TITLE

\v for Vertical Tab

VERSION

  Maintainer: Nicholas Clark <nick@talking.bollo.cx>
  Date: 26 Sep 2000
  Last Modified: 30 Sep 2000
  Mailing List: perl6-language@perl.org
  Number: 327
  Version: 3
  Status: Frozen

CHANGES

Reissued on perl6-language@perl.org - I goofed the list. So far no comments on it.

Summarised discussion and changed status to frozen

DISCUSSION

Very little comment was made. Russ Allbery notes

        The last time I used a vertical tab intentionally and for some
        productive purpose was about 1984.

but states that this isn't an objection to the RFC.

ABSTRACT

perl5 includes all of C's escapes except \v (vertical tab). Treating \v the same as \a \b \e \f \h \r \t would remove a special case.

DESCRIPTION

man perl says:

       Perl combines (in the author's opinion, anyway) some of
       the best features of C, sed, awk, and sh, so people
       familiar with those languages should have little
       difficulty with it.

However, lack of \v represents a special case for a C programmer to learn. \v isn't used for anything else in double quoted strings, nor is it used in regular expressions, so it won't require removal of an existing feature to add it. Currently a \v in a double quoted strings will be treated as v, with a warning about unknown escape issued if warnings are in force.

Vertical tab was also omitted from the range of characters considered whitespace by \s in regular expressions.

This RFC proposes

\v becomes a recognised escape for a vertical tab in interpolated contexts (double quoted strings and regular expressions)
That vertical tab is moved from the \S (non-whitespace) to \s (whitespace) class in regular expressions.

IMPLEMENTATION

Shouldn't be hard. Here are patches for perl 5.7.0

\v

    --- toke.c.orig     Fri Sep 15 04:14:57 2000
    +++ toke.c  Tue Sep 26 12:54:30 2000
    @@ -469,7 +469,7 @@
        for (t = s; !isSPACE(*t); t++) ;
        e = t;
         }
    -    while (SPACE_OR_TAB(*e) || *e == '\r' || *e == '\f')
    +    while (SPACE_OR_TAB(*e) || *e == '\r' || *e == '\f' || *e == '\v')
        e++;
         if (*e != '\n' && *e != '\0')
        return;         /* false alarm */
    @@ -1196,7 +1196,7 @@
        : UTF;
         const char *leaveit =  /* set of acceptably-backslashed characters */
        PL_lex_inpat
    -       ? "\\.^$@AGZdDwWsSbBpPXC+*?|()-nrtfeaxcz0123456789[{]} \t\n\r\f\v#"
    +       ? "\\.^$@AGZdDwWsSbBpPXC+*?|()-nrtfveaxcz0123456789[{]} \t\n\r\f\v#"
            : "";
     
         while (s < send || dorange) {
    @@ -1540,6 +1540,9 @@
            case 'f':
                *d++ = '\f';
                break;
    +       case 'v':
    +           *d++ = '\v';
    +           break;
            case 't':
                *d++ = '\t';
                break;
    @@ -2700,7 +2703,7 @@
        Perl_croak(aTHX_ 
           "\t(Maybe you didn't strip carriage returns after a network transfer?)\n");
     #endif
    -    case ' ': case '\t': case '\f': case 013:
    +    case ' ': case '\t': case '\f': case '\v':
     #ifdef MACOS_TRADITIONAL
         case '\312':
     #endif
    --- t/op/pat.t.orig Tue Aug 29 13:54:13 2000
    +++ t/op/pat.t      Tue Sep 26 12:56:36 2000
    @@ -469,27 +469,27 @@
     print "ok $test\n";
     $test++;
     
    -print "not " unless qr/\b\v$/i eq '(?i-xsm:\bv$)';
    +print "not " unless qr/\b\y$/i eq '(?i-xsm:\by$)';
     print "ok $test\n";
     $test++;
     
    -print "not " unless qr/\b\v$/s eq '(?s-xim:\bv$)';
    +print "not " unless qr/\b\y$/s eq '(?s-xim:\by$)';
     print "ok $test\n";
     $test++;
     
    -print "not " unless qr/\b\v$/m eq '(?m-xis:\bv$)';
    +print "not " unless qr/\b\y$/m eq '(?m-xis:\by$)';
     print "ok $test\n";
     $test++;
     
    -print "not " unless qr/\b\v$/x eq '(?x-ism:\bv$)';
    +print "not " unless qr/\b\y$/x eq '(?x-ism:\by$)';
     print "ok $test\n";
     $test++;
     
    -print "not " unless qr/\b\v$/xism eq '(?msix:\bv$)';
    +print "not " unless qr/\b\y$/xism eq '(?msix:\by$)';
     print "ok $test\n";
     $test++;
     
    -print "not " unless qr/\b\v$/ eq '(?-xism:\bv$)';
    +print "not " unless qr/\b\y$/ eq '(?-xism:\by$)';
     print "ok $test\n";
     $test++;

\S to \s

    --- handy.h.orig    Thu Sep 14 15:44:20 2000
    +++ handy.h Tue Sep 26 13:04:30 2000
    @@ -295,8 +295,9 @@
     #define isIDFIRST(c)       (isALPHA(c) || (c) == '_')
     #define isALPHA(c) (isUPPER(c) || isLOWER(c))
     #define isSPACE(c) \
    -   ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f')
    -#define isPSXSPC(c)        (isSPACE(c) || (c) == '\v')
    +   ((c) == ' ' || (c) == '\t' || (c) == '\n' || (c) =='\r' || (c) == '\f' \
    +    || (c) == '\v')
    +#define isPSXSPC(c)        isSPACE(c)
     #define isBLANK(c) ((c) == ' ' || (c) == '\t')
     #define isDIGIT(c) ((c) >= '0' && (c) <= '9')
     #ifdef EBCDIC
    --- t/op/pat.t.orig Tue Aug 29 13:54:13 2000
    +++ t/op/pat.t      Tue Sep 26 14:27:14 2000
    @@ -1064,15 +1064,14 @@
              cr    => "\r",
              lf    => "\n",
              ff    => "\f",
    -# The vertical tabulator seems miraculously be 12 both in ASCII and EBCDIC.
    -         vt    => chr(11),
    +         vt    => "\v",
              false => "space" );
     
     my @space0 = sort grep { $space{$_} =~ /\s/ }          keys %space;
     my @space1 = sort grep { $space{$_} =~ /[[:space:]]/ } keys %space;
     my @space2 = sort grep { $space{$_} =~ /[[:blank:]]/ } keys %space;
     
    -print "not " unless "@space0" eq "cr ff lf spc tab";
    +print "not " unless "@space0" eq "cr ff lf spc tab vt";
     print "ok $test\n";
     $test++;

To be strict the perl5 to perl6 converter would need to

replace \v with v in interpolated strings.
replace \s with [\t\n\r\f ] and \S with [^\t\n\r\f ] in regular expressions.

It might be considered acceptable to omit either or both conversions if the number of programs that would break were negligible.

REFERENCES

perlop manpage for interpolation

perlre manpage for \s and \S