Probably this commit causes the problem:
http://code.google.com/p/vim/source/detail?r=44ffd0a8abcdba441216514cb711649eb1169876
Vim version: 7.3.289
Looks like it only appears when "set completeopt+=longest" is set.
Steps to reproduce: it's very easy to reproduce this bug with the
option mentioned above + one of my omnifunc plugins: ccode (C/C++
completion based on clang) or gocode (Go completion). Both worked just
fine with previous version of vim + the interaction with vim is
minimal, therefore they shouldn't really cause a segfault by themselves.
Bug appears on both x86 and amd64. Plus confirmed by at least one other
person.
Yes, the motivation behind this patch looks somewhat counter-intuitive
(http://groups.google.com/group/vim_dev/browse_thread/thread/86643b0112664dfe)
Why on Earth a completion of "a" would be "foo1", "foo2", "foo3" but
"bar1", "bar2", "bar3" for "ab"?
If there really is a reason to behave like that we'd better to
implement this somehow else. E.g., the completion function could
return an extra value which says that the function wants to be invoked
every time a user types a letter.
--
Sergey Khorev
http://sites.google.com/site/khorser
Can anybody think of a good tagline I can steal?
[BTW: please use your real name on this maillist]
I'll revert the patch for now. There should be a better solution for
the original problem.
--
BEDEVERE: And that, my lord, is how we know the Earth to be banana-shaped.
"Monty Python and the Holy Grail" PYTHON (MONTY) PICTURES LTD
/// Bram Moolenaar -- Br...@Moolenaar.net -- http://www.Moolenaar.net \\\
/// sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\ an exciting new programming language -- http://www.Zimbu.org ///
\\\ help me help AIDS victims -- http://ICCF-Holland.org ///