Fix issue #516 #517
Fix issue #516 #517
Conversation
|
Note: I have tested this with vim + python2 and vim + python3, it works as expected with both. |
| @@ -489,6 +489,8 @@ function! jedi#complete_opened(is_popup_on_dot) | |||
| " option is set. | |||
| if g:jedi#popup_select_first && stridx(&completeopt, 'longest') > -1 | |||
| return "\<Down>" | |||
| else | |||
| return "" | |||
There was a problem hiding this comment.
However, I'm pretty sure this brings up new issues, because it doesn't respect popup_on_dot.
There was a problem hiding this comment.
@davidhalter Can you clarify how it breaks it? In my hands all combinations of popup_on_dot and popup_select_first work as expected. If you can tell me how you think it will cause an issue, I will be very happy to fix it.
There was a problem hiding this comment.
"popup_on_dot" is only used to trigger the completion (and therefore not relevant here), right?
There was a problem hiding this comment.
Oh, it's also an argument here (below) and probably for some reason.
There was a problem hiding this comment.
Have you tested it using the popup_on_dot feature and without "noinsert" or "noselect" in &completeopt?
That's when it's expected to return <c-p>, but would return "" now.
Adding a test for these combinations would be really nice btw.. :)
Make
g:jedi#popup_select_first=0work as expected by preventing theautoselection of the last item when first item not explicitly selected.
Fix for issue #516