On Oct 1, 6:30 am, jason <[email protected]> wrote:
> Hi Dominique
> I don't use omnicppcomplete-0.41, I use vim's default omnicomplete function.
> These is my .vimrc setting about omnicomplete function:
> set completeopt=menu,longest
> set omnifunc=ccomplete#Complete
> My vim version is 7.2.0 in Windows and 7.2.245 in Slackware 13.0.
> Ctags is 5.7. I just run "ctags -R ." in working dir.
> In your screenshot, you inserted a space before "sA.", which changed to
> "pb-> sA.", not "pb->sA.".
> My problem is that if I don't insert space before "sA.", just "pb->sA.",
> then put cursor behind "sA.", press C-X C-O,
> omnicomplete's completion information is members of struct tagBBB,
> but not members of struct tagAAA.
> Thanks!
> B.R.
Jason,
I see the same behavior as you. In fact, you will see the same list of
members even if you use aa or bb instead of sA. I haven't looked at it
in detail, but a quick glance at the ccomplete.vim file leads me to
suspect that it doesn't contain the recursive logic to traverse a
sequence of nested structures. I believe that for a . or ->
completion, it looks all the way back to the beginning of the WORD
(unbroken sequence of non-whitespace, including . and ->) to determine
the type whose members will be displayed in the popup. If you type...
pb->aa.bb.cc.something_silly.<C-X><C-O>
...you will still see the members of BB in the popup; i.e., only the
type of pb matters.
As Dominique pointed out, omnicppcomplete doesn't suffer from this
limitation, and it works for C and C++. Incidentally, does anyone know
whether Vissale Neang still maintains the omnicppcomplete plugin? I
haven't gotten responses to a post and a private email on several
omnicppcomplete issues...
Thanks,
Brett Stahlman
--~--~---------~--~----~------------~-------~--~----~
You received this message from the "vim_use" maillist.
For more information, visit http://www.vim.org/maillist.php
-~----------~----~----~----~------~----~------~--~---
