Closed lifepillar closed 8 years ago
Interestingly, this seems a regression (that's why I hadn't noticed before…) introduced by c173570ac8dd4a4046885408f5269c32e95d11fe.
@Konfekt @tjharman I have addressed this problem. Could you please try the current master and see how it works for you now?
Thanks. Works perfectly.
Some completion methods may be skipped in some circumstances, even when they have high priority and they return (possibly, not very quickly) some results. The problem can be reproduced with line completion. Set:
then insert this text:
Enable automatic completion and type
iSo
. Line completion is triggered, as expected, but µcomplete goes on trying (not local, btw!) keyword completion.This issue affects
'tags'
,'line'
,'dict'
(see https://github.com/lifepillar/vim-mucomplete/issues/7 and https://github.com/lifepillar/vim-mucomplete/issues/9) and omni/user completion from some plugins (notably, clang_complete).The reason is that
pumvisible()
returns0
when it is expected to return1
inmucomplete#complete_chain()
. Interestingly, if I “slow down”mucomplete#complete_chain()
by puttingsleep
commands, I can see thatpumvisible()
returns0
when the popup menu is actually visible.Not sure if I have hit a bug in Vim or it is just a problem of buffering/typeahead or another timing issue.