I asked this in another thread, no response maybe I can muster up one from
this thread ;)
I was informed of the following:
local ($^I, @ARGV) = (".bak", @sleepystart);
while (<>) {
$_ = "$db_name\n" if $. == 1;
print;
close ARGV if eof;
}
which will nab line one and change it to $db_name. I would like to ONLY muck
with the
files that have line one that differes from $db_name.
I thought of this
$_ = "$db_name\n" if ($. == 1 && $_ !~ /$db_name/);
and
$_ = "$db_name\n" if ($. == 1 && !~ /$db_name/); # which is the same
as above seeing it is still using $_ in the background
but no dice...
could someone kindle point me in the right direction?
regards
-Ron
> -----Original Message-----
> From: Michael R. Wolf [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, January 09, 2002 15:59
> To: [EMAIL PROTECTED]
> Subject: Re: Inline file edit
>
>
> [EMAIL PROTECTED] (Ronald Yacketta) writes:
>
> > This works great with one exception... the two values in
> the sed need to be
> > passed into the script...
> >
> > IE:
> > changedbsid.pl OLDSID NEWSID filename(s)
> >
> > could be ran against 1 or more files at a time.
> >
> > I havethis which works like a champ on a SINGLE file
> >
> > #!/usr/bin/perl -w
> >
> > $^I = ".bak";
> >
> > if ( @ARGV != 3 ) {
> > print "You must supply two command line options\n";
> > print "The OLD TOEKN and the NEW TOEKN\n";
> > print "\n IE: $0 _VALUTEST _QAP2\n";;
> > exit (1);
> > }
> >
> > $OLD = $ARGV[0];
> > $NEW = $ARGV[1];
> > $FILES = $ARGV[2];
> > @ARGV = glob $FILES;
> > while(<>) {
> > s/\Q$OLD\E/$NEW/g;
> > print;
> > }
> >
> > my problem is with multiple files, I could axe the ARGV
> check above.. but
> > that wont get me anywhere I do not think...
>
> The changes I made:
>
> 1. Lower case variable names. More Perl-ish, and less eye
> and shift-key strain.
>
> 2. Move the glob from Perl to the command line. That makes
> @ARGV pre-populated with the filenames. (This may not
> work on WinDOS systems. I don't know; I'm used to having
> glob done by the Korn shell in Unix.)
>
> 2a. The program now works on STDIN by specifying no
> filename arguments. (It may work by specifying "-" as
> an argument; I forget).
>
> 2b. The program now works for multple file specifications.
> It isn't limited to working with one file glob spec.
> Specify 'em all on the line, it'll process 'em all.
>
> 3. Conditional setting of in-place editing argument. I'd
> bet that it worked OK to set it for STDIN, but I don't
> have a Perl installation on my traveling laptop to check,
> so I played it safe by undef'ing it for STDIN
> processing.
>
> 4. Shift off the required leading arguments rather than
> accessing them directly through indices. This leaves
> @ARGV to contain only the filenames.
>
> The changes I didn't (really) make:
>
> 1. My check for the number of args is basically the same as
> yours. I mucked with it for my own personal style
> reasons since I was alrady playing with ARGV for
> globbing.
>
> Here's how I'd do it.
>
> ================================================================
> #!/usr/bin/perl -w
>
> $program_name = $0;
>
> # Print a usage message to STDERR
> sub usage {
> print STDERR "USAGE: $program_name old_token new_token
> [files...]\n";
> }
>
> unless ( @ARGV >= 2 ) {
> usage;
> exit 1;
> }
>
> $old_token = shift @ARGV;
> $new_token = shift @ARGV;
>
> # Elements left in @ARGV are filenames.
> # If @ARGV is empty, use STDIN.
> # Make backups if there are filenames to process.
> $^I = @ARGV ? ".bak" : undef;
>
> while(<>) {
> s/\Q$old_token\E/$new_token/g;
> print;
> }
>
> --
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
