Re: utf text in a GtkTextView

[Havoc Pennington <hp redhat com>]

On Sat, Jun 07, 2003 at 04:46:47PM -0400, Rajarshi Guha wrote: 
> Now as far as I know the text was UTF8 encoded - is there anyway to
> determine the encoding of a text file?

No. The way it needs to work is that either the file has a fixed
encoding determined by its MIME type, or the file has its encoding
labeled in some way. For example, XML and HTML files have a tag
indicating their encoding. ".txt" files are usually either in the
locale encoding or in UTF-8, but really all you can do is have a
"choose the encoding" option in the user interface allowing the user
to specify manually.

If you got that g_utf8_validate() warning, then your file was not
valid UTF-8. But even if the file isn't, you shouldn't get that
warning - you need to be validating untrusted data, such as file
contents, before putting it in a GtkTextView. If the file is not
validly encoded, you should open an error dialog and complain to the
user, or perhaps use g_convert_with_fallback() to try to load as much
of it as possible.

> Do I need to use any special GTK function to insert different encodings?
> Or is this something that is supposed to be managed outside of GTK

You need to get things in UTF-8 before passing them to GTK. GLib
offers a lot of functions to help with that, such as 
g_utf8_validate(), g_convert(), g_locale_to_utf8(), and so forth.

If you pass GTK non-UTF-8, then GTK will often crash but will at 
least spew warnings and not work. It's a bug to pass GTK any text that
isn't known to be valid UTF-8.

Havoc