UPDATED:
I edited Bcrypt.pm as below, and I can confirm this allows me to verify $2a$ or $2y$ bcrypt passwords.
=================
# diff -Naur /usr/lib64/perl5/site_perl/5.8.8/x86_64-linux-thread-mult
+i/Crypt/Eksblowfish/Bcrypt.pm.orig /usr/lib64/perl5/site_perl/5.8.8/x
+86_64-linux-thread-multi/Crypt/Eksblowfish/Bcrypt.pm
--- /usr/lib64/perl5/site_perl/5.8.8/x86_64-linux-thread-multi/Crypt/E
+ksblowfish/Bcrypt.pm.orig 2014-12-16 05:50:54.000000000 +0000
+++ /usr/lib64/perl5/site_perl/5.8.8/x86_64-linux-thread-multi/Crypt/E
+ksblowfish/Bcrypt.pm 2014-12-16 05:23:46.000000000 +0000
@@ -153,7 +153,7 @@
sub bcrypt($$) {
my($password, $settings) = @_;
croak "bad bcrypt settings"
- unless $settings =~ m#\A\$2(a?)\$([0-9]{2})\$
+ unless $settings =~ m#\A\$2([ay]?)\$([0-9]{2})\$
([./A-Za-z0-9]{22})#x;
my($key_nul, $cost, $salt_base64) = ($1, $2, $3);
my $hash = bcrypt_hash({
=============
So assuming I pull a $2y$10$ hashed password from the database, I can just s/^\$2y\$/\$2a\$/ and verify it using Crypt::Eksblowfish::Bcrypt, is that I'm hearing?
If that is the case, I support I can hack up the module itself and allow $2a or $2y without throwing an error.
| [reply]
[d/l] |
| [reply] |