created July 22, 2006 · complexity basic · author Tony Mechelynck · version 6.0
Usually, when opening an editfile while 'encoding' is set to UTF-8, Vim will try UTF-8 before Latin1. This is all good and proper since the UTF-8 heuristic can give an error result and the Latin1 one cannot. However, if the file contains (at the moment) only 7-bit ASCII data, it will be represented identically under Latin1 and UTF-8, and no one (not even a human) can decide without additional information.
If we want later-added bytes above 127 to be represented as Unicode, the solution is simple: we can add a BOM to the file by setting the 'bomb' option (":setlocal bomb"). But what if we want later data between 128 and 255 to be represented as Latin1? This is a little trickier, but the answer is simple; it's of the kind that makes one say "Now why didn't I think of that?". Here it is:
- If you want your file to be always interpreted by Vim as Latin1 and never as UTF-8, place in it a line with a number of "upper-ASCII" bytes in it (i.e. bytes in the range 128-255), then save the file after making sure that 'fileencoding' is set to Latin1, e.g. by doing ":setlocal fenc=latin1". In a program source file or a shell script, the line can be a comment; in a text file, it can be "fancy underlining". The following will do (for example):
# zim: set fenc=latin1 nomod : ÷÷÷÷÷÷
Place it where you would a modeline. Notice that with the letter v instead of z in column 3 it would be a modeline; but the problem with setting the 'fileencoding' by means of a modeline is that the latter's action happens only after the file has been read. This line will make the "typical" 'fileencodings' heuristics (as set, e.g., by ":set fileencodings=ucs-bom,utf-8,latin1") to give a "negative" result for UTF-8, so that Vim's automatic mechanisms will recognise the file as Latin1.
See also Working with Unicode.
Comments
If the purpose is to ensure the loader fails to match it UTF-8, why not just put something like #£ in the file. Why bother with the pseudo-modeline?
- You could. One possible way is to use Latin1 "divide-by signs" to underline the top title. It all depends what you want the file to look like. You may also want to make sure that none of your co-editors (if any) remove the upper-ASCII bytes, and that may require a comment. -- Tonymec 15:48, 29 November 2007 (UTC)
To force opening a file in Latin1 I just call vim this way: $ vim --cmd "set encoding=Latin1" file.txt -- Cocker68 2011-08-19 23:35:10 (UTC+2)