Re: libgettextpo: Ability to work with non-ASCII without a header

[Bruno Haible]

John Cowan wrote:
> If you view any non-ASCII file using
> a dumb (non-locale-sensitive) text editor, you will see mojibake unless
> your locale's encoding matches the file's.

I'm talking about
  - a normal (locale aware) text editor,
  - a user in ISO-8859-x or EUC-JP locale,
  - a PO file in UTF-8 encoding.
This yields mojibake.

Therefore, as long as users are not always in UTF-8 locales, we cannot make
all PO files UTF-8 encoded.

> Currently, there is a rule "if no encoding header and has high-bit
> bytes, then error"; I'm proposing changing that to "if no encoding
> header and has high-bit bytes, then UTF-8".

Not good. People will then start to write scripts using sed, grep or similar
tools that work only with UTF-8 encoded PO files.

I force you to put a header on each PO file. You don't like it. But it has
the effect that your scripts (if reasonably written) will work on PO files
in any encoding. And as long as not all PO files are in UTF-8, I insist on
that.

Bruno