Re: [Gnu-arch-users] How does arch/tla handle encodings?

[Marcus Sundman]

On Friday 27 August 2004 19:59, John Meinel wrote:
> In most cases, I think it just depends on what the other tools do. Most
> times, arch doesn't really care. It just asks diff to determine what to
> do. I'm guessing the standard GNU diff treats a UTF-16 file as binary
> (since for English text every other character is null.)
>
> You can argue for having a better diff tool, but I don't think the
> revision control system itself should care too much about what is in the
> file.

Even if it doesn't it still should keep track of that piece of metadata, so 
that it can tell diff what encoding the file uses so that diff can make a 
proper decision about what it should do.

> I'm thinking that you don't really want tla to say "Oh, you're using
> UTF-8, let me translate all of the UTF-16 files into your native
> encoding."
>
> Because more than likely you need UTF-16 for whatever is using the code,
> and translating is *not* what you want.

More than likely transcoding is *precisely* what you want. For text files, 
that is. There are some cases where you do not want this, e.g. when 
handling xml files (since the encoding is specified inside the xml file 
itself, a solution scorned by purists). This is obvious and it is why I 
said that the transcoding shouldn't be done for all files. However, 
regardles of whether there is transcoding or not, arch still has to keep 
track of the encoding.


- Marcus Sundman