SPUG:Best One-Liners and Scripts for UNIX

[Richard Anderson]

Well, this example of a scriptlet scores zero on the cleverness scale, but
it has been so useful to me that I'll run it up the flagpole.  When I'm
writing code I often find it faster to test a Perl feature rather than
reading the online or printed docs (which are sometimes misleading).  So I
run my psh (Perl shell) scriptlet:

#! /usr/bin/perl
while (<>) {
    eval;
    if ($@) { print $@ }
}

and type in the commands that test the feature.  No terminating semicolons
are needed on the commands and the if statement displays any Perl error
messages.

Trivial, but one of my more heavily-used scriptlets.

Richard Anderson
richard at richard-anderson.org
www.richard-anderson.org
----- Original Message -----
From: "Tim Maher" <tim at consultix-inc.com>
To: <spug-list at pm.org>
Sent: Wednesday, April 16, 2003 5:49 PM
Subject: SPUG:Best One-Liners and Scripts for UNIX


> SPUGsters,
>
> Seems like most people stayed home to work on their taxes last
> night, judging from the turnout at our April SPUG meeting 8-{
>
> But I gave my talk anyway -- well, at least *part* of it.
>
> It took me the full two hours to cover "Perl as a Better
> Grep, Sed, and Awk", so I didn't delve into the other
> advertised topics, which were:
>
>    * 5 Perl One-liners All Unix/Linux Users Should Know
>    * 3 Perl Scripts UNIX/Linux Users Shouldn't Live Without
>    * How Perl's Looping Facilities Compare to the Shell's
>
> I've got my own ideas about what the "5" and the "3" are, but I'm
> very interested to get input from you folks on your favorites!
>
> So please give some thought as to what Perl one-liners or small
> scripts you value for your UNIX/Linux work, and post them to the
> list so we can discuss them.  If I like your submissions, with
> your permission, I'll include them in my book, and you'll get
> "your foot-noted". 8-}
>
> I'll start off the discussion with one of each.
>
> In the one-liner category, a command to convert MEMO.TXT (transferred
> from some evil OS) to memo.txt:
>
> echo MEMO.TXT | perl -wlne 'rename $_, "\L$_\E"'
> (It's processing input because arg-handling takes more typing.)
>
> ls *[A-Z]* | perl -wlne 'rename $_, "\L$_\E"'
> (to rename all files in directory)
>
> Here's a *deluxe version* with error-checking:
> perl -wlne 'rename $_, "\L$_\E" or
> warn "rename failed for $_, $!\n"'
>
>     (Surprisingly, there's no warning for a missing file,
>     but there's a missing file warning for a permission error
>     -- at least on Linux!)
>
> In the script category, I offer one that sorts filenames by modification
> time, which is something
> find . -name 'x' | xargs ls -rdlt
> can't be trusted to do, because of the possibility of separate sorts for
> different argument sets with lots of filenames.
>
> I'll post the code samples from last night's talk on my web site soon,
> and announce that to the list when it happens.
>
> Here's the script UNIX SAs shouldn't live without:
>
> #! /usr/local/bin/perl -w
> # sort_files_by_mtime
> # Tim Maher, tim at teachmeperl.com
> #
> # Sorts files by mod times, using Schwartzian transform,
> # and prints filename and mod-time.
> # Gives correct answers for large collections of names,
> # where "find . -print | xargs ls -rdtl" cannot
>
> # NOTE: Use find or locate to provide input, or ls -d dir/*,
> # but *not* simply "ls dir" (dir won't be present in pathnanme,
> # which will trigger "invalid file" error
>
> # Sample invocations:
> #   locate '*.c' | sort_files_by_mtime
> #   find /local -name 'somescript' | sort_files_by_mtime
> #   sort_files_by_mtime < /tmp/filelist
>
> $USAGE="Usage: $0 < file_list OR find ... | $0\n" ;
> @ARGV  and  die "$USAGE" ;
> @files=<STDIN>  or  exit 0 ;    # no input is okay
> chomp @files ;
>
> $max=0 ;    # suppress warning
> for ($_=0; $_ < @files; $_++) {
>     if ($_ ne ""  and  -e $files[$_]) {
>         length $files[$_] > $max  and
>             $max = length $files[$_]
>     }
>     else {
>         die "$0: Invalid filename, '$files[$_]'\n" ;
>         delete $files[$_] ;
>     }
> }
> @files  or  exit 0 ;
>
> map { printf "%${max}s  %s\n", $_->[0], (scalar localtime $_->[1]) }
>     sort { $a->[1] <=> $b ->[1] }   # compare m-timestamps
>         map { [ $_ , (stat $_)[9] ]  } @files ;
>
> -Tim
> *------------------------------------------------------------*
> |  Tim Maher (206) 781-UNIX  (866) DOC-PERL  (866) DOC-UNIX  |
> |  CEO, JAWCAR ("Just Another White Camel Award Recipient")  |
> |  tim at Consultix-Inc.Com  TeachMeUnix.Com  TeachMePerl.Com   |
> *+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-*
> |  Watch for my  Book: "Minimal Perl for Shell Programmers"  |
> *------------------------------------------------------------*
> _____________________________________________________________
> Seattle Perl Users Group Mailing List
> POST TO: spug-list at mail.pm.org
> ACCOUNT CONFIG: http://mail.pm.org/mailman/listinfo/spug-list
> MEETINGS: 3rd Tuesdays, U-District, Seattle WA
> WEB PAGE: www.seattleperl.org
>
>