On Mon, Jan 24, 2005 at 04:22:47PM +0000, David Sumbler wrote: > I'm trying to get my head around keyboard and character coding, utf-8 > etc., particularly as regards the Linux consoles which I use much of > the time. > > I run Fedora Core 2, Emacs 21.3 and Gnus v5.10.6. The default console > font seems to be latarcyrheb-sun16, and the keyboard map is uk.map. > I'd be grateful if someone can explain to me: > > 1) Where do the "compose" combinations come from?
There is a default list of 68 compose combinations - see kernel source, drivers/char/defkeymap.*. You can set your own using your keymap (with loadkeys). See keymaps(5) and loadkeys(1). > 2) Although most, perhaps all, of the composed characters seem to > exist when I do 'showconsolefont', none of them actually shows up > when I do 'dumpkeys'. I just get lots of lines such as > > compose '`' 'A' to '' A design error. The actual symbols (bytes) are shown, and they are in some unspecified 8-bit character set assumed to be your local character set. (Otherwise, why would you want to use such compose definitions?) The default compose definitions are for ISO 8859-1. Today many people are running utf-8 and these 8-bit characters are not valid utf-8 symbols. Therefore the output of dumpkeys is not very enlightening. > 3) But if I do 'dumpkeys > filex' and then view filex in Emacs, the > characters do show up. > > 4) Also in Emacs, if I press, say, PrintScreen then ` then A, an A > with a grave accent appears (followed by a hyphen) in the > minibuffer, but not in the buffer I am editing. How do I get round > this? Such things depend on your local setup. > 5) I think I read somewhere that a console font can have a maximum of > 512 characters defined. Is this true? Yes. Andries -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
