I've happily used count prefixes on search commands in vim for probably a couple of decades, so I know they used to count correctly.
5n was equivalent to
n 5 times.
As of some recent version,
5n is sometimes equivalent to
n 5 times, and sometimes equivalent to
n 6 times.
I can no longer type
5n, see the cursor stay in the same place while the "search wrapped" message appears at the bottom of the screen, and know that there are 5 whatevers in the file. Which sounds really silly when I try to explain it, but it's apparently something I was doing subconsciously all the time, so the new version is driving me crazy.
What could be causing the search repeat count to behave inconsistently, and how can I fix it?
About my environment: The active options as reported by
autoindent expandtab ruler ttymouse=xterm background=dark fileformats= shiftwidth=2 viminfo= backspace=1 nofsync swapsync= wrapmargin=3 cpoptions=BceFs$t nojoinspaces notextauto noequalalways noloadplugins tildeop noesckeys nomore ttyfast completeopt=preview formatoptions=tcroq2b path=.,/usr/include,,,/usr/include/x86_64-linux-gnu printoptions=paper:letter runtimepath=~/.vim,/var/lib/vim/addons,/usr/share/vim/vimfiles,/usr/share/vim/ vim80,/usr/share/vim/vimfiles/after,/var/lib/vim/addons/after,~/.vim/after suffixes=.bak,~,.swp,.o,.info,.aux,.log,.dvi,.bbl,.blg,.brf,.cb,.ind,.idx,.ilg ,.inx,.out,.toc
Some of those settings come from the Debian packaging. The rest are my preferences. I don't do scripting, plugins, or any other fancy stuff.
ndoes have logic to repeat count+1 times (see
:h n) but the circumstances in which it happens is somewhat rare groups.google.com/forum/#!topic/vim_dev/cOjJrIk3pZg . Note that this change was made in early 2015.
cpoptionsflags (e.g. 'c') can change search fairly dramatically in some cases. Is it possible those flags changed recently? (Suggestion from Mass seems more relevant but leave no stone unturned...)