text modification

[yahve]

Hi,

I have a script that replaces the french accented caracters
by their html entities. I works great. I would like to improve it. I would like it to take the file name from the command line, back-up that file to whatever.bak and then modify the file and write it out to the original name.

How would I do that. How would I make my script back-up the file and then rewrite it with the modified text?

Thanks

 

[luciddream]

this is what i would do:

$file = shift || die "no input file\n";
$filebak = $file . ".bak";

open(FILE, "$file&quot || die;
@file = <FILE>;
close(FILE);

open(BACK, &quot;>$filebak&quot || die;
foreach(@file)
{
print BACK $_;
}
close(BACK);

open(FILE, &quot;$file&quot || die;
foreach(@file)
{
tr/stuff/otherstuff/g;
print FILE $_;
}
close(FILE);


or you could just use system() or `` to copy the file to filename.bak before you modify it.

 

[yahve]

Thanks Luciddream,

I appreciate your input but I went a different route:

#!/usr/bin/perl -w
$infile = shift || die;
$outfile = $infile . &quot;.NEW&quot;;
open(IN,$infile) || die;
open(OUT,&quot;>$outfile&quot || die;
while(<IN>) {
tr/old/new/;
print OUT $_;
}
close(IN);
close(OUT);

I think it is simpler. But maybe you can help me out an other way. Can you tell me, simply, what is the difference between s/// and tr/// ?

Thank you

 

[luciddream]

s/// does replacement. tr/// does translation..

with tr you can translate a range of characters to another range for example:

tr/[a-d]/[e-h]/g

translates every a to an e, every b to an f, every c to a g, and every d to an h.

with s you can only replace like this:

s/joe/dave/g

replaces every instance of joe with dave.

 

[yahve]

Ok Luciddream, I think I got that (maybe... english not being my 'forte'...).

If I can pick your brain some more:

1. how could I tell my script to process every file with a given extension (say .htm) in a directory?

I would like to give a path as the argument and have the script process each .htm file in that directory. Could I specify more than one extension? The first argument would be the directory and the second argument a LIST of extensions.

2. If in the script I change & to $#38; (s/&/$#38;/g) and also change é to $#233; (s/é/$#233;/g) how can I prevent the first command to change the & in the output of the second command? (I used $ because you'd see s/&/&/g if I did not)

Thanks


Vieux motard que jamais...

 

[luciddream]

#!/usr/bin/perl

$dir = shift;
$exts = shift;

@ext = split(',', $exts);

chdir &quot;$dir&quot;;
@files = glob('*');

foreach $file (@files)
{
foreach $ext (@ext)
{
($filename, $fileext) = split('.', $file);
if($fileext eq $ext)
{
push(@good_files, $file);
}
}
}

#@good_files now holds all of the files you want to modify.

foreach $file (@good_files)
{
open(FILE, &quot;$file&quot || die;
@lines = <FILE>;
close(FILE);
foreach(@lines)
{
s/\&/\$\#38;/g;
s/é/\$\#233;/g;
}
open(FILE, &quot;>$file) || die;
print FILE @lines;
close(FILE);
}


i'm not sure what you mean in the regex's but, that's how i would do it. there's also a way to open directories like filehandles. i don't recall off the top of my head how to do this but i'm sure it isn't hard to find out.