Markus Kuhn wrote: > > There is no standard for what you want to do, as this is getting > very far away from the classic VT100 / ISO 6429 terminal semantics. > No matter what you do, it will be your private encoding that isn't > compatible with anything else.
Assuming he takes your advice and does something other than corrupting UTF-8, there is an existing implementation that he could take some design hints from. DEC's VT420 and VT500 series terminals implemented what they called Key Position Mode, in which the terminal sent "key coordinates" from ISO 9995(?) instead of Latin-1 characters. The state of modifier keys was also encoded in the stream, to make it easier for the host to determine which keys were still held down. Multiple key pressed were packed into control strings, surrounded by the C1 control codes APC and ST. > Make sure that the ESC sequence that you use to activate this private > mark/break mode is as long and obscure as possible (at least 10 bytes, > but still within the ECMA-48 syntax for ESC sequences!) How about DECKPM: CSI 8 1 h - send key position reports CSI 8 1 l - send character codes - Paul -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/
