Often I need to search large xml mode files for the next occurrence of the word under the cursor but preferably not if it's a tag the closing tag
In the below example # is where the cursor is. Using * or # with iskeyword not including > or < will move between <Dealid> and </Dealid>.
<Deali#d>4444</Dealid>
... 50 lines
<Dealid>6666</Dealid>
... n lines
<Dealid>5643</Dealid>
I tried :set iskeyword+=< which worked fine for moving to the next match but it changed vey to include the < of the closing tag so moving to the start of 6666, for eg, and typing vey will yank 6666< rather than 6666.
Anyway of getting the best of both words or limit the iskeyword+=< to the * and # operators?
This should do what you want.
nnoremap * :set iskeyword+=<<CR>*:set iskeyword-=<<CR>
nnoremap # :set iskeyword+=<<CR>#:set iskeyword-=<<CR>
It sets changes iskeyword before executing * or # and then resets it.
Instead of changing 'iskeyword' maybe use a visual star search. This will allow you to visually select the text and then hit *. e.g. va<*
If you don't want to use a plugin you can put the following mapping in your ~/.vimrc file:
xnoremap * y/\V<c-r>=escape(@", '\')<cr><cr>
Note: this mapping clobbers the unnamed register so it would probably be best to use a more advanced mapping or a plugin.
来源：https://stackoverflow.com/questions/26007655/set-iskeyword-selectively