tagsrch.txt For Vim version 9.1. Last change: 2024 Aug 01
VIM REFERENCE MANUAL by Bram Moolenaar
Tags and special searches tags-and-searches
See section 29.1 of the user manual for an introduction.
1. Jump to a tag tag-commands
2. Tag stack tag-stack
3. Tag match list tag-matchlist
4. Tags details tag-details
5. Tags file format tags-file-format
6. Include file searches include-search
7. Using 'tagfunc' tag-function
==============================================================================
1. Jump to a tag tag-commands
tag tags
A tag is an identifier that appears in a "tags" file. It is a sort of label
that can be jumped to. For example: In C programs each function name can be
used as a tag. The "tags" file has to be generated by a program like ctags,
before the tag commands can be used.
With the ":tag" command the cursor will be positioned on the tag. With the
CTRL-] command, the keyword on which the cursor is standing is used as the
tag. If the cursor is not on a keyword, the first keyword to the right of the
cursor is used.
The ":tag" command works very well for C programs. If you see a call to a
function and wonder what that function does, position the cursor inside of the
function name and hit CTRL-]. This will bring you to the function definition.
An easy way back is with the CTRL-T command. Also read about the tag stack
below.
:ta :tag E426 E429
:[count]ta[g][!] {name}
Jump to the definition of {name}, using the
information in the tags file(s). Put {name} in the
tag stack. See tag-! for [!].
{name} can be a regexp pattern, see tag-regexp.
When there are several matching tags for {name}, jump
to the [count] one. When [count] is omitted the
first one is jumped to. See tag-matchlist for
jumping to other matching tags.
g<LeftMouse> g<LeftMouse>
<C-LeftMouse> <C-LeftMouse> CTRL-]
CTRL-] Jump to the definition of the keyword under the
cursor. Same as ":tag {name}", where {name} is the
keyword under or after cursor.
When there are several matching tags for {name}, jump
to the [count] one. When no [count] is given the
first one is jumped to. See tag-matchlist for
jumping to other matching tags.
v_CTRL-]
{Visual}CTRL-] Same as ":tag {name}", where {name} is the text that
is highlighted.
telnet-CTRL-]
CTRL-] is the default telnet escape key. When you type CTRL-] to jump to a
tag, you will get the telnet prompt instead. Most versions of telnet allow
changing or disabling the default escape key. See the telnet man page. You
can 'telnet -E {Hostname}' to disable the escape character, or 'telnet -e
{EscapeCharacter} {Hostname}' to specify another escape character. If
possible, try to use "ssh" instead of "telnet" to avoid this problem.
tag-priority
When there are multiple matches for a tag, this priority is used:
1. "FSC" A full matching static tag for the current file.
2. "F C" A full matching global tag for the current file.
3. "F " A full matching global tag for another file.
4. "FS " A full matching static tag for another file.
5. " SC" An ignore-case matching static tag for the current file.
6. " C" An ignore-case matching global tag for the current file.
7. " " An ignore-case matching global tag for another file.
8. " S " An ignore-case matching static tag for another file.
Note that when the current file changes, the priority list is mostly not
changed, to avoid confusion when using ":tnext". It is changed when using
":tag {name}".
The ignore-case matches are not found for a ":tag" command when:
- 'tagcase' is "followic" and the 'ignorecase' option is off
- 'tagcase' is "followscs" and the 'ignorecase' option is off and the
'smartcase' option is off or the pattern contains an upper case character.
- 'tagcase' is "match"
- 'tagcase' is "smart" and the pattern contains an upper case character.
The ignore-case matches are found when:
- a pattern is used (starting with a "/")
- for ":tselect"
- when 'tagcase' is "followic" and 'ignorecase' is on
- when 'tagcase' is "followscs" and 'ignorecase' is on or the 'smartcase'
option is on and the pattern does not contain an upper case character
- when 'tagcase' is "ignore"
- when 'tagcase' is "smart" and the pattern does not contain an upper case
character
Note that using ignore-case tag searching disables binary searching in the
tags file, which causes a slowdown. This can be avoided by fold-case sorting
the tag file. See the 'tagbsearch' option for an explanation.
==============================================================================
2. Tag stack tag-stack tagstack E425
On the tag stack is remembered which tags you jumped to, and from where.
Tags are only pushed onto the stack when the 'tagstack' option is set.
g<RightMouse> g<RightMouse>
<C-RightMouse> <C-RightMouse> CTRL-T
CTRL-T Jump to [count] older entry in the tag stack
(default 1).
:po :pop E555 E556
:[count]po[p][!] Jump to [count] older entry in tag stack (default 1).
See tag-! for [!].
:[count]ta[g][!] Jump to [count] newer entry in tag stack (default 1).
See tag-! for [!].
:tags
:tags Show the contents of the tag stack. The active
entry is marked with a '>'.
The output of ":tags" looks like this:
# TO tag FROM line in file/text
1 1 main 1 harddisk2:text/vim/test
> 2 2 FuncA 58 i = FuncA(10);
3 1 FuncC 357 harddisk2:text/vim/src/amiga.c
This list shows the tags that you jumped to and the cursor position before
that jump. The older tags are at the top, the newer at the bottom.
The '>' points to the active entry. This is the tag that will be used by the
next ":tag" command. The CTRL-T and ":pop" command will use the position
above the active entry.
Below the "TO" is the number of the current match in the match list. Note
that this doesn't change when using ":pop" or ":tag".
The line number and file name are remembered to be able to get back to where
you were before the tag command. The line number will be correct, also when
deleting/inserting lines, unless this was done by another program (e.g.
another instance of Vim).
For the current file, the "file/text" column shows the text at the position.
An indent is removed and a long line is truncated to fit in the window.
You can jump to previously used tags with several commands. Some examples:
":pop" or CTRL-T to position before previous tag
{count}CTRL-T to position before {count} older tag
":tag" to newer tag
":0tag" to last used tag
The most obvious way to use this is while browsing through the call graph of
a program. Consider the following call graph:
main ---> FuncA ---> FuncC
---> FuncB
(Explanation: main calls FuncA and FuncB; FuncA calls FuncC).
You can get from main to FuncA by using CTRL-] on the call to FuncA. Then
you can CTRL-] to get to FuncC. If you now want to go back to main you can
use CTRL-T twice. Then you can CTRL-] to FuncB.
If you issue a ":ta {name}" or CTRL-] command, this tag is inserted at the
current position in the stack. If the stack was full (it can hold up to 20
entries), the oldest entry is deleted and the older entries shift one
position up (their index number is decremented by one). If the last used
entry was not at the bottom, the entries below the last used one are
deleted. This means that an old branch in the call graph is lost. After the
commands explained above the tag stack will look like this:
# TO tag FROM line in file/text
1 1 main 1 harddisk2:text/vim/test
2 1 FuncB 59 harddisk2:text/vim/src/main.c
The gettagstack() function returns the tag stack of a specified window. The
settagstack() function modifies the tag stack of a window.
tagstack-examples
Write to the tag stack just like :tag but with a user-defined
jumper#jump_to_tag function:
Set current index of the tag stack to 4:
Push a new item onto the tag stack:
E73
When you try to use the tag stack while it doesn't contain anything you will
get an error message.
==============================================================================
3. Tag match list tag-matchlist E427 E428
When there are several matching tags, these commands can be used to jump
between them. Note that these commands don't change the tag stack, they keep
the same entry.
:ts :tselect
:ts[elect][!] [name] List the tags that match [name], using the
information in the tags file(s).
When [name] is not given, the last tag name from the
tag stack is used.
See tag-! for [!].
With a '>' in the first column is indicated which is
the current position in the list (if there is one).
[name] can be a regexp pattern, see tag-regexp.
See tag-priority for the priorities used in the
listing.
Example output:
See tag-priority for the "pri" column. Note that
this depends on the current file, thus using
":tselect xxx" can produce different results.
The "kind" column gives the kind of tag, if this was
included in the tags file.
The "info" column shows information that could be
found in the tags file. It depends on the program
that produced the tags file.
When the list is long, you may get the more-prompt.
If you already see the tag you want to use, you can
type 'q' and enter the number.
:sts :stselect
:sts[elect][!] [name] Does ":tselect[!] [name]" and splits the window for
the selected tag.
g]
g] Like CTRL-], but use ":tselect" instead of ":tag".
v_g]
{Visual}g] Same as "g]", but use the highlighted text as the
identifier.
:tj :tjump
:tj[ump][!] [name] Like ":tselect", but jump to the tag directly when
there is only one match.
:stj :stjump
:stj[ump][!] [name] Does ":tjump[!] [name]" and splits the window for the
selected tag.
g_CTRL-]
g CTRL-] Like CTRL-], but use ":tjump" instead of ":tag".
v_g_CTRL-]
{Visual}g CTRL-] Same as "g CTRL-]", but use the highlighted text as
the identifier.
:tn :tnext
:[count]tn[ext][!] Jump to [count] next matching tag (default 1). See
tag-! for [!].
:tp :tprevious
:[count]tp[revious][!] Jump to [count] previous matching tag (default 1).
See tag-! for [!].
:tN :tNext
:[count]tN[ext][!] Same as ":tprevious".
:tr :trewind
:[count]tr[ewind][!] Jump to first matching tag. If [count] is given, jump
to [count]th matching tag. See tag-! for [!].
:tf :tfirst
:[count]tf[irst][!] Same as ":trewind".
:tl :tlast
:tl[ast][!] Jump to last matching tag. See tag-! for [!].
:lt :ltag
:lt[ag][!] [name] Jump to tag [name] and add the matching tags to a new
location list for the current window. [name] can be
a regexp pattern, see tag-regexp. When [name] is
not given, the last tag name from the tag stack is
used. The search pattern to locate the tag line is
prefixed with "\V" to escape all the special
characters (very nomagic). The location list showing
the matching tags is independent of the tag stack.
See tag-! for [!].
When there is no other message, Vim shows which matching tag has been jumped
to, and the number of matching tags:
When using ":tnext" a few times, or with ":tlast", more matches may be found.
When you didn't see this message because of some other message, or you just
want to know where you are, this command will show it again (and jump to the
same tag as last time):
tag-skip-file
When a matching tag is found for which the file doesn't exist, this match is
skipped and the next matching tag is used. Vim reports this, to notify you of
missing files. When the end of the list of matches has been reached, an error
message is given.
tag-preview
The tag match list can also be used in the preview window. The commands are
the same as above, with a "p" prepended.
{not available when compiled without the +quickfix feature}
:pts :ptselect
:pts[elect][!] [name] Does ":tselect[!] [name]" and shows the new tag in a
"Preview" window. See :ptag for more info.
:ptj :ptjump
:ptj[ump][!] [name] Does ":tjump[!] [name]" and shows the new tag in a
"Preview" window. See :ptag for more info.
:ptn :ptnext
:[count]ptn[ext][!] ":tnext" in the preview window. See :ptag.
:ptp :ptprevious
:[count]ptp[revious][!] ":tprevious" in the preview window. See :ptag.
:ptN :ptNext
:[count]ptN[ext][!] Same as ":ptprevious".
:ptr :ptrewind
:[count]ptr[ewind][!] ":trewind" in the preview window. See :ptag.
:ptf :ptfirst
:[count]ptf[irst][!] Same as ":ptrewind".
:ptl :ptlast
:ptl[ast][!] ":tlast" in the preview window. See :ptag.
==============================================================================
4. Tags details tag-details
static-tag
A static tag is a tag that is defined for a specific file. In a C program
this could be a static function.
In Vi jumping to a tag sets the current search pattern. This means that the
"n" command after jumping to a tag does not search for the same pattern that
it did before jumping to the tag. Vim does not do this as we consider it to
be a bug. If you really want the old Vi behavior, set the 't' flag in
'cpoptions'.
tag-binary-search
Vim uses binary searching in the tags file to find the desired tag quickly
(when enabled at compile time +tag_binary). But this only works if the
tags file was sorted on ASCII byte value. Therefore, if no match was found,
another try is done with a linear search. If you only want the linear search,
reset the 'tagbsearch' option. Or better: Sort the tags file!
Note that the binary searching is disabled when not looking for a tag with a
specific name. This happens when ignoring case and when a regular expression
is used that doesn't start with a fixed string. Tag searching can be a lot
slower then. The former can be avoided by case-fold sorting the tags file.
See 'tagbsearch' for details.
tag-regexp
The ":tag" and ":tselect" commands accept a regular expression argument. See
pattern for the special characters that can be used.
When the argument starts with '/', it is used as a pattern. If the argument
does not start with '/', it is taken literally, as a full tag name.
Examples:
When the argument both exists literally, and match when used as a regexp, a
literal match has a higher priority. For example, ":tag /open" matches "open"
before "open_file" and "file_open".
When using a pattern case is ignored. If you want to match case use "\C" in
the pattern.
tag-!
If the tag is in the current file this will always work. Otherwise the
performed actions depend on whether the current file was changed, whether a !
is added to the command and on the 'autowrite' and 'winfixbuf' options:
tag in file winfixbuf autowrite
current file changed ! option option action
-----------------------------------------------------------------------------
yes x x off x goto tag
no no x off x read other file, goto tag
no yes yes off x abandon current file,
read other file, goto tag
no yes no off on write current file,
read other file, goto tag
no yes no off off fail
yes x yes x x goto tag
no no no on x fail
no yes no on x fail
no yes no on on fail
no yes no on off fail
-----------------------------------------------------------------------------
- If the tag is in the current file, the command will always work.
- If the tag is in another file and the current file was not changed, the
other file will be made the current file and read into the buffer.
- If the tag is in another file, the current file was changed and a ! is
added to the command, the changes to the current file are lost, the other
file will be made the current file and read into the buffer.
- If the tag is in another file, the current file was changed and the
'autowrite' option is on, the current file will be written, the other
file will be made the current file and read into the buffer.
- If the tag is in another file, the current file was changed and the
'autowrite' option is off, the command will fail. If you want to save
the changes, use the ":w" command and then use ":tag" without an argument.
This works because the tag is put on the stack anyway. If you want to lose
the changes you can use the ":tag!" command.
- If the tag is in another file and the window includes 'winfixbuf', the
command will fail. If the tag is in the same file then it may succeed.
tag-security
Note that Vim forbids some commands, for security reasons. This works like
using the 'secure' option for exrc/vimrc files in the current directory. See
trojan-horse and sandbox.
When the {tagaddress} changes a buffer, you will get a warning message:
"WARNING: tag command changed a buffer!!!"
In a future version changing the buffer will be impossible. All this for
security reasons: Somebody might hide a nasty command in the tags file, which
would otherwise go unnoticed. Example:
In Vi the ":tag" command sets the last search pattern when the tag is searched
for. In Vim this is not done, the previous search pattern is still remembered,
unless the 't' flag is present in 'cpoptions'.
emacs-tags emacs_tags E430
Emacs style tag files are only supported if Vim was compiled with the
+emacs_tags feature enabled. Sorry, there is no explanation about Emacs tag
files here, it is only supported for backwards compatibility :-).
Lines in Emacs tags files can be very long. Vim only deals with lines of up
to about 510 bytes. To see whether lines are ignored set 'verbose' to 5 or
higher. Non-Emacs tags file lines can be any length.
tags-option
The 'tags' option is a list of file names. Each of these files is searched
for the tag. This can be used to use a different tags file than the default
file "tags". It can also be used to access a common tags file.
The next file in the list is not used when:
- A matching static tag for the current buffer has been found.
- A matching global tag has been found.
This also depends on whether case is ignored. Case is ignored when:
- 'tagcase' is "followic" and 'ignorecase' is set
- 'tagcase' is "ignore"
- 'tagcase' is "smart" and the pattern only contains lower case
characters.
- 'tagcase' is "followscs" and 'smartcase' is set and the pattern only
contains lower case characters.
If case is not ignored, and the tags file only has a match without matching
case, the next tags file is searched for a match with matching case. If no
tag with matching case is found, the first match without matching case is
used. If case is ignored, and a matching global tag with or without matching
case is found, this one is used, no further tags files are searched.
When a tag file name starts with "./", the '.' is replaced with the path of
the current file. This makes it possible to use a tags file in the directory
where the current file is (no matter what the current directory is). The idea
of using "./" is that you can define which tag file is searched first: In the
current directory ("tags,./tags") or in the directory of the current file
("./tags,tags").
For example:
In this example the tag will first be searched for in the file "tags" in the
directory where the current file is. Next the "tags" file in the current
directory. If it is not found there, then the file "/home/user/commontags"
will be searched for the tag.
This can be switched off by including the 'd' flag in 'cpoptions', to make
it Vi compatible. "./tags" will then be the tags file in the current
directory, instead of the tags file in the directory where the current file
is.
Instead of the comma a space may be used. Then a backslash is required for
the space to be included in the string option:
To include a space in a file name use three backslashes. To include a comma
in a file name use two backslashes. For example, use:
for the files "tag file" and "/home/user/common,tags". The 'tags' option will
have the value "tag\ file,/home/user/common\,tags".
If the 'tagrelative' option is on (which is the default) and using a tag file
in another directory, file names in that tag file are relative to the
directory where the tag file is.
==============================================================================
5. Tags file format tags-file-format E431
ctags jtags
A tags file can be created with an external command, for example "ctags". It
will contain a tag for each function. Some versions of "ctags" will also make
a tag for each "#defined" macro, typedefs, enums, etc.
Some programs that generate tags files:
ctags As found on most Unix systems. Only supports C. Only
does the basic work.
universal ctags A maintained version of ctags based on exuberant
ctags. See https://ctags.io.
Exuberant_ctags
exuberant ctags This is a very good one. It works for C, C++, Java,
Fortran, Eiffel and others. It can generate tags for
many items. See http://ctags.sourceforge.net.
No new version since 2009.
etags Connected to Emacs. Supports many languages.
JTags For Java, in Java. It can be found at
http://www.fleiner.com/jtags/.
ptags.py For Python, in Python. Found in your Python source
directory at Tools/scripts/ptags.py.
ptags For Perl, in Perl. It can be found at (link seems
dead):
http://www.eleves.ens.fr:8080/home/nthiery/Tags/.
gnatxref For Ada. See http://www.gnuada.org/. gnatxref is
part of the gnat package.
The lines in the tags file must have one of these two formats:
1. {tagname} {TAB} {tagfile} {TAB} {tagaddress}
2. {tagname} {TAB} {tagfile} {TAB} {tagaddress} {term} {field} ..
Previously an old format was supported, see tag-old-static.
The first format is a normal tag, which is completely compatible with Vi. It
is the only format produced by traditional ctags implementations. This is
often used for functions that are global, also referenced in other files.
The lines in the tags file can end in <NL> or <CR><NL>. On the Macintosh <CR>
also works. The <CR> and <NL> characters can never appear inside a line.
The second format is new. It includes additional information in optional
fields at the end of each line. It is backwards compatible with Vi. It is
only supported by new versions of ctags (such as Universal ctags or Exuberant
ctags).
{tagname} The identifier. Normally the name of a function, but it can
be any identifier. It cannot contain a <Tab>.
{TAB} One <Tab> character. Note: previous versions allowed any
white space here. This has been abandoned to allow spaces in
{tagfile}.
{tagfile} The file that contains the definition of {tagname}. It can
have an absolute or relative path. It may contain environment
variables and wildcards (although the use of wildcards is
doubtful). It cannot contain a <Tab>.
{tagaddress} The Ex command that positions the cursor on the tag. It can
be any Ex command, although restrictions apply (see
tag-security). Posix only allows line numbers and search
commands, which are mostly used.
{term} ;" The two characters semicolon and double quote. This is
interpreted by Vi as the start of a comment, which makes the
following be ignored. This is for backwards compatibility
with Vi, it ignores the following fields. Example:
{term} must be |;". Here the bar ends the command (excluding
the bar) and ;" is used to have Vi ignore the rest of the
line. Example:
{field} .. A list of optional fields. Each field has the form:
<Tab>{fieldname}:{value}
The {fieldname} identifies the field, and can only contain
alphabetical characters [a-zA-Z].
The {value} is any string, but cannot contain a <Tab>.
These characters are special:
"\t" stands for a <Tab>
"\r" stands for a <CR>
"\n" stands for a <NL>
"\\" stands for a single '\' character
There is one field that doesn't have a ':'. This is the kind
of the tag. It is handled like it was preceded with "kind:".
In the above example, this was "kind:v" (typically variable).
See the documentation of ctags for the kinds it produces, with
ctags you can use ctags --list-kinds .
The only other field currently recognized by Vim is "file:"
(with an empty value). It is used for a static tag.
The first lines in the tags file can contain lines that start with
!_TAG_
These are sorted to the first lines, only rare tags that start with "!" can
sort to before them. Vim recognizes two items. The first one is the line
that indicates if the file was sorted. When this line is found, Vim uses
binary searching for the tags file:
!_TAG_FILE_SORTED<Tab>1<Tab>{anything}
A tag file may be case-fold sorted to avoid a linear search when case is
ignored. (Case is ignored when 'ignorecase' is set and 'tagcase' is
"followic", or when 'tagcase' is "ignore".) See 'tagbsearch' for details.
The value '2' should be used then:
!_TAG_FILE_SORTED<Tab>2<Tab>{anything}
The other tag that Vim recognizes is the encoding of the tags file:
!_TAG_FILE_ENCODING<Tab>utf-8<Tab>{anything}
Here "utf-8" is the encoding used for the tags. Vim will then convert the tag
being searched for from 'encoding' to the encoding of the tags file. And when
listing tags the reverse happens. When the conversion fails the unconverted
tag is used.
tag-search
The command can be any Ex command, but often it is a search command.
Examples:
tag1 file1 /^main(argc, argv)/
tag2 file2 108
The command is always executed with 'magic' not set. The only special
characters in a search pattern are "^" (begin-of-line) and "$" (<EOL>).
See pattern. Note that you must put a backslash before each backslash in
the search text. This is for backwards compatibility with Vi.
E434 E435
If the command is a normal search command (it starts and ends with "/" or
"?"), some special handling is done:
- Searching starts on line 1 of the file.
The direction of the search is forward for "/", backward for "?".
Note that 'wrapscan' does not matter, the whole file is always searched.
- If the search fails, another try is done ignoring case. If that fails too,
a search is done for:
names, this will find the function name when it is in column 0. This will
help when the arguments to the function have changed since the tags file was
made. If this search also fails another search is done with:
followed by white space and a '('. This will find macro names and function
names with a type prepended.
tag-old-static
Until March 2019 (patch 8.1.1092) an outdated format was supported:
{tagfile}:{tagname} {TAB} {tagfile} {TAB} {tagaddress}
This format is for a static tag only. It is obsolete now, replaced by
the second format. It is only supported by Elvis 1.x, older Vim versions and
a few versions of ctags. A static tag is often used for functions that are
local, only referenced in the file {tagfile}. Note that for the static tag,
the two occurrences of {tagfile} must be exactly the same. Also see
tags-option below, for how static tags are used.
The support was removed, since when you can update to the new Vim version you
should also be able to update ctags to one that supports the second format.
==============================================================================
6. Include file searches include-search definition-search
E387 E388 E389
These commands look for a string in the current file and in all encountered
included files (recursively). This can be used to find the definition of a
variable, function or macro. If you only want to search in the current
buffer, use the commands listed at pattern-searches.
These commands are not available when the +find_in_path feature was disabled
at compile time.
When a line is encountered that includes another file, that file is searched
before continuing in the current buffer. Files included by included files are
also searched. When an include file could not be found it is silently
ignored. Use the :checkpath command to discover which files could not be
found, possibly your 'path' option is not set up correctly. Note: the
included file is searched, not a buffer that may be editing that file. Only
for the current file the lines in the buffer are used.
The string can be any keyword or a defined macro. For the keyword any match
will be found. For defined macros only lines that match with the 'define'
option will be found. The default is "^#\s*define", which is for C programs.
For other languages you probably want to change this. See 'define' for an
example for C++. The string cannot contain an end-of-line, only matches
within a line are found.
When a match is found for a defined macro, the displaying of lines continues
with the next line when a line ends in a backslash.
The commands that start with "[" start searching from the start of the current
file. The commands that start with "]" start at the current cursor position.
The 'include' option is used to define a line that includes another file. The
default is "\^#\s*include", which is for C programs. Note: Vim does not
recognize C syntax, if the 'include' option matches a line inside
"#ifdef/#endif" or inside a comment, it is searched anyway. The 'isfname'
option is used to recognize the file name that comes after the matched
pattern.
The 'path' option is used to find the directory for the include files that
do not have an absolute path.
The 'comments' option is used for the commands that display a single line or
jump to a line. It defines patterns that may start a comment. Those lines
are ignored for the search, unless [!] is used. One exception: When the line
matches the pattern "^# *define" it is not considered to be a comment.
If you want to list matches, and then select one to jump to, you could use a
mapping to do that for you. Here is an example:
[i
[i Display the first line that contains the keyword
under the cursor. The search starts at the beginning
of the file. Lines that look like a comment are
ignored (see 'comments' option). If a count is given,
the count'th matching line is displayed, and comment
lines are not ignored.
]i
]i like "[i", but start at the current cursor position.
:is :isearch
:[range]is[earch][!] [count] [/]pattern[/]
Like "[i" and "]i", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
[I
[I Display all lines that contain the keyword under the
cursor. Filenames and line numbers are displayed
for the found lines. The search starts at the
beginning of the file.
]I
]I like "[I", but start at the current cursor position.
:il :ilist
:[range]il[ist][!] [/]pattern[/]
Like "[I" and "]I", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
[_CTRL-I
[ CTRL-I Jump to the first line that contains the keyword
under the cursor. The search starts at the beginning
of the file. Lines that look like a comment are
ignored (see 'comments' option). If a count is given,
the count'th matching line is jumped to, and comment
lines are not ignored.
]_CTRL-I
] CTRL-I like "[ CTRL-I", but start at the current cursor
position.
:ij :ijump
:[range]ij[ump][!] [count] [/]pattern[/]
Like "[ CTRL-I" and "] CTRL-I", but search in
[range] lines (default: whole file).
See :search-args for [/] and [!].
CTRL-W CTRL-I CTRL-W_CTRL-I CTRL-W_i
CTRL-W i Open a new window, with the cursor on the first line
that contains the keyword under the cursor. The
search starts at the beginning of the file. Lines
that look like a comment line are ignored (see
'comments' option). If a count is given, the count'th
matching line is jumped to, and comment lines are not
ignored.
:isp :isplit
:[range]isp[lit][!] [count] [/]pattern[/]
Like "CTRL-W i" and "CTRL-W i", but search in
[range] lines (default: whole file).
See :search-args for [/] and [!].
[d
[d Display the first macro definition that contains the
macro under the cursor. The search starts from the
beginning of the file. If a count is given, the
count'th matching line is displayed.
]d
]d like "[d", but start at the current cursor position.
:ds :dsearch
:[range]ds[earch][!] [count] [/]string[/]
Like "[d" and "]d", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
[D
[D Display all macro definitions that contain the macro
under the cursor. Filenames and line numbers are
displayed for the found lines. The search starts
from the beginning of the file.
]D
]D like "[D", but start at the current cursor position.
:dli :dlist
:[range]dli[st][!] [/]string[/]
Like [D and ]D, but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
Note that :dl works like :delete with the "l"
flag, not :dlist.
[_CTRL-D
[ CTRL-D Jump to the first macro definition that contains the
keyword under the cursor. The search starts from
the beginning of the file. If a count is given, the
count'th matching line is jumped to.
]_CTRL-D
] CTRL-D like "[ CTRL-D", but start at the current cursor
position.
:dj :djump
:[range]dj[ump][!] [count] [/]string[/]
Like "[ CTRL-D" and "] CTRL-D", but search in
[range] lines (default: whole file).
See :search-args for [/] and [!].
CTRL-W CTRL-D CTRL-W_CTRL-D CTRL-W_d
CTRL-W d Open a new window, with the cursor on the first
macro definition line that contains the keyword
under the cursor. The search starts from the
beginning of the file. If a count is given, the
count'th matching line is jumped to.
:dsp :dsplit
:[range]dsp[lit][!] [count] [/]string[/]
Like "CTRL-W d", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
:che :chec :check :checkpath
:che[ckpath] List all the included files that could not be found.
:che[ckpath]! List all the included files.
:search-args
Common arguments for the commands above:
[!] When included, find matches in lines that are recognized as comments.
When excluded, a match is ignored when the line is recognized as a
comment (according to 'comments'), or the match is in a C comment
(after "//" or inside /* */). Note that a match may be missed if a
line is recognized as a comment, but the comment ends halfway the line.
And if the line is a comment, but it is not recognized (according to
'comments') a match may be found in it anyway. Example:
a comment (even though syntax highlighting does recognize it).
Note: Since a macro definition mostly doesn't look like a comment, the
[!] makes no difference for ":dlist", ":dsearch" and ":djump".
[/] A pattern can be surrounded by '/'. Without '/' only whole words are
matched, using the pattern "\<pattern\>". Only after the second '/' a
next command can be appended with '|'. Example:
is used as a literal string, not as a search pattern.
==============================================================================
7. Using 'tagfunc' tag-function
It is possible to provide Vim with a function which will generate a list of
tags used for commands like :tag, :tselect and Normal mode tag commands
like CTRL-].
The function used for generating the taglist is specified by setting the
'tagfunc' option. The function will be called with three arguments:
pattern The tag identifier or pattern used during the tag search.
flags String containing flags to control the function behavior.
info Dict containing the following entries:
buf_ffname Full filename which can be used for priority.
user_data Custom data String, if stored in the tag
stack previously by tagfunc.
Note that in a legacy function "a:" needs to be prepended to the argument name
when using it.
Currently up to three flags may be passed to the tag function:
'c' The function was invoked by a normal command being processed
(mnemonic: the tag function may use the context around the
cursor to perform a better job of generating the tag list.)
'i' In Insert mode, the user was completing a tag (with
i_CTRL-X_CTRL-] or 'completeopt' contains t).
'r' The first argument to tagfunc should be interpreted as a
pattern (see tag-regexp), such as when using:
If this flag is not present, the argument is usually taken
literally as the full tag name.
Note that when 'tagfunc' is set, the priority of the tags described in
tag-priority does not apply. Instead, the priority is exactly as the
ordering of the elements in the list returned by the function.
E987
The function should return a List of Dict entries. Each Dict must at least
include the following entries and each value must be a string:
name Name of the tag.
filename Name of the file where the tag is defined. It is
either relative to the current directory or a full path.
cmd Ex command used to locate the tag in the file. This
can be either an Ex search pattern or a line number.
Note that the format is similar to that of taglist(), which makes it possible
to use its output to generate the result.
The following fields are optional:
kind Type of the tag.
user_data String of custom data stored in the tag stack which
can be used to disambiguate tags between operations.
If the function returns v:null instead of a List, a standard tag lookup will
be performed instead.
It is not allowed to change the tagstack from inside 'tagfunc'. E986
It is not allowed to close a window or change window from inside 'tagfunc'.
E1299
The following is a hypothetical example of a function used for 'tagfunc'. It
uses the output of taglist() to generate the result: a list of tags in the
inverse order of file names.
vim:tw=78:ts=8:noet:ft=help:norl:
VIM REFERENCE MANUAL by Bram Moolenaar
Tags and special searches tags-and-searches
See section 29.1 of the user manual for an introduction.
1. Jump to a tag tag-commands
2. Tag stack tag-stack
3. Tag match list tag-matchlist
4. Tags details tag-details
5. Tags file format tags-file-format
6. Include file searches include-search
7. Using 'tagfunc' tag-function
==============================================================================
1. Jump to a tag tag-commands
tag tags
A tag is an identifier that appears in a "tags" file. It is a sort of label
that can be jumped to. For example: In C programs each function name can be
used as a tag. The "tags" file has to be generated by a program like ctags,
before the tag commands can be used.
With the ":tag" command the cursor will be positioned on the tag. With the
CTRL-] command, the keyword on which the cursor is standing is used as the
tag. If the cursor is not on a keyword, the first keyword to the right of the
cursor is used.
The ":tag" command works very well for C programs. If you see a call to a
function and wonder what that function does, position the cursor inside of the
function name and hit CTRL-]. This will bring you to the function definition.
An easy way back is with the CTRL-T command. Also read about the tag stack
below.
:ta :tag E426 E429
:[count]ta[g][!] {name}
Jump to the definition of {name}, using the
information in the tags file(s). Put {name} in the
tag stack. See tag-! for [!].
{name} can be a regexp pattern, see tag-regexp.
When there are several matching tags for {name}, jump
to the [count] one. When [count] is omitted the
first one is jumped to. See tag-matchlist for
jumping to other matching tags.
g<LeftMouse> g<LeftMouse>
<C-LeftMouse> <C-LeftMouse> CTRL-]
CTRL-] Jump to the definition of the keyword under the
cursor. Same as ":tag {name}", where {name} is the
keyword under or after cursor.
When there are several matching tags for {name}, jump
to the [count] one. When no [count] is given the
first one is jumped to. See tag-matchlist for
jumping to other matching tags.
v_CTRL-]
{Visual}CTRL-] Same as ":tag {name}", where {name} is the text that
is highlighted.
telnet-CTRL-]
CTRL-] is the default telnet escape key. When you type CTRL-] to jump to a
tag, you will get the telnet prompt instead. Most versions of telnet allow
changing or disabling the default escape key. See the telnet man page. You
can 'telnet -E {Hostname}' to disable the escape character, or 'telnet -e
{EscapeCharacter} {Hostname}' to specify another escape character. If
possible, try to use "ssh" instead of "telnet" to avoid this problem.
tag-priority
When there are multiple matches for a tag, this priority is used:
1. "FSC" A full matching static tag for the current file.
2. "F C" A full matching global tag for the current file.
3. "F " A full matching global tag for another file.
4. "FS " A full matching static tag for another file.
5. " SC" An ignore-case matching static tag for the current file.
6. " C" An ignore-case matching global tag for the current file.
7. " " An ignore-case matching global tag for another file.
8. " S " An ignore-case matching static tag for another file.
Note that when the current file changes, the priority list is mostly not
changed, to avoid confusion when using ":tnext". It is changed when using
":tag {name}".
The ignore-case matches are not found for a ":tag" command when:
- 'tagcase' is "followic" and the 'ignorecase' option is off
- 'tagcase' is "followscs" and the 'ignorecase' option is off and the
'smartcase' option is off or the pattern contains an upper case character.
- 'tagcase' is "match"
- 'tagcase' is "smart" and the pattern contains an upper case character.
The ignore-case matches are found when:
- a pattern is used (starting with a "/")
- for ":tselect"
- when 'tagcase' is "followic" and 'ignorecase' is on
- when 'tagcase' is "followscs" and 'ignorecase' is on or the 'smartcase'
option is on and the pattern does not contain an upper case character
- when 'tagcase' is "ignore"
- when 'tagcase' is "smart" and the pattern does not contain an upper case
character
Note that using ignore-case tag searching disables binary searching in the
tags file, which causes a slowdown. This can be avoided by fold-case sorting
the tag file. See the 'tagbsearch' option for an explanation.
==============================================================================
2. Tag stack tag-stack tagstack E425
On the tag stack is remembered which tags you jumped to, and from where.
Tags are only pushed onto the stack when the 'tagstack' option is set.
g<RightMouse> g<RightMouse>
<C-RightMouse> <C-RightMouse> CTRL-T
CTRL-T Jump to [count] older entry in the tag stack
(default 1).
:po :pop E555 E556
:[count]po[p][!] Jump to [count] older entry in tag stack (default 1).
See tag-! for [!].
:[count]ta[g][!] Jump to [count] newer entry in tag stack (default 1).
See tag-! for [!].
:tags
:tags Show the contents of the tag stack. The active
entry is marked with a '>'.
The output of ":tags" looks like this:
# TO tag FROM line in file/text
1 1 main 1 harddisk2:text/vim/test
> 2 2 FuncA 58 i = FuncA(10);
3 1 FuncC 357 harddisk2:text/vim/src/amiga.c
This list shows the tags that you jumped to and the cursor position before
that jump. The older tags are at the top, the newer at the bottom.
The '>' points to the active entry. This is the tag that will be used by the
next ":tag" command. The CTRL-T and ":pop" command will use the position
above the active entry.
Below the "TO" is the number of the current match in the match list. Note
that this doesn't change when using ":pop" or ":tag".
The line number and file name are remembered to be able to get back to where
you were before the tag command. The line number will be correct, also when
deleting/inserting lines, unless this was done by another program (e.g.
another instance of Vim).
For the current file, the "file/text" column shows the text at the position.
An indent is removed and a long line is truncated to fit in the window.
You can jump to previously used tags with several commands. Some examples:
":pop" or CTRL-T to position before previous tag
{count}CTRL-T to position before {count} older tag
":tag" to newer tag
":0tag" to last used tag
The most obvious way to use this is while browsing through the call graph of
a program. Consider the following call graph:
main ---> FuncA ---> FuncC
---> FuncB
(Explanation: main calls FuncA and FuncB; FuncA calls FuncC).
You can get from main to FuncA by using CTRL-] on the call to FuncA. Then
you can CTRL-] to get to FuncC. If you now want to go back to main you can
use CTRL-T twice. Then you can CTRL-] to FuncB.
If you issue a ":ta {name}" or CTRL-] command, this tag is inserted at the
current position in the stack. If the stack was full (it can hold up to 20
entries), the oldest entry is deleted and the older entries shift one
position up (their index number is decremented by one). If the last used
entry was not at the bottom, the entries below the last used one are
deleted. This means that an old branch in the call graph is lost. After the
commands explained above the tag stack will look like this:
# TO tag FROM line in file/text
1 1 main 1 harddisk2:text/vim/test
2 1 FuncB 59 harddisk2:text/vim/src/main.c
The gettagstack() function returns the tag stack of a specified window. The
settagstack() function modifies the tag stack of a window.
tagstack-examples
Write to the tag stack just like :tag but with a user-defined
jumper#jump_to_tag function:
" Store where we're jumping from before we jump.
let tag = expand('<cword>')
let pos = [bufnr()] + getcurpos()[1:]
let item = {'bufnr': pos[0], 'from': pos, 'tagname': tag}
if jumper#jump_to_tag(tag)
" Jump was successful, write previous location to tag stack.
let winid = win_getid()
let stack = gettagstack(winid)
let stack['items'] = [item]
call settagstack(winid, stack, 't')
endif
let tag = expand('<cword>')
let pos = [bufnr()] + getcurpos()[1:]
let item = {'bufnr': pos[0], 'from': pos, 'tagname': tag}
if jumper#jump_to_tag(tag)
" Jump was successful, write previous location to tag stack.
let winid = win_getid()
let stack = gettagstack(winid)
let stack['items'] = [item]
call settagstack(winid, stack, 't')
endif
Set current index of the tag stack to 4:
call settagstack(1005, {'curidx' : 4})
Push a new item onto the tag stack:
let pos = [bufnr('myfile.txt'), 10, 1, 0]
let newtag = [{'tagname' : 'mytag', 'from' : pos}]
call settagstack(2, {'items' : newtag}, 'a')
let newtag = [{'tagname' : 'mytag', 'from' : pos}]
call settagstack(2, {'items' : newtag}, 'a')
E73
When you try to use the tag stack while it doesn't contain anything you will
get an error message.
==============================================================================
3. Tag match list tag-matchlist E427 E428
When there are several matching tags, these commands can be used to jump
between them. Note that these commands don't change the tag stack, they keep
the same entry.
:ts :tselect
:ts[elect][!] [name] List the tags that match [name], using the
information in the tags file(s).
When [name] is not given, the last tag name from the
tag stack is used.
See tag-! for [!].
With a '>' in the first column is indicated which is
the current position in the list (if there is one).
[name] can be a regexp pattern, see tag-regexp.
See tag-priority for the priorities used in the
listing.
Example output:
# pri kind tag file
1 F f mch_delay os_amiga.c
mch_delay(msec, ignoreinput)
> 2 F f mch_delay os_msdos.c
mch_delay(msec, ignoreinput)
3 F f mch_delay os_unix.c
mch_delay(msec, ignoreinput)
Type number and <Enter> (empty cancels):
1 F f mch_delay os_amiga.c
mch_delay(msec, ignoreinput)
> 2 F f mch_delay os_msdos.c
mch_delay(msec, ignoreinput)
3 F f mch_delay os_unix.c
mch_delay(msec, ignoreinput)
Type number and <Enter> (empty cancels):
See tag-priority for the "pri" column. Note that
this depends on the current file, thus using
":tselect xxx" can produce different results.
The "kind" column gives the kind of tag, if this was
included in the tags file.
The "info" column shows information that could be
found in the tags file. It depends on the program
that produced the tags file.
When the list is long, you may get the more-prompt.
If you already see the tag you want to use, you can
type 'q' and enter the number.
:sts :stselect
:sts[elect][!] [name] Does ":tselect[!] [name]" and splits the window for
the selected tag.
g]
g] Like CTRL-], but use ":tselect" instead of ":tag".
v_g]
{Visual}g] Same as "g]", but use the highlighted text as the
identifier.
:tj :tjump
:tj[ump][!] [name] Like ":tselect", but jump to the tag directly when
there is only one match.
:stj :stjump
:stj[ump][!] [name] Does ":tjump[!] [name]" and splits the window for the
selected tag.
g_CTRL-]
g CTRL-] Like CTRL-], but use ":tjump" instead of ":tag".
v_g_CTRL-]
{Visual}g CTRL-] Same as "g CTRL-]", but use the highlighted text as
the identifier.
:tn :tnext
:[count]tn[ext][!] Jump to [count] next matching tag (default 1). See
tag-! for [!].
:tp :tprevious
:[count]tp[revious][!] Jump to [count] previous matching tag (default 1).
See tag-! for [!].
:tN :tNext
:[count]tN[ext][!] Same as ":tprevious".
:tr :trewind
:[count]tr[ewind][!] Jump to first matching tag. If [count] is given, jump
to [count]th matching tag. See tag-! for [!].
:tf :tfirst
:[count]tf[irst][!] Same as ":trewind".
:tl :tlast
:tl[ast][!] Jump to last matching tag. See tag-! for [!].
:lt :ltag
:lt[ag][!] [name] Jump to tag [name] and add the matching tags to a new
location list for the current window. [name] can be
a regexp pattern, see tag-regexp. When [name] is
not given, the last tag name from the tag stack is
used. The search pattern to locate the tag line is
prefixed with "\V" to escape all the special
characters (very nomagic). The location list showing
the matching tags is independent of the tag stack.
See tag-! for [!].
When there is no other message, Vim shows which matching tag has been jumped
to, and the number of matching tags:
tag 1 of 3 or more
The " or more" is used to indicate that Vim didn't try all the tags files yet.When using ":tnext" a few times, or with ":tlast", more matches may be found.
When you didn't see this message because of some other message, or you just
want to know where you are, this command will show it again (and jump to the
same tag as last time):
:0tn
tag-skip-file
When a matching tag is found for which the file doesn't exist, this match is
skipped and the next matching tag is used. Vim reports this, to notify you of
missing files. When the end of the list of matches has been reached, an error
message is given.
tag-preview
The tag match list can also be used in the preview window. The commands are
the same as above, with a "p" prepended.
{not available when compiled without the +quickfix feature}
:pts :ptselect
:pts[elect][!] [name] Does ":tselect[!] [name]" and shows the new tag in a
"Preview" window. See :ptag for more info.
:ptj :ptjump
:ptj[ump][!] [name] Does ":tjump[!] [name]" and shows the new tag in a
"Preview" window. See :ptag for more info.
:ptn :ptnext
:[count]ptn[ext][!] ":tnext" in the preview window. See :ptag.
:ptp :ptprevious
:[count]ptp[revious][!] ":tprevious" in the preview window. See :ptag.
:ptN :ptNext
:[count]ptN[ext][!] Same as ":ptprevious".
:ptr :ptrewind
:[count]ptr[ewind][!] ":trewind" in the preview window. See :ptag.
:ptf :ptfirst
:[count]ptf[irst][!] Same as ":ptrewind".
:ptl :ptlast
:ptl[ast][!] ":tlast" in the preview window. See :ptag.
==============================================================================
4. Tags details tag-details
static-tag
A static tag is a tag that is defined for a specific file. In a C program
this could be a static function.
In Vi jumping to a tag sets the current search pattern. This means that the
"n" command after jumping to a tag does not search for the same pattern that
it did before jumping to the tag. Vim does not do this as we consider it to
be a bug. If you really want the old Vi behavior, set the 't' flag in
'cpoptions'.
tag-binary-search
Vim uses binary searching in the tags file to find the desired tag quickly
(when enabled at compile time +tag_binary). But this only works if the
tags file was sorted on ASCII byte value. Therefore, if no match was found,
another try is done with a linear search. If you only want the linear search,
reset the 'tagbsearch' option. Or better: Sort the tags file!
Note that the binary searching is disabled when not looking for a tag with a
specific name. This happens when ignoring case and when a regular expression
is used that doesn't start with a fixed string. Tag searching can be a lot
slower then. The former can be avoided by case-fold sorting the tags file.
See 'tagbsearch' for details.
tag-regexp
The ":tag" and ":tselect" commands accept a regular expression argument. See
pattern for the special characters that can be used.
When the argument starts with '/', it is used as a pattern. If the argument
does not start with '/', it is taken literally, as a full tag name.
Examples:
:tag main
jumps to the tag "main" that has the highest priority. :tag /^get
jumps to the tag that starts with "get" and has the highest priority. :tag /norm
lists all the tags that contain "norm", including "id_norm".When the argument both exists literally, and match when used as a regexp, a
literal match has a higher priority. For example, ":tag /open" matches "open"
before "open_file" and "file_open".
When using a pattern case is ignored. If you want to match case use "\C" in
the pattern.
tag-!
If the tag is in the current file this will always work. Otherwise the
performed actions depend on whether the current file was changed, whether a !
is added to the command and on the 'autowrite' and 'winfixbuf' options:
tag in file winfixbuf autowrite
current file changed ! option option action
-----------------------------------------------------------------------------
yes x x off x goto tag
no no x off x read other file, goto tag
no yes yes off x abandon current file,
read other file, goto tag
no yes no off on write current file,
read other file, goto tag
no yes no off off fail
yes x yes x x goto tag
no no no on x fail
no yes no on x fail
no yes no on on fail
no yes no on off fail
-----------------------------------------------------------------------------
- If the tag is in the current file, the command will always work.
- If the tag is in another file and the current file was not changed, the
other file will be made the current file and read into the buffer.
- If the tag is in another file, the current file was changed and a ! is
added to the command, the changes to the current file are lost, the other
file will be made the current file and read into the buffer.
- If the tag is in another file, the current file was changed and the
'autowrite' option is on, the current file will be written, the other
file will be made the current file and read into the buffer.
- If the tag is in another file, the current file was changed and the
'autowrite' option is off, the command will fail. If you want to save
the changes, use the ":w" command and then use ":tag" without an argument.
This works because the tag is put on the stack anyway. If you want to lose
the changes you can use the ":tag!" command.
- If the tag is in another file and the window includes 'winfixbuf', the
command will fail. If the tag is in the same file then it may succeed.
tag-security
Note that Vim forbids some commands, for security reasons. This works like
using the 'secure' option for exrc/vimrc files in the current directory. See
trojan-horse and sandbox.
When the {tagaddress} changes a buffer, you will get a warning message:
"WARNING: tag command changed a buffer!!!"
In a future version changing the buffer will be impossible. All this for
security reasons: Somebody might hide a nasty command in the tags file, which
would otherwise go unnoticed. Example:
:$d|/tag-function-name/
In Vi the ":tag" command sets the last search pattern when the tag is searched
for. In Vim this is not done, the previous search pattern is still remembered,
unless the 't' flag is present in 'cpoptions'.
emacs-tags emacs_tags E430
Emacs style tag files are only supported if Vim was compiled with the
+emacs_tags feature enabled. Sorry, there is no explanation about Emacs tag
files here, it is only supported for backwards compatibility :-).
Lines in Emacs tags files can be very long. Vim only deals with lines of up
to about 510 bytes. To see whether lines are ignored set 'verbose' to 5 or
higher. Non-Emacs tags file lines can be any length.
tags-option
The 'tags' option is a list of file names. Each of these files is searched
for the tag. This can be used to use a different tags file than the default
file "tags". It can also be used to access a common tags file.
The next file in the list is not used when:
- A matching static tag for the current buffer has been found.
- A matching global tag has been found.
This also depends on whether case is ignored. Case is ignored when:
- 'tagcase' is "followic" and 'ignorecase' is set
- 'tagcase' is "ignore"
- 'tagcase' is "smart" and the pattern only contains lower case
characters.
- 'tagcase' is "followscs" and 'smartcase' is set and the pattern only
contains lower case characters.
If case is not ignored, and the tags file only has a match without matching
case, the next tags file is searched for a match with matching case. If no
tag with matching case is found, the first match without matching case is
used. If case is ignored, and a matching global tag with or without matching
case is found, this one is used, no further tags files are searched.
When a tag file name starts with "./", the '.' is replaced with the path of
the current file. This makes it possible to use a tags file in the directory
where the current file is (no matter what the current directory is). The idea
of using "./" is that you can define which tag file is searched first: In the
current directory ("tags,./tags") or in the directory of the current file
("./tags,tags").
For example:
:set tags=./tags,tags,/home/user/commontags
In this example the tag will first be searched for in the file "tags" in the
directory where the current file is. Next the "tags" file in the current
directory. If it is not found there, then the file "/home/user/commontags"
will be searched for the tag.
This can be switched off by including the 'd' flag in 'cpoptions', to make
it Vi compatible. "./tags" will then be the tags file in the current
directory, instead of the tags file in the directory where the current file
is.
Instead of the comma a space may be used. Then a backslash is required for
the space to be included in the string option:
:set tags=tags\ /home/user/commontags
To include a space in a file name use three backslashes. To include a comma
in a file name use two backslashes. For example, use:
:set tags=tag\\\ file,/home/user/common\\,tags
for the files "tag file" and "/home/user/common,tags". The 'tags' option will
have the value "tag\ file,/home/user/common\,tags".
If the 'tagrelative' option is on (which is the default) and using a tag file
in another directory, file names in that tag file are relative to the
directory where the tag file is.
==============================================================================
5. Tags file format tags-file-format E431
ctags jtags
A tags file can be created with an external command, for example "ctags". It
will contain a tag for each function. Some versions of "ctags" will also make
a tag for each "#defined" macro, typedefs, enums, etc.
Some programs that generate tags files:
ctags As found on most Unix systems. Only supports C. Only
does the basic work.
universal ctags A maintained version of ctags based on exuberant
ctags. See https://ctags.io.
Exuberant_ctags
exuberant ctags This is a very good one. It works for C, C++, Java,
Fortran, Eiffel and others. It can generate tags for
many items. See http://ctags.sourceforge.net.
No new version since 2009.
etags Connected to Emacs. Supports many languages.
JTags For Java, in Java. It can be found at
http://www.fleiner.com/jtags/.
ptags.py For Python, in Python. Found in your Python source
directory at Tools/scripts/ptags.py.
ptags For Perl, in Perl. It can be found at (link seems
dead):
http://www.eleves.ens.fr:8080/home/nthiery/Tags/.
gnatxref For Ada. See http://www.gnuada.org/. gnatxref is
part of the gnat package.
The lines in the tags file must have one of these two formats:
1. {tagname} {TAB} {tagfile} {TAB} {tagaddress}
2. {tagname} {TAB} {tagfile} {TAB} {tagaddress} {term} {field} ..
Previously an old format was supported, see tag-old-static.
The first format is a normal tag, which is completely compatible with Vi. It
is the only format produced by traditional ctags implementations. This is
often used for functions that are global, also referenced in other files.
The lines in the tags file can end in <NL> or <CR><NL>. On the Macintosh <CR>
also works. The <CR> and <NL> characters can never appear inside a line.
The second format is new. It includes additional information in optional
fields at the end of each line. It is backwards compatible with Vi. It is
only supported by new versions of ctags (such as Universal ctags or Exuberant
ctags).
{tagname} The identifier. Normally the name of a function, but it can
be any identifier. It cannot contain a <Tab>.
{TAB} One <Tab> character. Note: previous versions allowed any
white space here. This has been abandoned to allow spaces in
{tagfile}.
{tagfile} The file that contains the definition of {tagname}. It can
have an absolute or relative path. It may contain environment
variables and wildcards (although the use of wildcards is
doubtful). It cannot contain a <Tab>.
{tagaddress} The Ex command that positions the cursor on the tag. It can
be any Ex command, although restrictions apply (see
tag-security). Posix only allows line numbers and search
commands, which are mostly used.
{term} ;" The two characters semicolon and double quote. This is
interpreted by Vi as the start of a comment, which makes the
following be ignored. This is for backwards compatibility
with Vi, it ignores the following fields. Example:
APP file /^static int APP;$/;" v
When {tagaddress} is not a line number or search pattern, then{term} must be |;". Here the bar ends the command (excluding
the bar) and ;" is used to have Vi ignore the rest of the
line. Example:
APP file.c call cursor(3, 4)|;" v
{field} .. A list of optional fields. Each field has the form:
<Tab>{fieldname}:{value}
The {fieldname} identifies the field, and can only contain
alphabetical characters [a-zA-Z].
The {value} is any string, but cannot contain a <Tab>.
These characters are special:
"\t" stands for a <Tab>
"\r" stands for a <CR>
"\n" stands for a <NL>
"\\" stands for a single '\' character
There is one field that doesn't have a ':'. This is the kind
of the tag. It is handled like it was preceded with "kind:".
In the above example, this was "kind:v" (typically variable).
See the documentation of ctags for the kinds it produces, with
ctags you can use ctags --list-kinds .
The only other field currently recognized by Vim is "file:"
(with an empty value). It is used for a static tag.
The first lines in the tags file can contain lines that start with
!_TAG_
These are sorted to the first lines, only rare tags that start with "!" can
sort to before them. Vim recognizes two items. The first one is the line
that indicates if the file was sorted. When this line is found, Vim uses
binary searching for the tags file:
!_TAG_FILE_SORTED<Tab>1<Tab>{anything}
A tag file may be case-fold sorted to avoid a linear search when case is
ignored. (Case is ignored when 'ignorecase' is set and 'tagcase' is
"followic", or when 'tagcase' is "ignore".) See 'tagbsearch' for details.
The value '2' should be used then:
!_TAG_FILE_SORTED<Tab>2<Tab>{anything}
The other tag that Vim recognizes is the encoding of the tags file:
!_TAG_FILE_ENCODING<Tab>utf-8<Tab>{anything}
Here "utf-8" is the encoding used for the tags. Vim will then convert the tag
being searched for from 'encoding' to the encoding of the tags file. And when
listing tags the reverse happens. When the conversion fails the unconverted
tag is used.
tag-search
The command can be any Ex command, but often it is a search command.
Examples:
tag1 file1 /^main(argc, argv)/
tag2 file2 108
The command is always executed with 'magic' not set. The only special
characters in a search pattern are "^" (begin-of-line) and "$" (<EOL>).
See pattern. Note that you must put a backslash before each backslash in
the search text. This is for backwards compatibility with Vi.
E434 E435
If the command is a normal search command (it starts and ends with "/" or
"?"), some special handling is done:
- Searching starts on line 1 of the file.
The direction of the search is forward for "/", backward for "?".
Note that 'wrapscan' does not matter, the whole file is always searched.
- If the search fails, another try is done ignoring case. If that fails too,
a search is done for:
"^tagname[ \t]*("
(the tag with '^' prepended and "[ \t]*(" appended). When using functionnames, this will find the function name when it is in column 0. This will
help when the arguments to the function have changed since the tags file was
made. If this search also fails another search is done with:
"^[#a-zA-Z_].*\<tagname[ \t]*("
This means: A line starting with '#' or an identifier and containing the tagfollowed by white space and a '('. This will find macro names and function
names with a type prepended.
tag-old-static
Until March 2019 (patch 8.1.1092) an outdated format was supported:
{tagfile}:{tagname} {TAB} {tagfile} {TAB} {tagaddress}
This format is for a static tag only. It is obsolete now, replaced by
the second format. It is only supported by Elvis 1.x, older Vim versions and
a few versions of ctags. A static tag is often used for functions that are
local, only referenced in the file {tagfile}. Note that for the static tag,
the two occurrences of {tagfile} must be exactly the same. Also see
tags-option below, for how static tags are used.
The support was removed, since when you can update to the new Vim version you
should also be able to update ctags to one that supports the second format.
==============================================================================
6. Include file searches include-search definition-search
E387 E388 E389
These commands look for a string in the current file and in all encountered
included files (recursively). This can be used to find the definition of a
variable, function or macro. If you only want to search in the current
buffer, use the commands listed at pattern-searches.
These commands are not available when the +find_in_path feature was disabled
at compile time.
When a line is encountered that includes another file, that file is searched
before continuing in the current buffer. Files included by included files are
also searched. When an include file could not be found it is silently
ignored. Use the :checkpath command to discover which files could not be
found, possibly your 'path' option is not set up correctly. Note: the
included file is searched, not a buffer that may be editing that file. Only
for the current file the lines in the buffer are used.
The string can be any keyword or a defined macro. For the keyword any match
will be found. For defined macros only lines that match with the 'define'
option will be found. The default is "^#\s*define", which is for C programs.
For other languages you probably want to change this. See 'define' for an
example for C++. The string cannot contain an end-of-line, only matches
within a line are found.
When a match is found for a defined macro, the displaying of lines continues
with the next line when a line ends in a backslash.
The commands that start with "[" start searching from the start of the current
file. The commands that start with "]" start at the current cursor position.
The 'include' option is used to define a line that includes another file. The
default is "\^#\s*include", which is for C programs. Note: Vim does not
recognize C syntax, if the 'include' option matches a line inside
"#ifdef/#endif" or inside a comment, it is searched anyway. The 'isfname'
option is used to recognize the file name that comes after the matched
pattern.
The 'path' option is used to find the directory for the include files that
do not have an absolute path.
The 'comments' option is used for the commands that display a single line or
jump to a line. It defines patterns that may start a comment. Those lines
are ignored for the search, unless [!] is used. One exception: When the line
matches the pattern "^# *define" it is not considered to be a comment.
If you want to list matches, and then select one to jump to, you could use a
mapping to do that for you. Here is an example:
:map <F4> [I:let nr = input("Which one: ")<Bar>exe "normal " .. nr .. "[\t"<CR>
[i
[i Display the first line that contains the keyword
under the cursor. The search starts at the beginning
of the file. Lines that look like a comment are
ignored (see 'comments' option). If a count is given,
the count'th matching line is displayed, and comment
lines are not ignored.
]i
]i like "[i", but start at the current cursor position.
:is :isearch
:[range]is[earch][!] [count] [/]pattern[/]
Like "[i" and "]i", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
[I
[I Display all lines that contain the keyword under the
cursor. Filenames and line numbers are displayed
for the found lines. The search starts at the
beginning of the file.
]I
]I like "[I", but start at the current cursor position.
:il :ilist
:[range]il[ist][!] [/]pattern[/]
Like "[I" and "]I", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
[_CTRL-I
[ CTRL-I Jump to the first line that contains the keyword
under the cursor. The search starts at the beginning
of the file. Lines that look like a comment are
ignored (see 'comments' option). If a count is given,
the count'th matching line is jumped to, and comment
lines are not ignored.
]_CTRL-I
] CTRL-I like "[ CTRL-I", but start at the current cursor
position.
:ij :ijump
:[range]ij[ump][!] [count] [/]pattern[/]
Like "[ CTRL-I" and "] CTRL-I", but search in
[range] lines (default: whole file).
See :search-args for [/] and [!].
CTRL-W CTRL-I CTRL-W_CTRL-I CTRL-W_i
CTRL-W i Open a new window, with the cursor on the first line
that contains the keyword under the cursor. The
search starts at the beginning of the file. Lines
that look like a comment line are ignored (see
'comments' option). If a count is given, the count'th
matching line is jumped to, and comment lines are not
ignored.
:isp :isplit
:[range]isp[lit][!] [count] [/]pattern[/]
Like "CTRL-W i" and "CTRL-W i", but search in
[range] lines (default: whole file).
See :search-args for [/] and [!].
[d
[d Display the first macro definition that contains the
macro under the cursor. The search starts from the
beginning of the file. If a count is given, the
count'th matching line is displayed.
]d
]d like "[d", but start at the current cursor position.
:ds :dsearch
:[range]ds[earch][!] [count] [/]string[/]
Like "[d" and "]d", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
[D
[D Display all macro definitions that contain the macro
under the cursor. Filenames and line numbers are
displayed for the found lines. The search starts
from the beginning of the file.
]D
]D like "[D", but start at the current cursor position.
:dli :dlist
:[range]dli[st][!] [/]string[/]
Like [D and ]D, but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
Note that :dl works like :delete with the "l"
flag, not :dlist.
[_CTRL-D
[ CTRL-D Jump to the first macro definition that contains the
keyword under the cursor. The search starts from
the beginning of the file. If a count is given, the
count'th matching line is jumped to.
]_CTRL-D
] CTRL-D like "[ CTRL-D", but start at the current cursor
position.
:dj :djump
:[range]dj[ump][!] [count] [/]string[/]
Like "[ CTRL-D" and "] CTRL-D", but search in
[range] lines (default: whole file).
See :search-args for [/] and [!].
CTRL-W CTRL-D CTRL-W_CTRL-D CTRL-W_d
CTRL-W d Open a new window, with the cursor on the first
macro definition line that contains the keyword
under the cursor. The search starts from the
beginning of the file. If a count is given, the
count'th matching line is jumped to.
:dsp :dsplit
:[range]dsp[lit][!] [count] [/]string[/]
Like "CTRL-W d", but search in [range] lines
(default: whole file).
See :search-args for [/] and [!].
:che :chec :check :checkpath
:che[ckpath] List all the included files that could not be found.
:che[ckpath]! List all the included files.
:search-args
Common arguments for the commands above:
[!] When included, find matches in lines that are recognized as comments.
When excluded, a match is ignored when the line is recognized as a
comment (according to 'comments'), or the match is in a C comment
(after "//" or inside /* */). Note that a match may be missed if a
line is recognized as a comment, but the comment ends halfway the line.
And if the line is a comment, but it is not recognized (according to
'comments') a match may be found in it anyway. Example:
/* comment
foobar */
A match for "foobar" is found, because this line is not recognized asfoobar */
a comment (even though syntax highlighting does recognize it).
Note: Since a macro definition mostly doesn't look like a comment, the
[!] makes no difference for ":dlist", ":dsearch" and ":djump".
[/] A pattern can be surrounded by '/'. Without '/' only whole words are
matched, using the pattern "\<pattern\>". Only after the second '/' a
next command can be appended with '|'. Example:
:isearch /string/ | echo "the last one"
For a ":djump", ":dsplit", ":dlist" and ":dsearch" command the patternis used as a literal string, not as a search pattern.
==============================================================================
7. Using 'tagfunc' tag-function
It is possible to provide Vim with a function which will generate a list of
tags used for commands like :tag, :tselect and Normal mode tag commands
like CTRL-].
The function used for generating the taglist is specified by setting the
'tagfunc' option. The function will be called with three arguments:
pattern The tag identifier or pattern used during the tag search.
flags String containing flags to control the function behavior.
info Dict containing the following entries:
buf_ffname Full filename which can be used for priority.
user_data Custom data String, if stored in the tag
stack previously by tagfunc.
Note that in a legacy function "a:" needs to be prepended to the argument name
when using it.
Currently up to three flags may be passed to the tag function:
'c' The function was invoked by a normal command being processed
(mnemonic: the tag function may use the context around the
cursor to perform a better job of generating the tag list.)
'i' In Insert mode, the user was completing a tag (with
i_CTRL-X_CTRL-] or 'completeopt' contains t).
'r' The first argument to tagfunc should be interpreted as a
pattern (see tag-regexp), such as when using:
:tag /pat
It is also given when completing in insert mode.If this flag is not present, the argument is usually taken
literally as the full tag name.
Note that when 'tagfunc' is set, the priority of the tags described in
tag-priority does not apply. Instead, the priority is exactly as the
ordering of the elements in the list returned by the function.
E987
The function should return a List of Dict entries. Each Dict must at least
include the following entries and each value must be a string:
name Name of the tag.
filename Name of the file where the tag is defined. It is
either relative to the current directory or a full path.
cmd Ex command used to locate the tag in the file. This
can be either an Ex search pattern or a line number.
Note that the format is similar to that of taglist(), which makes it possible
to use its output to generate the result.
The following fields are optional:
kind Type of the tag.
user_data String of custom data stored in the tag stack which
can be used to disambiguate tags between operations.
If the function returns v:null instead of a List, a standard tag lookup will
be performed instead.
It is not allowed to change the tagstack from inside 'tagfunc'. E986
It is not allowed to close a window or change window from inside 'tagfunc'.
E1299
The following is a hypothetical example of a function used for 'tagfunc'. It
uses the output of taglist() to generate the result: a list of tags in the
inverse order of file names.
function TagFunc(pattern, flags, info)
function CompareFilenames(item1, item2)
let f1 = a:item1['filename']
let f2 = a:item2['filename']
return f1 >=# f2 ?
\ -1 : f1 <=# f2 ? 1 : 0
endfunction
function CompareFilenames(item1, item2)
let f1 = a:item1['filename']
let f2 = a:item2['filename']
return f1 >=# f2 ?
\ -1 : f1 <=# f2 ? 1 : 0
endfunction
let result = taglist(a:pattern)
call sort(result, "CompareFilenames")
call sort(result, "CompareFilenames")
return result
endfunc
set tagfunc=TagFunc
endfunc
set tagfunc=TagFunc
vim:tw=78:ts=8:noet:ft=help:norl: