files
under "/usr/share/scripts/" are all "ruby" files, but don't have the expected
file name extension.  Adding this to the example above: >

	augroup filetypedetect
	au BufNewFile,BufRead *.xyz			setf xyz
	au BufNewFile,BufRead /usr/share/scripts/*	setf ruby
	augroup END

However, if you now edit a file /usr/share/scripts/README.txt, this is not a
ruby file.  The danger of a pattern ending in "*" is that it quickly matches
too many files.  To avoid trouble with this, put the filetype.vim file in
another directory, one that is at the end of 'runtimepath'.  For Unix for
example, you could use "~/.config/nvim/after/filetype.vim".
   You now put the detection of text files in ~/.config/nvim/filetype.vim: >

	augroup filetypedetect
	au BufNewFile,BufRead *.txt			setf text
	augroup END

That file is found in 'runtimepath' first.  Then use this in
~/.config/nvim/after/filetype.vim, which is found last: >

	augroup filetypedetect
	au BufNewFile,BufRead /usr/share/scripts/*	setf ruby
	augroup END

What will happen now is that Vim searches for "filetype.vim" files in each
directory in 'runtimepath'.  First ~/.config/nvim/filetype.vim is found.  The
autocommand to catch `*.txt` files is defined there.  Then Vim finds the
filetype.vim file in $VIMRUNTIME, which is halfway 'runtimepath'.  Finally
~/.config/nvim/after/filetype.vim is found and the autocommand for detecting
ruby files in /usr/share/scripts is added.
   When you now edit /usr/share/scripts/README.txt, the autocommands are
checked in the order in which they were defined.  The `*.txt` pattern matches,
thus "setf text" is executed to set the filetype to "text".  The pattern for
ruby matches too, and the "setf ruby" is executed.  But since 'filetype' was
already set to "text", nothing happens here.
   When you edit the file /usr/share/scripts/foobar the same autocommands are
checked.  Only the one for ruby matches and "setf ruby" sets 'filetype' to
ruby.


RECOGNIZING BY CONTENTS

If your file cannot be recognized by its file name, you might be able to
recognize it by its contents.  For example, many script files start with a
line like:

	#!/bin/xyz ~

To recognize this script create a file "scripts.vim" in your runtime directory
(same place where filetype.vim goes).  It might look like this: >

	if did_filetype()
	  finish
	endif
	if getline(1) =~ '^#!.*[/\\]xyz\>'
	  setf xyz
	endif

The first check with did_filetype() is to avoid that you will check the
contents of files for which the filetype was already detected by the file
name.  That avoids wasting time on checking the file when the "setf" command
won't do anything.
   The scripts.vim file is sourced by an autocommand in the default
filetype.vim file.  Therefore, the order of checks is:

	1. filetype.vim files before $VIMRUNTIME in 'runtimepath'
	2. first part of $VIMRUNTIME/filetype.vim
	3. all scripts.vim files in 'runtimepath'
	4. remainder of $VIMRUNTIME/filetype.vim
	5. filetype.vim files after $VIMRUNTIME in 'runtimepath'

If this is not sufficient for you, add an autocommand that matches all files
and sources a script or executes a function to check the contents of the file.

==============================================================================

Next chapter: |usr_44.txt|  Your own syntax highlighted

Copyright: see |manual-copyright|  vim:tw=78:ts=8:noet:ft=help:norl: