Re^3: using colors with print()

[pbeckingham]

I believe there is a need to print the newline character after the reset, not before, or in other words, do not let color attributes span unnecessary lines, and those last lines dealing with $command are redundant.

sub textcolor ($$$);

print textcolor (RESET, RED, BLACK),
      'In color',
      textcolor (RESET, WHITE, BLACK),
      "\n";

sub textcolor ($$$) 
{
  my ($attr, $fg, $bg) = @_;
  sprintf "\e[%d;%d;%dm", $attr, $fg + 30, $bg + 40;
}

[tachyon]

Actually the ANSI escapes span lines so newline position is really immaterial (except from a neatness point of view). In fact AFAIK they remain active until reset, and this status outlives the life of your program as you are effectively manipulating defaults on your term.

 

$ cat col.pl #!/usr/bin/perl # default black background use constant BLUE => "\e[0;34;40m"; use constant RED => "\e[0;31;40m"; use constant DEFAULT => "\e[0;37;40m"; print "The flag is: ", RED, " red ", DEFAULT, " white & ", BLUE," blue. ANSI color escapes will span", DEFAULT, " at least with putty as the terminal emulation. ", BLUE; $ ./col.pl The flag is: red white & blue. ANSI color escapes will span at least with putty as the terminal emulation. $ $ echo Oops forgot to reset blue mode Oops forgot to reset blue mode $ perl -e 'print "\e[0;37;40m"'; $ echo Fixed.... Fixed....

[pbeckingham]

Oh, you're absolutely right, the sequences do span lines, and indeed should, but I had problems on Solaris 8/xterm when I relied on this. Perhaps I shouldn't have even mentioned this.

I will mention though, that the longer those sequences remain active, the greater the chance of the userr hitting Crtl-C and having their terminal in need of an \e[0m.