NAME
App::bmkpasswd - bcrypt-capable mkpasswd(1) and exported helpers
SYNOPSIS
bmkpasswd --help
## Generate bcrypted passwords
## Defaults to work cost factor '08':
bmkpasswd
bmkpasswd --workcost='06'
## Use other methods:
bmkpasswd --method='md5'
# SHA requires Crypt::Passwd::XS or glibc2.7+
bmkpasswd --method='sha512'
## Compare a hash:
bmkpasswd --check=HASH
## Check hash generation times:
bmkpasswd --benchmark
DESCRIPTION
App::bmkpasswd is a simple bcrypt-enabled mkpasswd. (Helper functions
are also exported for use in other applications; see "EXPORTED".)
See "bmkpasswd --help" for usage information.
Uses Crypt::Eksblowfish::Bcrypt for bcrypted passwords. Bcrypt hashes
come with a configurable work-cost factor; that allows hash generation
to become configurably slower as computers get faster, thereby impeding
brute-force hash generation attempts.
See for more on
why you ought to be using bcrypt or similar "adaptive" techniques.
SHA-256 and SHA-512 are supported if available. You'll need either
Crypt::Passwd::XS or a system crypt() that can handle SHA, such as
glibc-2.7+ or newer FreeBSD builds.
MD5 support is fairly universal, but it is known insecure and there is
really no valid excuse to be using it; it is included here for
compatibility with ancient hashes.
Salts are randomly generated.
EXPORTED
You can use the exported mkpasswd and passwdcmp functions in other Perl
modules/applications:
use App::bmkpasswd qw/mkpasswd passwdcmp/;
## Generate a bcrypted passwd with work-cost 08:
$bcrypted = mkpasswd($passwd);
## Generate a bcrypted passwd with other work-cost:
$bcrypted = mkpasswd($passwd, 'bcrypt', '06');
## SHA:
$crypted = mkpasswd($passwd, 'sha256');
$crypted = mkpasswd($passwd, 'sha512');
## Compare a password against a hash
## passwdcmp() will return the hash if it is a match
if ( passwdcmp($passwd, $hash) ) {
## Successful match
} else {
## Failed match
}
BUGS
There is currently no easy way to pass your own salt; frankly, this
thing is aimed at some projects of mine where that issue is unlikely to
come up and randomized is appropriate. If that's a problem, patches
welcome? ;-)
AUTHOR
Jon Portnoy