Terminal control/Unicode output

The task is to check that the terminal supports Unicode output, before outputting a Unicode character. If the terminal supports Unicode, then the terminal should output a Unicode delta (U+25b3). If the terminal does not support Unicode, then an appropriate error should be raised.

AWK

<lang awk>#!/usr/bin/awk -f BEGIN {

 unicodeterm=1   # Assume Unicode support
 if (ENVIRON["LC_ALL"] !~ "UTF") {
   if (ENVIRON["LC_ALL"] != ""
     unicodeterm=0    # LC_ALL is the boss, and it says nay
   else {
     # Check other locale settings if LC_ALL override not set
     if (ENVIRON["LC_CTYPE"] !~ "UTF") {
       if (ENVIRON["LANG"] !~ "UTF")
         unicodeterm=0    # This terminal does not support Unicode
     }        
   }    
 }

 if (unicodeterm) {
     # This terminal supports Unicode
     # We need a Unicode compatible printf, so we source this externally
     # printf might not know \u or \x, so use octal.
     # U+25B3 => UTF-8 342 226 263
     "/usr/bin/printf \\342\\226\\263\\n"
 } else {
     print "HW65001 This program requires a Unicode compatible terminal"|"cat 1>&2"
   exit 252    # Incompatible hardware
 }</lang>

Mathematica

<lang Mathematica>If[StringMatchQ[$CharacterEncoding, "UTF*"], Print[FromCharacterCode[30000]], Print["UTF-8 capable terminal required"]] ->田</lang>

PicoLisp

<lang PicoLisp>(if (sub? "UTF-8" (or (sys "LC_ALL") (sys "LC_CTYPE") (sys "LANG")))

  (prinl (char (hex "25b3")))
  (quit "UTF-8 capable terminal required") )</lang>

Bold text

Tcl

Tcl configures the standard output channel to use the system encoding by default. The system encoding is formally the encoding for use when communicating with the OS (e.g., for filenames) but is virtually always correlated with the default terminal encoding. <lang tcl># Check if we're using one of the UTF or "unicode" encodings if {[string match utf-* [encoding system]] || [string match *unicode* [encoding system]]} {

   puts "\u25b3"

} else {

   error "terminal does not support unicode (probably)"

}</lang> Note that idiomatic Tcl code would not perform such a check; it would just produce the output which would be translated as best as possible (possibly into the target encoding's placeholder character).

UNIX Shell

This script only checks if the name of the locale contains "UTF-8". This often works because many UTF-8 locales have names like "en_US.UTF-8". This script will fail to recognize a Unicode terminal if:

The locale is a UTF-8 locale, but does not have "UTF-8" in its name.
The locale uses some other Unicode Transformation Format, such as GB18030.

Works with: Bourne Shell

<lang bash>unicode_tty() {

 # LC_ALL supersedes LC_CTYPE, which supersedes LANG.
 # Set $1 to environment value.
 case y in
 ${LC_ALL:+y})		set -- "$LC_ALL";;
 ${LC_CTYPE:+y})	set -- "$LC_CTYPE";;
 ${LANG:+y})		set -- "$LANG";;
 y)			return 1;;  # Assume "C" locale not UTF-8.
 esac
 # We use 'case' to perform pattern matching against a string.
 case "$1" in
 *UTF-8*)		return 0;;
 *)			return 1;;
 esac

}

if unicode_tty; then

 # printf might not know \u or \x, so use octal.
 # U+25B3 => UTF-8 342 226 263
 printf "\342\226\263\n"

else

 echo "HW65001 This program requires a Unicode compatible terminal" >&2
 exit 252    # Incompatible hardware

fi</lang>

The terminal might support UTF-8, but its fonts might not have every Unicode character. Unless they have U+25B3, the output will not look correct. Greek letters like U+25B3 tend to be common, but some fonts might not have Chinese characters (for example), and almost no fonts have dead scripts such as Cuneiform.

ZX Spectrum Basic

<lang zxbasic>10 REM There is no Unicode delta in ROM 20 REM So we first define a custom character 30 FOR l=0 TO 7 40 READ n 50 POKE USR "d"+l,n 60 NEXT l 70 REM our custom character is a user defined d 80 PRINT CHR$(147): REM this outputs our delta 9500 REM data for our custom delta 9510 DATA 0,0,8,20,34,65,127,0 </lang>