Thank you. I will study this more perlish style and upgrade my
program. Chee
On Fri, 2 Apr 2004, Wiggins d'Anconia wrote:
> Date: Fri, 02 Apr 2004 20:21:11 -0500
> From: Wiggins d'Anconia <[EMAIL PROTECTED]>
> To: Charlotte Hee <[EMAIL PROTECTED]>
> Cc: [EMAIL PROTECTED]
> Subject: Re: illegal octal digit '8'?
>
> Charlotte Hee wrote:
> >
> > In my perl script I use the unix command 'date' to make a time stamp
> > because eventually I want the date format to be DD-MON-YYYY, where DD is
> > the number of the day (1-31), MON is the 3-char spelling of month (i.e.
> > JAN), and YYYY is the 4-digit year. In this perl script I'm just testing
> > for the date so the day and year aren't in this script.
> > The date command returns the month as a two digit number which I pass to
> > my routine. My routine then changes the two digit number into the
> > corresponding 3-char letters for that month. I get an error message like
> >
> > Illegal octal digit '8'
> > Illegal octal digit '9'
> >
> > when my routine uses the number 08 and 09. I don't understand what
> > this error means. If I change my routine to use 8 and 9 I don't see this
> > error message. The number for April is 04 and that doesn't seem to cause a
> > problem. What is special about numbers 08 and 09?
> >
>
> In general this is a bad way of going about all of this, but we will get
> to that in a moment...
>
>
> > Here is the code: ( thanks, Chee )
> >
> >
> > #!/usr/local/bin/perl -w
> > #
> >
> > use strict;
>
> A good start...
>
> > use vars qw($mm $mon);
> >
>
> perldoc -f our 'use vars is a bit dated, unless you know why to use them
> or have an older Perl, stick to 'our'. Having said that do they need to
> be package variables in the 'main' namespace? Better to make them
> lexically scoped with 'my'
>
> perldoc -f my
>
> > $mm = `date +%m`;
>
> There is no reason to shell out to get the date components *especially*
> if you aren't going to use full paths, chomp the output, check for error
> codes, etc.
>
> perldoc -f localtime
>
> > $mon = &convertMO($mm);
> >
>
> Drop the & until you know why you don't.
>
> > print "month is $mon\n";
> >
> > exit;
> >
> > sub convertMO {
> > my ($tmp) = @_;
> > my $xmon;
> >
> > if ( $tmp == 01 ){ $xmon = 'JAN'; }
> > if ( $tmp == 02 ){ $xmon = 'FEB'; }
> > if ( $tmp == 03 ){ $xmon = 'MAR'; }
> > if ( $tmp == 04 ){ $xmon = 'APR'; }
> > if ( $tmp == 05 ){ $xmon = 'MAY'; }
> > if ( $tmp == 06 ){ $xmon = 'JUN'; }
> > if ( $tmp == 07 ){ $xmon = 'JUL'; }
> > if ( $tmp == 08 ){ $xmon = 'AUG'; } #<-- illegal error
> > if ( $tmp == 09 ){ $xmon = 'SEP'; } #<-- illegal error
>
> Since you are using the numerical comparison on numbers that start with
> a '0' which Perl takes to mean as numbers in octal rather than decimal.
> As soon as you hit 8 and 9 the digits are not expected in an octal
> number. You would be better off using string comparison, or dispensing
> with this subrouine anyways. It would be much simpler and faster to
> store the month abbreviations in a hash and index into that hash using a
> key, better still use an array since it is already ordered with numbered
> indexes.
>
>
> > if ( $tmp == 10 ){ $xmon = 'OCT'; }
> > if ( $tmp == 11 ){ $xmon = 'NOV'; }
> > if ( $tmp == 12 ){ $xmon = 'DEC'; }
> >
> > return $xmon;
> > }
> >
>
>
> So a Perlish way to do this,
>
> my (@months) =
> ('jan','feb','mar','apr','may','jun','jul','aug','sep','oct','nov','dec');
> my ($mday, $mon, $year) = (localtime)[3,4,5];
>
> my $date = sprintf('%02d-%3s-%04d', uc $mday, $months[$mon] , $year+1900);
>
> print "Date: $date\n";
>
> perldoc -f uc
> perldoc -f sprintf
>
> HTH,
>
> http://danconia.org
>
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
<http://learn.perl.org/> <http://learn.perl.org/first-response>
