l*
1045\L	non-lowercase character:	[^a-z]		*/\L*
1046\u	uppercase character:		[A-Z]		*/\u*
1047\U	non-uppercase character:	[^A-Z]		*/\U*
1048
1049	NOTE: Using the atom is faster than the [] form.
1050
1051	NOTE: 'ignorecase', "\c" and "\C" are not used by character classes.
1052
1053			*/\_* *E63* */\_i* */\_I* */\_k* */\_K* */\_f* */\_F*
1054			*/\_p* */\_P* */\_s* */\_S* */\_d* */\_D* */\_x* */\_X*
1055			*/\_o* */\_O* */\_w* */\_W* */\_h* */\_H* */\_a* */\_A*
1056			*/\_l* */\_L* */\_u* */\_U*
1057\_x	Where "x" is any of the characters above: The character class with
1058	end-of-line added
1059(end of character classes)
1060
1061\e	matches <Esc>					*/\e*
1062\t	matches <Tab>					*/\t*
1063\r	matches <CR>					*/\r*
1064\b	matches <BS>					*/\b*
1065\n	matches an end-of-line				*/\n*
1066	When matching in a string instead of buffer text a literal newline
1067	character is matched.
1068
1069~	matches the last given substitute string	*/~* */\~*
1070
1071\(\)	A pattern enclosed by escaped parentheses.	*/\(* */\(\)* */\)*
1072	E.g., "\(^a\)" matches 'a' at the start of a line.
1073	*E51* *E54* *E55* *E872* *E873*
1074
1075\1      Matches the same string that was matched by	*/\1* *E65*
1076	the first sub-expression in \( and \).
1077	Example: "\([a-z]\).\1" matches "ata", "ehe", "tot", etc.
1078\2      Like "\1", but uses second sub-expression,	*/\2*
1079   ...							*/\3*
1080\9      Like "\1", but uses ninth sub-expression.	*/\9*
1081	Note: The numbering of groups is done based on which "\(" comes first
1082	in the pattern (going left to right), NOT based on what is matched
1083	first.
1084
1085\%(\)	A pattern enclosed by escaped parentheses.	*/\%(\)* */\%(* *E53*
1086	Just like \(\), but without counting it as a sub-expression.  This
1087	allows using more groups and it's a little bit faster.
1088
1089x	A single character, with no special meaning, matches itself
1090
1091							*/\* */\\*
1092\x	A backslash followed by a single character, with no special meaning,
1093	is reserved for future expansions
1094
1095[]	(with 'nomagic': \[])		*/[]* */\[]* */\_[]* */collection*
1096\_[]
1097	A collection.  This is a sequence of characters enclosed in square
1098	brackets.  It matches any single character in the collection.
1099	Example		matches ~
1100	[xyz]		any 'x', 'y' or 'z'
1101	[a-zA-Z]$	any alphabetic character at the end of a line
1102	\c[a-z]$	same
1103	[А-яЁё]		Russian alphabet (with utf-8 and cp1251)
1104
1105								*/[\n]*
1106	With "\_" prepended the collection also includes the end-of-line.
1107	The same can be done by including "\n" in the collection.  The
1108	end-of-line is also matched when the collection starts with "^"!  Thus
1109	"\_[^ab]" matches the end-of-line and any character but "a" and "b".
1110	This makes it Vi compatible: Without the "\_" or "\n" the collection
1111	does not match an end-of-line.
1112								*E769*
1113	When the ']' is not there Vim will not give an error message but
1114	assume no collection is used.  Useful to search for '['.  However, you
1115	do get E769 for internal searching.  And be aware that in a
1116	`:substitute` command the whole command becomes the pattern.  E.g.
1117	":s/[/x/" searches for "[/x" and replaces it with nothing.  It does
1118	not search for "[" and replaces it with "x"!
1119
1120								*E944* *E945*
1121	If the sequence begins with "^", it matches any single character NOT
1122	in the collection: "[^xyz]" matches anything but 'x', 'y' and 'z'.
1123	- If two characters in the sequence are separated by '-', this is
1124	  shorthand for the full list of ASCII characters between them.  E.g.,
1125	  "[0-9]" matches any decimal digit. If the starting character exceeds
1126	  the ending character, e.g. [c-a], E944 occurs. Non-ASCII characters
1127	  can be used, but the character values must not be more than 256 apart
1128	  in the old regexp engine. For example, searching by [\u3000-\u4000]
1129	  after setting re=1 emits a E945 error. Prepending \%#=2 will fix it.
1130	- A character class expression is evaluated to the set of characters
1131	  belonging to that character class.  The following character classes
1132	  are supported:
1133		  Name	      Func	Contents ~
1134*[:alnum:]*	  [:alnum:]   isalnum	ASCII letters and digits
1135*[:alpha:]*	  [:alpha:]   isalpha  	ASCII letters
1136*[:blank:]*	  [:blank:]     	space and tab
1137*[:cntrl:]*	  [:cntrl:]   iscntrl 	ASCII control characters
1138*[:digit:]*	  [:digit:]     	decimal digits '0' to '9'
1139*[:graph:]*	  [:graph:]   isgraph	ASCII printable characters excluding
1140					space
1141*[:lower:]*	  [:lower:]   (1)	lowercase letters (all letters when
1142					'ignorecase' is used)
1143*[:print:]*	  [:print:]   (2) 	printable characters including space
1144*[:punct:]*	  [:punct:]   ispunct	ASCII punctuation characters
1145*[:space:]*	  [:space:]     	whitespace characters: space, tab, CR,
1146					NL, vertical tab, form feed
1147*[:upper:]*	  [:upper:]   (3)	uppercase letters (all letters when
1148					'ignorecase' is used)
1149*[:xdigit:]*	  [:xdigit:]    	hexadecimal digits: 0-9, a-f, A-F
1150*[:return:]*	  [:return:]		the <CR> character
1151*[:tab:]*	  [:tab:]		the <Tab> character
1152*[:escape:]*	  [:escape:]		the <Esc> character
1153*[:backspace:]*	  [:backspace:]		the <BS> character
1154*[:ident:]*	  [:ident:]		identifier character (same as "\i")
1155*[:keyword:]*	  [:keyword:]		keyword character (same as "\k")
1156*[:fname:]*	  [:fname:]		file name character (same as "\f")
1157	  The square brackets in character class expressions are additional to
1158	  the square brackets delimiting a collection.  For example, the
1159	  following is a plausible pattern for a UNIX filename:
1160	  "[-./[:alnum:]_~]\+".  That is, a list of at least one character,
1161	  each of which is either '-', '.', '/', alphabetic, numeric, '_' or
1162	  '~'.
1163	  These items only work for 8-bit characters, except [:lower:] and
1164	  [:upper:] also work for multibyte characters when using the new
1165	  regexp engine.  See |two-engines|.  In the future these items may
1166	  work for multibyte characters.  For now, to get all "alpha"
1167	  characters you can use: [[:lower:][:upper:]].
1168
1169	  The "Func" column shows what library function is used.  The
1170	  implementation depends on the system.  Otherwise:
1171	  (1) Uses islower() for ASCII and Vim builtin rules for other
1172	  characters.
1173	  (2) Uses Vim builtin rules
1174	  (3) As with (1) but using isupper()
1175							*/[[=* *[==]*
1176	- An equivalence class.  This means that characters are matched that
1177	  have almost the same meaning, e.g., when ignoring accents.  This
1178	  only works for Unicode, latin1 and latin9.  The form is:
1179		[=a=]
1180							*/[[.* *[..]*
1181	- A collation element.  This currently simply accepts a single
1182	  character in the form:
1183		[.a.]
1184							  */\]*
1185	- To include a literal ']', '^', '-' or '\' in the collection, put a
1186	  backslash before it: "[xyz\]]", "[\^xyz]", "[xy\-z]" and "[xyz\\]".
1187	  (Note: POSIX does not support the use of a backslash this way).  For
1188	  ']' you can also make it the first character (following a possible
1189	  "^"):  "[]xyz]" or "[^]xyz]".
1190	  For '-' you can also make it the first or last character: "[-xyz]",
1191	  "[^-xyz]" or "[xyz-]".  For '\' you can also let it be followed by
1192	  any character that's not in "^]-\bdertnoUux".  "[\xyz]" matches '\',
1193	  'x', 'y' and 'z'.  It's better to use "\\" though, future expansions
1194	  may use other characters after '\'.
1195	- Omitting the trailing ] is not considered an error. "[]" works like
1196	  "[]]", it matches the ']' character.
1197	- The following translations are accepted when the 'l' flag is not
1198	  included in 'cpoptions':
1199		\e	<Esc>
1200		\t	<Tab>
1201		\r	<CR>	(NOT end-of-line!)
1202		\b	<BS>
1203		\n	line break, see above |/[\n]|
1204		\d123	decimal number of character
1205		\o40	octal number of character up to 0o377
1206		\x20	hexadecimal number of character up to 0xff
1207		\u20AC	hex. number of multibyte character up to 0xffff
1208		\U1234	hex. number of multibyte character up to 0xffffffff
1209	  NOTE: The other backslash codes mentioned above do not work inside
1210	  []!
1211	- Matching with a collection can be slow, because each character in
1212	  the text has to be compared with each character in the collection.
1213	  Use one of the other atoms above when possible.  Example: "\d" is
1214	  much faster than "[0-9]" and matches the same characters.  However,
1215	  the new |NFA| regexp engine deals with this better than the old one.
1216
1217						*/\%[]* *E69* *E70* *E369*
1218\%[]	A sequence of optionally matched atoms.  This always matches.
1219	It matches as much of the list of atoms it contains as possible.  Thus
1220	it stops at the first atom that doesn't match.  For example: >
1221		/r\%[ead]
1222<	matches "r", "re", "rea" or "read".  The longest that matches is used.
1223	To match the Ex command "function", where "fu" is required and
1224	"nction" is optional, this would work: >
1225		/\<fu\%[nction]\>
1226<	The end-of-word atom "\>" is used to avoid matching "fu" in "full".
1227	It gets more complicated when the atoms are not ordinary characters.
1228	You don't often have to use it, but it is possible.  Example: >
1229		/\<r\%[[eo]ad]\>
1230<	Matches the words "r", "re", "ro", "rea", "roa", "read" and "road".
1231	There can be no \(\), \%(\) or \z(\) items inside the [] and \%[] does
1232	not nest.
1233	To include a "[" use "[[]" and for "]" use []]", e.g.,: >
1234		/index\%[[[]0[]]]
1235<	matches "index" "index[", "index[0" and "index[0]".
1236	{not available when compiled without the |+syntax| feature}
1237
1238				*/\%d* */\%x* */\%o* */\%u* */\%U* *E678*
1239
1240\%d123	Matches the character specified with a decimal number.  Must be
1241	followed by a non-digit.
1242\%o40	Matches the character specified with an octal number up to 0o377.
1243	Numbers below 0o40 must be followed by a non-octal digit or a
1244	non-digit.
1245\%x2a	Matches the character specified with up to two hexadecimal characters.
1246\%u20AC	Matches the character specified with up to four hexadecimal
1247	characters.
1248\%U1234abcd	Matches the character specified with up to eight hexadecimal
1249	characters, up to 0x7fffffff
1250
1251==============================================================================
12527. Ignoring case in a pattern					*/ignorecase*
1253
1254If the 'ignorecase' option is on, the case of normal letters is ignored.
1255'smartcase' can be set to ignore case when the pattern contains lowercase
1256letters only.
1257							*/\c* */\C*
1258When "\c" appears anywhere in the pattern, the whole pattern is handled like
1259'ignorecase' is on.  The actual value of 'ignorecase' and 'smartcase' is
1260ignored.  "\C" does the opposite: Force matching case for the whole pattern.
1261{only Vim supports \c and \C}
1262Note that 'ignorecase', "\c" and "\C" are not used for the character classes.
1263
1264Examples:
1265      pattern	'ignorecase'  'smartcase'	matches ~
1266	foo	  off		-		foo
1267	foo	  on		-		foo Foo FOO
1268	Foo	  on		off		foo Foo FOO
1269	Foo	  on		on		    Foo
1270	\cfoo	  -		-		foo Foo FOO
1271	foo\C	  -		-		foo
1272
1273Technical detail:				*NL-used-for-Nul*
1274<Nul> characters in the file are stored as <NL> in memory.  In the display
1275they are shown as "^@".  The translation is done when reading and writing
1276files.  To match a <Nul> with a search pattern you can just enter CTRL-@ or
1277"CTRL-V 000".  This is probably just what you expect.  Internally the
1278character is replaced with a <NL> in the search pattern.  What is unusual is
1279that typing CTRL-V CTRL-J also inserts a <NL>, thus also searches for a <Nul>
1280in the file.
1281
1282						*CR-used-for-NL*
1283When 'fileformat' is "mac", <NL> characters in the file are stored as <CR>
1284characters internally.  In the text they are shown as "^J".  Otherwise this
1285works similar to the usage of <NL> for a <Nul>.
1286
1287When working with expression evaluation, a <NL> character in the pattern
1288matches a <NL> in the string.  The use of "\n" (backslash n) to match a <NL>
1289doesn't work there, it only works to match text in the buffer.
1290
1291				*pattern-multi-byte* *pattern-multibyte*
1292Patterns will also work with multibyte characters, mostly as you would
1293expect.  But invalid bytes may cause trouble, a pattern with an invalid byte
1294will probably never match.
1295
1296==============================================================================
12978. Composing characters					*patterns-composing*
1298
1299							*/\Z*
1300When "\Z" appears anywhere in the pattern, all composing characters are
1301ignored.  Thus only the base characters need to match, the composing
1302characters may be different and the number of composing characters may differ.
1303Only relevant when 'encoding' is "utf-8".
1304Exception: If the pattern starts with one or more composing characters, these
1305must match.
1306							*/\%C*
1307Use "\%C" to skip any composing characters.  For example, the pattern "a" does
1308not match in "càt" (where the a has the composing character 0x0300), but
1309"a\%C" does.  Note that this does not match "cát" (where the á is character
13100xe1, it does not have a compositing character).  It does match "cat" (where
1311the a is just an a).
1312
1313When a composing character appears at the start of the pattern or after an
1314item that doesn't include the composing character, a match is found at any
1315character that includes this composing character.
1316
1317When using a dot and a composing character, this works the same as the
1318composing character by itself, except that it doesn't matter what comes before
1319this.
1320
1321The order of composing characters does not matter.  Also, the text may have
1322more composing characters than the pattern, it still matches.  But all
1323composing characters in the pattern must be found in the text.
1324
1325Suppose B is a base character and x and y are composing characters:
1326	pattern		text		match ~
1327	Bxy		Bxy		yes (perfect match)
1328	Bxy		Byx		yes (order ignored)
1329	Bxy		By		no (x missing)
1330	Bxy		Bx		no (y missing)
1331	Bx		Bx		yes (perfect match)
1332	Bx		By		no (x missing)
1333	Bx		Bxy		yes (extra y ignored)
1334	Bx		Byx		yes (extra y ignored)
1335
1336==============================================================================
13379. Compare with Perl patterns				*perl-patterns*
1338
1339Vim's regexes are most similar to Perl's, in terms of what you can do.  The
1340difference between them is mostly just notation;  here's a summary of where
1341they differ:
1342
1343Capability			in Vimspeak	in Perlspeak ~
1344----------------------------------------------------------------
1345force case insensitivity	\c		(?i)
1346force case sensitivity		\C		(?-i)
1347backref-less grouping		\%(atom\)	(?:atom)
1348conservative quantifiers	\{-n,m}		*?, +?, ??, {}?
13490-width match			atom\@=		(?=atom)
13500-width non-match		atom\@!		(?!atom)
13510-width preceding match		atom\@<=	(?<=atom)
13520-width preceding non-match	atom\@<!	(?<!atom)
1353match without retry		atom\@>		(?>atom)
1354
1355Vim and Perl handle newline characters inside a string a bit differently:
1356
1357In Perl, ^ and $ only match at the very beginning and end of the text,
1358by default, but you can set the 'm' flag, which lets them match at
1359embedded newlines as well.  You can also set the 's' flag, which causes
1360a . to match newlines as well.  (Both these flags can be changed inside
1361a pattern using the same syntax used for the i flag above, BTW.)
1362
1363On the other hand, Vim's ^ and $ always match at embedded newlines, and
1364you get two separate atoms, \%^ and \%$, which only match at the very
1365start and end of the text, respectively.  Vim solves the second problem
1366by giving you the \_ "modifier":  put it in front of a . or a character
1367class, and they will match newlines as well.
1368
1369Finally, these constructs are unique to Perl:
1370- execution of arbitrary code in the regex:  (?{perl code})
1371- conditional expressions:  (?(condition)true-expr|false-expr)
1372
1373...and these are unique to Vim:
1374- changing the magic-ness of a pattern:  \v \V \m \M
1375   (very useful for avoiding backslashitis)
1376- sequence of optionally matching atoms:  \%[atoms]
1377- \& (which is to \| what "and" is to "or";  it forces several branches
1378   to match at one spot)
1379- matching lines/columns by number:  \%5l \%5c \%5v
1380- setting the start and end of the match:  \zs \ze
1381
1382==============================================================================
138310. Highlighting matches				*match-highlight*
1384
1385							*:mat* *:match*
1386:mat[ch] {group} /{pattern}/
1387		Define a pattern to highlight in the current window.  It will
1388		be highlighted with {group}.  Example: >
1389			:highlight MyGroup ctermbg=green guibg=green
1390			:match MyGroup /TODO/
1391<		Instead of // any character can be used to mark the start and
1392		end of the {pattern}.  Watch out for using special characters,
1393		such as '"' and '|'.
1394
1395		{group} must exist at the moment this command is executed.
1396
1397		The {group} highlighting still applies when a character is
1398		to be highlighted for 'hlsearch', as the highlighting for
1399		matches is given higher priority than that of 'hlsearch'.
1400		Syntax highlighting (see 'syntax') is also overruled by
1401		matches.
1402
1403		Note that highlighting the last used search pattern with
1404		'hlsearch' is used in all windows, while the pattern defined
1405		with ":match" only exists in the current window.  It is kept
1406		when switching to another buffer.
1407
1408		'ignorecase' does not apply, use |/\c| in the pattern to
1409		ignore case.  Otherwise case is not ignored.
1410
1411		'redrawtime' defines the maximum time searched for pattern
1412		matches.
1413
1414		When matching end-of-line and Vim redraws only part of the
1415		display you may get unexpected results.  That is because Vim
1416		looks for a match in the line where redrawing starts.
1417
1418		Also see |matcharg()| and |getmatches()|. The former returns
1419		the highlight group and pattern of a previous |:match|
1420		command.  The latter returns a list with highlight groups and
1421		patterns defined by both |matchadd()| and |:match|.
1422
1423		Highlighting matches using |:match| are limited to three
1424		matches (aside from |:match|, |:2match| and |:3match| are
1425		available). |matchadd()| does not have this limitation and in
1426		addition makes it possible to prioritize matches.
1427
1428		Another example, which highlights all characters in virtual
1429		column 72 and more: >
1430			:highlight rightMargin term=bold ctermfg=blue guifg=blue
1431			:match rightMargin /.\%>72v/
1432<		To highlight all character that are in virtual column 7: >
1433			:highlight col8 ctermbg=grey guibg=grey
1434			:match col8 /\%<8v.\%>7v/
1435<		Note the use of two items to also match a character that
1436		occupies more than one virtual column, such as a TAB.
1437
1438:mat[ch]
1439:mat[ch] none
1440		Clear a previously defined match pattern.
1441
1442
1443:2mat[ch] {group} /{pattern}/					*:2match*
1444:2mat[ch]
1445:2mat[ch] none
1446:3mat[ch] {group} /{pattern}/					*:3match*
1447:3mat[ch]
1448:3mat[ch] none
1449		Just like |:match| above, but set a separate match.  Thus
1450		there can be three matches active at the same time.  The match
1451		with the lowest number has priority if several match at the
1452		same position.
1453		The ":3match" command is used by the |matchparen| plugin.  You
1454		are suggested to use ":match" for manual matching and
1455		":2match" for another plugin.
1456
1457==============================================================================
145811. Fuzzy matching					*fuzzy-match*
1459
1460Fuzzy matching refers to matching strings using a non-exact search string.
1461Fuzzy matching will match a string, if all the characters in the search string
1462are present anywhere in the string in the same order. Case is ignored.  In a
1463matched string, other characters can be present between two consecutive
1464characters in the search string. If the search string has multiple words, then
1465each word is matched separately. So the words in the search string can be
1466present in any order in a string.
1467
1468Fuzzy matching assigns a score for each matched string based on the following
1469criteria:
1470    - The number of sequentially matching characters.
1471    - The number of characters (distance) between two consecutive matching
1472      characters.
1473    - Matches at the beginning of a word
1474    - Matches at a camel case character (e.g. Case in CamelCase)
1475    - Matches after a path separator or a hyphen.
1476    - The number of unmatched characters in a string.
1477The matching string with the highest score is returned first.
1478
1479For example, when you search for the "get pat" string using fuzzy matching, it
1480will match the strings "GetPattern", "PatternGet", "getPattern", "patGetter",
1481"getSomePattern", "MatchpatternGet" etc.
1482
1483The functions |matchfuzzy()| and |matchfuzzypos()| can be used to fuzzy search
1484a string in a List of strings. The matchfuzzy() function returns a List of
1485matching strings. The matchfuzzypos() functions returns the List of matches,
1486the matching positions and the fuzzy match scores.
1487
1488The "f" flag of `:vimgrep` enables fuzzy matching.
1489
1490
1491
1492 vim:tw=78:ts=8:noet:ft=help:norl:
1493