Ad blocker interference detected!
Wikia is a free-to-use site that makes money from advertising. We have a modified experience for viewers using ad blockers
Wikia is not accessible if you’ve made further modifications. Remove the custom ad blocker rule(s) and the page will load as expected.
created 2006 · complexity basic · author Gerald Lai · version n/a
Vim 7's popup menu for completion (omni or otherwise) takes a little getting used to when it comes to selecting the completed word. See :help ins-completion. The completed word can be selected using the arrow, PgUp/PgDn, Enter and Esc keys. In addition to that, there is Ctrl-E, Ctrl-Y, Ctrl-N and Ctrl-P.
Here's an alternative setup that feels a little more like the completion menu of other IDEs
inoremap <expr> <Esc> pumvisible() ? "\<C-e>" : "\<Esc>" inoremap <expr> <CR> pumvisible() ? "\<C-y>" : "\<CR>" inoremap <expr> <Down> pumvisible() ? "\<C-n>" : "\<Down>" inoremap <expr> <Up> pumvisible() ? "\<C-p>" : "\<Up>" inoremap <expr> <PageDown> pumvisible() ? "\<PageDown>\<C-p>\<C-n>" : "\<PageDown>" inoremap <expr> <PageUp> pumvisible() ? "\<PageUp>\<C-p>\<C-n>" : "\<PageUp>"
Use the Up/Down arrow and PgUp/PgDn keys to scroll through the popup menu. The word in text will change as you scroll through. Press Enter to accept the selected word. Hit Esc to cancel completion and go back to the original word.
Those who prefer Esc to work like Enter can use this
inoremap <silent> <Esc> <C-r>=pumvisible() ? "\<C-y>" : "\<Esc>"<CR>
Word to the wise, mapping over <Esc> can cause an issue in terminal vim where arrow keys will print letters A through D along with a newline rather than moving. To avoid the issue, simply omit the mapping over <Esc>.
For scrolling using PgUp/PgDn, one can also map Ctrl-u and Ctrl-d to PgUp and PgDn respectively
inoremap <expr> <C-d> pumvisible() ? "\<PageDown>\<C-p>\<C-n>" : "\<C-d>" inoremap <expr> <C-u> pumvisible() ? "\<PageUp>\<C-p>\<C-n>" : "\<C-u>"
The following option inserts only the longest common text of the matches. :help 'completeopt'
- Fixed mappings. Replaced original <C-r> mappings with the <expr> mappings in the comments. Merged the comments with the tip, and removed some that didn't belong.
- I think the name of this tip should be changed to something more suitable.
- --Spiiph 14:41, 18 January 2009 (UTC)
- Thanks. Any suggestions for a better title? Do you not like "Improve"? --JohnBeckett 02:51, 19 January 2009 (UTC)
- I actually like the behavior of <Up> and <Down> but using <C-P> and <C-N> is more convenient, so prefer to map <C-N> to <Down> etc. rather than the other way. However, for some reason, imaps on <C-N> and <C-P> get ignored when omni-completion popup is visible (but work fine for other completion popups). I asked a question about this long back in vim (e.g., see http://markmail.org/thread/rz6bpanuwneavfym) and didn't get a conclusive response. Anyone knows how to workaround this issue?
- --Haridsv 01:34, 12 March 2009 (UTC)
- Mappings only work in the generic 'complete' completion (<C-N>/<C-P>), not in any completion that is started with <C-X>. This seems to be a design decision. Cp. the notes below :help complete_CTRL-Y.
- To work around this (apart from patching Vim itself), you would need to break out of <C-X> mode. The only way I can think of (and I haven't tried this), is to append a
<C-R>=MyInterceptor()<CR>mapping behind every mapping that can enter completion mode (i.e. all <C-X><C-...> mappings:
inoremap <C-X><C-U> <C-X><C-U><C-R>=MyInterceptor()<CR>). Inside the MyInterceptor() function, use :help inputsave() and :help getchar() to query the completion key from the user. Then, you'll be able to translate <C-N> to <Down>, etc., and return any other keys unfiltered.
- I was thinking of defaulting to <Down> instead of <C-N>, too, but then stuck with the default mappings, which are useful, too. (And the auto-insertion of matches sometimes saves one from explicitly acknowledging the match, thus saving one keystroke.) -- Inkarkat 13:31, 16 June 2009 (UTC)
- I had some issues with the <Esc> mapping given here, and some github bugs back up the idea that it is not uncommon, so I added a caveat about it to the article. 18.104.22.168 18:15, May 2, 2013 (UTC)