The name of the parser: any one
The command line you used to run ctags:
$ ctags --options=NONE -R -x
The content of input file:
$ git clone https://github.com/liuchengxu/vista.vim
$ cd vista.vim
$ ctags --options=NONE -R -x
The tags output you are not satisfied with:
I have no idea how to parse the output of ctags -R -x
. I want to filter the tagname and scope from the output of ctags -R -x
, but don't know how to do properly.
The tags output you expect:
It's viable if there is also a specific format for ctags -R
:
{tagname}<Tab>{tagfile}<Tab>{tagaddress}[;"<Tab>{tagfield}..]
The version of ctags:
$ ctags --version
Universal Ctags 0.0.0(1208e45), Copyright (C) 2015 Universal Ctags Team
Universal Ctags is derived from Exuberant Ctags.
Exuberant Ctags 5.8, Copyright (C) 1996-2009 Darren Hiebert
Compiled: Mar 11 2019, 17:01:27
URL: https://ctags.io/
Optional compiled features: +wildcards, +regex, +iconv, +option-directory, +xpath, +case-insensitive-filenames, +packcc
How do you get ctags binary:
brew install ctags
on macOS
I don't understand your question.
Are you satisfied with the output of ctags -R
? If yes, why don't you use -x
option?
I would like to know the background of the use case more to answer your question.
I don't know vim script. However, as far as reading vim.c, the vim parser of ctags, it doesn't fill scope fields. If you think it is a bug, tell us the vim parser fills scope field for what kind of input.
Sorry for not making it clear.
why use
-x
option?
Although ctags -R
would meet my need, I don't want to the generate the tag file, so I use the -x
option.
my use case.
Currently vista.vim only supports show the tags in some file, but some users would like to show all the tags in a project(https://github.com/liuchengxu/vista.vim/issues/14). I tried to add ctags -R
support, but don't know how to extract these output correctly.
In a word, how to extract vista#util#Trim
, function
, 18
, autoload/vista/util.vim
, vista#util#Trim(str)
respectively?
ctags -R -x
:
...
vista#util#Trim function 18 autoload/vista/util.vim function! vista#util#Trim(str)
vista#util#Truncate function 13 autoload/vista/util.vim function! vista#util#Truncate(msg) abort
vista#util#Warning function 90 autoload/vista/util.vim function! vista#util#Warning(msg) abort
vista#viewer#Display function 90 autoload/vista/viewer.vim function! vista#viewer#Display(data) abort
vista#viewer#Render function 86 autoload/vista/viewer.vim function! vista#viewer#Render(data) abort
vista#viewer#prefixes function 94 autoload/vista/viewer.vim function! vista#viewer#prefixes() abort
With -o -
option, ctags emits the output to stdout like:
[yamato@slave]/tmp/vista.vim% u-ctags -R -o - | head
/ ftplugin/vista.vim /^nnoremap <buffer> <silent> \/ :<c-u>call vista#finder#fzf#Run()<CR>$/;" m
<CR> ftplugin/vista.vim /^nnoremap <buffer> <silent> <CR> :<c-u>call vista#cursor#FoldOrJump()<CR>$/;" m
Commands README.md /^### Commands$/;" S
Contributing README.md /^## Contributing$/;" s
Features README.md /^## Features$/;" s
Installation README.md /^## Installation$/;" s
Introduction README.md /^## Introduction$/;" s
License README.md /^## License$/;" s
NeoVim README.md /^#### NeoVim$/;" t
Options README.md /^### Options$/;" S
You can customize the output with --_xformat=
option:
u-ctags -R -x --_xformat='TAGNAME:%N @@@ KIND:%K !!! LINE:%n +++ INPUT-FILE:%F === PATTERN:%P' | head
TAGNAME:/ @@@ KIND:map !!! LINE:38 +++ INPUT-FILE:ftplugin/vista.vim === PATTERN:/^nnoremap <buffer> <silent> \/ :<c-u>call vista#finder#fzf#Run()<CR>$/
TAGNAME:<CR> @@@ KIND:map !!! LINE:37 +++ INPUT-FILE:ftplugin/vista.vim === PATTERN:/^nnoremap <buffer> <silent> <CR> :<c-u>call vista#cursor#FoldOrJump()<CR>$/
TAGNAME:Commands @@@ KIND:subsection !!! LINE:101 +++ INPUT-FILE:README.md === PATTERN:/^### Commands$/
TAGNAME:Contributing @@@ KIND:section !!! LINE:164 +++ INPUT-FILE:README.md === PATTERN:/^## Contributing$/
TAGNAME:Features @@@ KIND:section !!! LINE:37 +++ INPUT-FILE:README.md === PATTERN:/^## Features$/
TAGNAME:Installation @@@ KIND:section !!! LINE:69 +++ INPUT-FILE:README.md === PATTERN:/^## Installation$/
TAGNAME:Introduction @@@ KIND:section !!! LINE:25 +++ INPUT-FILE:README.md === PATTERN:/^## Introduction$/
TAGNAME:License @@@ KIND:section !!! LINE:168 +++ INPUT-FILE:README.md === PATTERN:/^## License$/
TAGNAME:NeoVim @@@ KIND:subsubsection !!! LINE:90 +++ INPUT-FILE:README.md === PATTERN:/^#### NeoVim$/
TAGNAME:Options @@@ KIND:subsection !!! LINE:113 +++ INPUT-FILE:README.md === PATTERN:/^### Options$/
See http://docs.ctags.io/en/latest/news.html?highlight=--_xformat#customizing-xref-output .
If you don't want to write a parser, --output-format=json
can be a choice:
u-ctags -R --output-format=json | head
{"_type": "tag", "name": "/", "path": "ftplugin/vista.vim", "pattern": "/^nnoremap <buffer> <silent> \\/ :<c-u>call vista#finder#fzf#Run()<CR>$/", "kind": "map"}
{"_type": "tag", "name": "<CR>", "path": "ftplugin/vista.vim", "pattern": "/^nnoremap <buffer> <silent> <CR> :<c-u>call vista#cursor#FoldOrJump()<CR>$/", "kind": "map"}
{"_type": "tag", "name": "Commands", "path": "README.md", "pattern": "/^### Commands$/", "kind": "subsection"}
{"_type": "tag", "name": "Contributing", "path": "README.md", "pattern": "/^## Contributing$/", "kind": "section"}
{"_type": "tag", "name": "Features", "path": "README.md", "pattern": "/^## Features$/", "kind": "section"}
{"_type": "tag", "name": "Installation", "path": "README.md", "pattern": "/^## Installation$/", "kind": "section"}
{"_type": "tag", "name": "Introduction", "path": "README.md", "pattern": "/^## Introduction$/", "kind": "section"}
{"_type": "tag", "name": "License", "path": "README.md", "pattern": "/^## License$/", "kind": "section"}
{"_type": "tag", "name": "NeoVim", "path": "README.md", "pattern": "/^#### NeoVim$/", "kind": "subsubsection"}
{"_type": "tag", "name": "Options", "path": "README.md", "pattern": "/^### Options$/", "kind": "subsection"}
Thank you! I would try the --_xformat=
option, since ctags
installed by brew doesn't support --output-format=json
by default, whether brew install ctags
or brew install --HEAD universal-ctags/universal-ctags/universal-ctags
.
@KazuakiM, isn't it possile to link ctags to libjansson.so.4?
u-ctags has json output mode. It is available if u-ctags is linked to libjansson.
Tell me if I should report this as a new issue on homebrew-universal-ctags repo.
@liuchengxu, tags file, the native format of ctags can be parserd with functions in readtags.c.
readtags command and python-ctags3, a python library uses readtags.c.
@masatake
Hi, homebrew-universal-ctags supportes jansson.
$ brew info universal-ctags
universal-ctags/universal-ctags/universal-ctags: HEAD
Maintained ctags implementation
https://github.com/universal-ctags/ctags
Conflicts with:
ctags (because this formula installs the same executable as the ctags formula)
/usr/local/Cellar/universal-ctags/HEAD-1b747b3 (325 files, 4.3MB) *
Built from source on 2019-02-02 at 16:45:17
From: https://github.com/universal-ctags/homebrew-universal-ctags/blob/master/universal-ctags.rb
==> Dependencies
Build: autoconf ✔, automake ✔, pkg-config ✔
Optional: jansson ✔, libyaml ✔
==> Options
--with-jansson
Build with jansson support
--with-libyaml
Build with libyaml support
--without-doc
Compile without man pages
--without-xml
Compile without libxml2
--HEAD
Install HEAD version
==> Caveats
Under some circumstances, emacs and ctags can conflict. By default,
emacs provides an executable `ctags` that would conflict with the
executable of the same name that ctags provides. To prevent this,
Homebrew removes the emacs `ctags` and its manpage before linking.
However, if you install emacs with the `--keep-ctags` option, then
the `ctags` emacs provides will not be removed. In that case, you
won't be able to install ctags successfully. It will build but not
link.
@KazuakiM Good catch, I missed the help. JSON support needs the option --with-jansson
.
$ brew reinstall --with-jansson universal-ctags/universal-ctags/universal-ctags