ctrlp still searches the ignored directory - vim

Question

I tried to put ignored setting in .vimrc
But when I used the ctrlp to search under rails app folder
It still search the vendor folder, so it took lots of time.
But when the search was done, I couldn't search anything under the vendor
It was so strange! How to fix it.
Here is my .vimrc setting file.
http://d.pr/i/yMtK
http://d.pr/i/Hy4u
" Sane Ignore For ctrlp
let g:ctrlp_custom_ignore = {
\ 'dir': '\.git$|vendor\|\.hg$\|\.svn$\|\.yardoc\|public\/images\|public\/system\|data\|log\|tmp$',
\ 'file': '\.exe$\|\.so$\|\.dat$'
\ }
When I appended the code in the end of .vimrc
217 let g:NERDTreeIgnore=['\~$', 'vendor']
218 set wildignore+=*\\vendor\\**
It worked when my first time to use the CTRLP to search under the RAILS app folder,
But still NOT worked in the following times.
I guess maybe there are some settings will disable the ignored setting ？
Here are the structure of my folder
.
├── Gemfile
├── Gemfile.lock
├── README.rdoc
├── Rakefile
├── app
│   ├── assets
│   ├── controllers
│   ├── helpers
│   ├── mailers
│   ├── models
│   ├── uploaders
│   ├── views
│   └── workers
├── auto.sh
├── config
│   ├── application.rb
│   ├── application.yml
│   ├── boot.rb
│   ├── database.yml
│   ├── environment.rb
│   ├── environments
│   ├── initializers
│   ├── locales
│   ├── macbookair_whenever_schedule.rb
│   ├── menu_navigation.rb
│   ├── navigation.rb
│   ├── resque.god
│   ├── resque_schedule.yml
│   ├── routes.rb
│   ├── schedule.rb -> ubuntu_whenever_schedule.rb
│   ├── tinymce.yml
│   └── ubuntu_whenever_schedule.rb
├── config.ru
├── db
│   ├── development.sqlite3
│   ├── migrate
│   ├── migrate_should_be_skip
│   ├── production.sqlite3
│   ├── schema.rb
│   └── seeds.rb
├── doc
│   └── README_FOR_APP
├── lib
│   ├── assets
│   ├── auto_tools
│   ├── tasks
│   └── url_automation_module.rb
├── log
│   ├── apalog
│   ├── development.log
│   ├── passenger.80.log
│   ├── production.log
│   └── prodution.log
├── output_name
├── public
│   ├── 404.html
│   ├── 422.html
│   ├── 500.html
│   ├── exports
│   ├── favicon.ico
│   ├── results.zip
│   ├── robots.txt
│   ├── sandbox
│   └── uploads
├── script
│   ├── delayed_job
│   └── rails
├── test
│   ├── fixtures
│   ├── functional
│   ├── integration
│   ├── performance
│   ├── test_helper.rb
│   └── unit
├── test.sh
├── tmp
│   ├── cache
│   ├── pids
│   ├── restart.txt
│   ├── sessions
│   └── sockets
├── tmplog
└── vendor
└── bundle

Answer 1

If you type :help ctrlp-options and read a bit, you will find:
Note #1: by default, wildignore and g:ctrlp_custom_ignore only
apply when globpath() is used to scan for files, thus these options
do not apply when a command defined with g:ctrlp_user_command is
being used.
Thus, you may need to unlet g:ctrlp_user_command (possibly set to a default command) to actually use wildignore as advised by #TomCammann. For instance, in your ~/.vimrc, add:
if exists("g:ctrlp_user_command")
unlet g:ctrlp_user_command
endif
set wildignore+=*\\vendor\\**
After that, you need to refresh your ctrlp cache: in Vim, press F5 in ctrlp mode, or run :CtrlPClearAllCaches, or remove the cache directory directly in your shell:
rm -r ~/.cache/ctrlp/ # On Linux

Answer 2

part of my .vimrc file . perhaps it will help
set wildignore+=*/.git/*,*/.hg/*,*/.svn/*,*/.idea/*,*/.DS_Store,*/vendor

Answer 3

You can use the wildignore vim setting which CtrlP will pick up on.
set wildignore+=*\\vendor\\**

Answer 4

Check if you are using some specific search command, like:
let g:ctrlp_user_command = 'find %s -type f' " MacOSX/Linux
let g:ctrlp_user_command = 'dir %s /-n /b /s /a-d' " Windows
This kind of configuration ignores the g:ctrlp_custom_ignore option.

Answer 5

wildignore may be used by other commands, the reason of failure for g:ctrlp_custom_ignore is g:ctrlp_user_command, for example, here is mine:
if executable('rg')
let g:ctrlp_user_command = 'rg %s --files --hidden --color=never --glob ""'
endif
For this case, rg has own ignore way, just put .git to .gitignore, rg will not search any files in .gitignore.