From Code to Community: Sponsoring The Perl and Raku Conference 2025 Learn more

package Proc::Find;
our $DATE = '2019-11-23'; # DATE
our $VERSION = '0.051'; # VERSION
use 5.010001;
use strict;
use List::Util qw(first);
require Exporter;
our @ISA = qw(Exporter);
our @EXPORT_OK = qw(
find_proc
find_any_proc
find_all_proc
proc_exists
);
our $CACHE = 0;
my $_table_res;
sub _table {
state $pt = do {
Proc::ProcessTable->new;
};
if (!$CACHE || !$_table_res) {
$_table_res = $pt->table;
}
$_table_res;
}
sub _match {
my ($target, $cond, $is_numeric) = @_;
if (ref $cond eq 'Regexp') {
return 0 unless $target =~ $cond;
} elsif (ref $cond eq 'ARRAY') {
return 0 unless grep { $is_numeric ? $target == $_ : $target eq $_ } @$cond;
} else {
return 0 unless $is_numeric ? $target == $cond : $target eq $cond;
}
1;
}
sub find_proc {
my %args = @_;
my @unknown_args = grep {!/\A(
filter|
pid|name|cmndline|exec|
user|uid|euser|euid|
table|detail|
result_max
)\z/x} keys %args;
die "Unknown arguments to find_proc(): ".join(", ", @unknown_args)
if @unknown_args;
my $table = $args{table} // _table();
my ($arg_uid, $arg_euid);
my @res;
for my $p (@$table) {
# create extra fields
$p->{name} = $p->{cmndline};
$p->{name} =~ s/\s.*//;
$p->{name} =~ s!.+/!!;
my $cond = 0;
COND:
{
if (defined $args{filter}) {
local $_ = $p;
last COND unless $args{filter}->($p);
}
if (defined $args{pid}) { last COND unless _match($p->{pid} , $args{pid} , 1) }
if (defined $args{name}) { last COND unless _match($p->{name} , $args{name} ) }
if (defined $args{cmndline}) { last COND unless _match($p->{cmndline}, $args{cmndline}) }
if (defined $args{exec}) {
my $exec = $p->{exec} // '';
unless ($args{exec} =~ m!/!) {
$exec =~ s!.+/!!;
}
last COND unless _match($exec, $args{exec});
}
if (defined($args{user}) || defined($args{uid})) {
my $cond = $args{user} // $args{uid};
if ($cond eq 'Regexp') {
last COND unless _match($p->{uid}, $cond, 1); # XXX allow matching against username?
} else {
my @uids;
for my $val (ref $cond eq 'ARRAY' ? @$cond : $cond) {
if ($val =~ /\A\d+\z/) {
push @uids, $val;
} else {
my @pw = getpwnam($val);
push @uids, @pw ? $pw[2] : -1;
}
}
last COND unless _match($p->{uid}, \@uids, 1);
}
}
if (defined($args{euser}) || defined($args{euid})) {
my $cond = $args{euser} // $args{euid};
if ($cond eq 'Regexp') {
last COND unless _match($p->{euid}, $cond, 1); # XXX allow matching against username?
} else {
my @uids;
for my $val (ref $cond eq 'ARRAY' ? @$cond : $cond) {
if ($val =~ /\A\d+\z/) {
push @uids, $val;
} else {
my @pw = getpwnam($val);
push @uids, @pw ? $pw[2] : -1;
}
}
last COND unless _match($p->{euid}, \@uids, 1);
}
}
$cond = 1;
}
$cond = !$cond if $args{inverse};
next unless $cond;
if ($args{detail}) {
push @res, { %$p }; # unbless
} else {
push @res, $p->{pid};
}
if (defined $args{result_max}) {
last if @res >= $args{result_max};
}
}
\@res;
}
sub proc_exists {
@{ find_proc(@_, result_max=>1) } > 0 ? 1:0;
}
sub find_any_proc {
return [] unless @_;
my $detail = $_[0]->{detail};
my $table = $_[0]->{table} // _table();
my @allres;
for my $crit (@_) {
my $res = find_proc(%$crit, table=>$table, detail=>$detail);
ITEM:
for my $item (@$res) {
# skip duplicate process
if ($detail) {
next ITEM if first {$_->{pid} == $res->{pid}} @allres;
} else {
next ITEM if first {$_ == $res} @allres;
}
push @allres, @$res;
}
}
\@allres;
}
sub find_all_proc {
return [] unless @_;
my $detail = $_[0]->{detail};
my $table = $_[0]->{table} // _table();
my @allres;
CRIT:
for my $crit (@_) {
my $res = find_proc(%$crit, table=>$table);
if (!@allres) {
push @allres, @$res;
next CRIT;
}
@allres = grep {
my $p = $_;
$detail ?
((first {$p->{pid} == $_->{pid}} @$res) ? 1:0) :
((first {$p == $_} @$res) ? 1:0)
} @allres;
}
\@allres;
}
1;
# ABSTRACT: Find processes by name, PID, or some other attributes
__END__
=pod
=encoding UTF-8
=head1 NAME
Proc::Find - Find processes by name, PID, or some other attributes
=head1 VERSION
This document describes version 0.051 of Proc::Find (from Perl distribution Proc-Find), released on 2019-11-23.
=head1 SYNOPSIS
use Proc::Find qw(find_proc proc_exists);
# list all of a user's processes
my $procs = find_proc(user=>'ujang', detail=>1);
# check if a program is running
die "Sorry, xscreensaver is not running"
unless proc_exists(name=>'xscreensaver').
=head1 DESCRIPTION
This module provides a simple routine, C<proc_exists()>, to check a process'
existence by name, something that is commonly done in shell scripts using:
ps ax | grep name
pgrep name
and also some routines, C<find_*()>, to list processes matching some criteria.
=head1 VARIABLES
=head2 $Proc::Find::CACHE => bool (default: 0)
If set to true, will cache the call to C<Proc::ProcessTable>'s C<table()> so
subsequent invocation to C<find_proc()> or C<proc_exists> doesn't have to call
the method again. But this also means that the process check/listing will be
done on a past/stale process table.
=head1 FUNCTIONS
=head2 find_proc(%args) => \@pids (or \@procs)
Find process by name, PID, or some other attributes. Return an arrayref of
PID's, or an empty arrayref if none match the criteria.
Currently use L<Proc::ProcessTable> to list the processes.
Arguments:
=over
=item * filter => code
Filter by a coderef. The coderef will receive the process record (hashref).
=item * pid => int|array[int]|regex
Find by PID. Note that if you only want to check whether a PID exists, there are
cheaper methods (see L</"SEE ALSO">).
=item * name => str|array[str]|regex
Match against process' "name". Name is taken from the first word of the
cmndline, with path stripped.
If value is regex, will do a regex match instead of exact string comparison.
Example:
find_proc(name => "bash")
find_proc(name => qr/^(Thunar|dolphin|konqueror)$/)
=item * cmndline => str|array[str]|regex
Match against full cmndline.
If value is regex, will do a regex match instead of exact string comparison.
=item * exec => str|array[str]|regex
Match against program (executable/binary)'s path. If value does not contain a
path separator character, will be matched against program's name.
Example:
find_proc(exec => "perl") # find any perl
find_proc(exec => "/usr/bin/perl") # find only a specific perl
=item * user => int|str|array[int|str]|regex
List processes owned by specified user/UID.
If given a username which does not exist, will simply not match.
=item * uid => int|str|array[int|str]|regex
Same as C<user>.
=item * euser => int|str|array[int|str]|regex
List processes running as certain effective user/UID (will look against
C<euid>).
If given a username which does not exist, will simply not match.
=item * euid => int|str|array[int|str]|regex
Same as C<euser>.
=item * inverse => bool
If set to true, then will return all processes I<not> matching the criteria.
=item * table => obj
Supply result from C<Proc::ProcessTable> object's C<table()>. This can be used
to reuse the C<table()> cached result instead of repeatedly call C<table()> on
every invocation.
See also C<$Proc::Find::CACHE>.
=item * detail => bool (default: 0)
Instead of returning just the PID for each result, return a hash (record) of
process information instead. Currently this is just the entry from
C<Proc::ProcTable> object's C<table()> result.
=back
=head2 proc_exists(%args) => bool
Shortcut for:
@{ find_proc(%args) } > 0
=head2 find_all_proc(\%args, \%args2, ...) => \@pids (or \@procs)
Given multiple criteria, perform an AND search. Will only call
C<Proc::ProcessTable>'s C<table()> method once.
# find all processes matching mutiple criteria (although the same thing can
# also be accomplished by find_proc() and combining the criteria)
find_all_proc([{name=>'mplayer'}, {cmndline=>qr/mp3/}]);
=head2 find_any_proc(\%args, \%args2, ...) => \@pids (or \@procs)
Given multiple criteria, perform an OR search. Will only call
C<Proc::ProcessTable>'s C<table()> method once.
# find all processes belonging to either user
find_any_proc([{user=>'ujang'}, {user=>'titin'}]);
=head1 HOMEPAGE
Please visit the project's homepage at L<https://metacpan.org/release/Proc-Find>.
=head1 SOURCE
=head1 BUGS
Please report any bugs or feature requests on the bugtracker website L<https://rt.cpan.org/Public/Dist/Display.html?Name=Proc-Find>
When submitting a bug or request, please include a test-file or a
patch to an existing test-file that illustrates the bug or desired
feature.
=head1 SEE ALSO
L<Proc::Exists> can be used to check if one or more PIDs exist. If you are only
concerned with POSIX systems, you can just do C<kill 0, $pid> to accomplish the
same.
B<pgrep> Unix command.
=head1 AUTHOR
perlancar <perlancar@cpan.org>
=head1 COPYRIGHT AND LICENSE
This software is copyright (c) 2019, 2015, 2014 by perlancar@cpan.org.
This is free software; you can redistribute it and/or modify it under
the same terms as the Perl 5 programming language system itself.
=cut