Customizing device character mapping

Seven-bit terminals

The seven-bit ASCII code set as defined by ISO 646 is used by many terminals. Using this code set, support for European languages is limited and involves modifying the ISO 646 code set. Some of the "non-essential" characters are changed to represent specific unsupported characters. The characters that are replaced are shown in the following table. The applications used with these terminals have to use the new character value assignments. The generic term for such variations of the seven-bit ISO 646 code set is "seven-bit ISO 646 national variant" code sets and one exists for each Western European language.

Seven-bit ISO 646 national variants

23 24 40 5b 5c 5d 5e 60 7b 7c 7d 7e

# $ @ [ \ ] ^ ` { | } ~

DE # $ § Ä Ö Ü ^ ` ä ö ü ß

DK # $ @ Æ Ø Å ^ ` æ ø å ~

ES £ $ § ¡ Ñ ¿ ^ ` ° ñ ç ~

FR £ $ à ° ç § ^ ` é ù è ¨

GB £ $ @ [ \ ] ^ ` { | } ~

IT £ $ § ° ç é ^ ù à ò è ì

NO # $ @ Æ Ø Å ^ ` æ ø å ~

SE # $ É Ä Ö Å Ü é ä ö å ü

A problem involving terminal control sequences may be introduced if code set mapping is used to support an ISO 646 national variant code set on a seven-bit terminal. It is introduced because the internal (system) code set can be different from the external code set for the terminal.

The problem arises when the terminal control sequence includes a character that has a different interpretation to that in the ISO 646 code set. An ISO 646 national variant code set gives certain character values new interpretations. For example, the French ISO 646 national variant associates the value normally reserved for the left brace, { with e-acute, é.

The problem manifests itself in the following two ways:

On input, the KBD module converts the external (terminal) code set to the internal code set (ISO 8859-1). For example, on a French ISO 646 national variant terminal, the value normally assigned to the left bracket, [ (0x5b) is re-assigned to the degree sign, °. The KBD module maps this value (0x5b) to the ISO 8859-1 code for degree sign (0xb0). Thus, the up arrow key on a French vt100 terminal will send the control sequence "ESC [ A" (0x1b, 0x5b, 0x41), but the receiving process will receive the sequence "ESC ° A" (0x1b, 0xb0, 0x41).
On output the KBD module converts the internal code set (ISO 8859-1) to the external code set for the terminal. Some characters are not valid on some terminals. For example, on a French terminal the left bracket, [ is not present (it is replaced by degree sign, ° ). The KBD module recognizes that [ is not a valid character and substitutes a default character (usually the underscore, _). Hence, if a process sends the control sequence for up arrow for a French vt100 terminal, "ESC [ A" (0x1b, 0x5b, 0x41), the terminal actually receives the sequence: "ESC _ A" (0x1b, 0x5f, 0x41).

This problem only affects seven-bit terminals using the ISO 646 national variant code sets. The Language Supplement provides terminfo terminal database specifications for the ISO 646 national variants shown in ``terminfo entries supported''. For example, in the French ISO 646 national variant terminal the up arrow control sequence is defined as (0x1b, 0x5b, 0x41). Hence, the terminal database entry for up arrow"should be "\E°A" (0x1b 0xb0 0x41). The KBD module will then map 0xb0 to 0x5b and the terminal will function correctly.

terminfo entries supported

terminfo Entry Country

vt100DK Denmark

vt100FR France

vt100DE Germany

vt100IT Italy

vt100NO Norway

vt100ES Spain

vt100SE Sweden

vt100GB Great Britain

terminfo Entry	Country
vt100DK	Denmark
vt100FR	France
vt100DE	Germany
vt100IT	Italy
vt100NO	Norway
vt100ES	Spain
vt100SE	Sweden
vt100GB	Great Britain