How to delete mutiple line from many pm files at one time

[Priti24]

suppose i have a folder name 'rough' having 10-15 pm files.i have to delete some common lines from all pm files.like...

my $t0 = [gettimeofday];
    my $time = tv_interval ( $t0, [gettimeofday]);
    my @timeData = localtime(time);
        $time= join(':', @timeData);
        if($time =~m#^(.*?)\:(\d+)\:(\d+).*?$#)
        {
            $time = $3 . ':' . $2 . ':' . $1;
        }

how should we do that?

Plz help me out

[Corion]

You show code, but you don't explain how the code you have relates to the problem you state. Please help us help you better by explaining in prose what the code you have should do, and what it does, and how the two differ.

Maybe consider looking at Tie::File to treat a file like an array. Or at File::Inplace or at the $^I special variable.

[johngg]

Do you mean that you want to remove the code lines you show from all of the *.pm files? If so, you could so this with in-place editing and a flip-flop (..) operator. You have to make sure you escape any regular expression meta-characters in the code lines you use to delineate start and end of the section to remove.

knoppix@Microknoppix:~/perl/Monks$ ls d971401/
File1.pm  File2.pm  File3.pm  File4.pm  File5.pm
knoppix@Microknoppix:~/perl/Monks$

Read more... (2 kB)

knoppix@Microknoppix:~/perl/Monks$ perl -ni.BAK -e '
print unless m{^my \$t0 = \[gettimeofday\];} .. m{^\s+\}};' d971401/*.
+pm
knoppix@Microknoppix:~/perl/Monks$ ls d971401/*
d971401/File1.pm      d971401/File3.pm      d971401/File5.pm
d971401/File1.pm.BAK  d971401/File3.pm.BAK  d971401/File5.pm.BAK
d971401/File2.pm      d971401/File4.pm
d971401/File2.pm.BAK  d971401/File4.pm.BAK
knoppix@Microknoppix:~/perl/Monks$

Read more... (759 Bytes)

I hope that I have guessed correctly and that this is helpful.

[NetWallah]

Here is a one-liner for this:

 perl -p0777 -i.BAK -e 's/my \$t0 = \[gettimeofday\];.+?\.\s*\$1;\s+}/
+/s;' Your.module.names.here

This takes advantage of the -0 option to set the input record separator to 0x777 (which is not a valid character, and can not occur in your files), and the "s" option in the regex to span matching a '.' across multiple lines.

The explanation in this debian admin article is mostly applicable - except they use the regex option 'm' to handle a slightly different use-case.

If you wanted to match line boundaries, in addition to matching '.' across lines, the 's' and 'm' options can be used together.
Update: Minor Aesthetic change: -iBAK to -i.BAK