Re:Padrino Framework 0.11 is here
by ronnqvist on Saturday March 23, 2013 @03:52AM
(#5427562)
Thankyou so much! I've been waiting for this release in order to get Twitter bootstrap for a project that I'll just have to get out the door! Give me a bitcoin address and I'll tip you one. :)
Re:Padrino Framework 0.11 is here
by pspeter3 on Friday March 22, 2013 @01:25PM
(#5424277)
I used to use padrino for my projects. I'm one of the few people who actually finds thinking in callbacks easier and I like using the same language in both the client and server so I switched to node. I think you guys are a fantastic project though!
Re:Padrino Framework 0.11 is here
by tekacs on Friday March 22, 2013 @01:50PM
(#5424435)
Whilst I think that the Padrino framework is an excellent effort, I would like to remind everyone (chiefly the project organisers) that as things stand, the framework has no meaningful notion of security for the admin interface.
I flagged this up about two years ago, now, but stopped using the framework shortly thereafter and never pushed any fixes of my own. The complete lack of meaningful authentication appears to still be there all this time later.
Though I would hope that Padrino admin isn't in fact being used in production by anybody who hasn't read the source, a quick search at the time showed otherwise.
Re:Padrino Framework 0.11 is here
by tekacs on Friday March 22, 2013 @01:59PM
(#5424497)
Huzzah, with regards to the point of BCrypt (presumably the second link goes to currently dead, but left-in-the-codebase helpers).
On the other hand, the first (original) point (the GitHub issue) still stands to this day - it would make sense to either improve the default implementation of Auth to be usable (as per most every sane framework out there) or to provide a warning somewhere (at least on the Padrino guide for admin & auth?) would it not?
Re:Padrino Framework 0.11 is here
by DAddYE on Friday March 22, 2013 @02:03PM
(#5424529)
Tekacs,
you are right, current auth is the most simple thing we have in mind. For sure can be better and so we planned to follow more an API standard so to be interchangeable with others auth systems.
That helper is 'dead'. Will be removed before 1.0.
If you have some others suggestions or you think there is something insecure, please share your thoughts. In this release we did a lot of stuff to improve that area, so you are welcome!
Re:Padrino Framework 0.11 is here
by tekacs on Friday March 22, 2013 @02:10PM
(#5424563)
I would be very glad to hear that a session_secret is being used to encrypt session cookies (which would prevent the (working-at-the-time) 'attack' I posted in the abovementioned issue).
I can't tell at a glance whether this change was made after the issue and if the problem is now overcome, but I'm glad to hear that security is at least on your minds! Just switching to BCrypt from the old DES'd passwords is a wonderful step!
Re:Padrino Framework 0.11 is here
by DAddYE on Friday March 22, 2013 @02:13PM
(#5424589)
Thanks! BTW, I'm far to be that is 'secure' nothing is 'secure', but we are putting a lot of efforts in that way, crf tokens, rack-protection and so on ... so jump in the community and help us to improve security aspects.
When I saw you guys at the meetup in SF, there was some chatter about concurrency features being added, did these make it into this release?
I flagged this up about two years ago, now, but stopped using the framework shortly thereafter and never pushed any fixes of my own. The complete lack of meaningful authentication appears to still be there all this time later.
For reference, see:
https://github.com/padrino/padrino-framework/issues/384
and
http://tkcs.in/Nlrc
Though I would hope that Padrino admin isn't in fact being used in production by anybody who hasn't read the source, a quick search at the time showed otherwise.
1) We use bcrypt: https://github.com/padrino/padrino-framework/blob/master/pad...
2) Auth is an API we generate that for YOU, but you can change the code generated to fit your needs.
On the other hand, the first (original) point (the GitHub issue) still stands to this day - it would make sense to either improve the default implementation of Auth to be usable (as per most every sane framework out there) or to provide a warning somewhere (at least on the Padrino guide for admin & auth?) would it not?
you are right, current auth is the most simple thing we have in mind. For sure can be better and so we planned to follow more an API standard so to be interchangeable with others auth systems.
That helper is 'dead'. Will be removed before 1.0.
If you have some others suggestions or you think there is something insecure, please share your thoughts. In this release we did a lot of stuff to improve that area, so you are welcome!
I can't tell at a glance whether this change was made after the issue and if the problem is now overcome, but I'm glad to hear that security is at least on your minds! Just switching to BCrypt from the old DES'd passwords is a wonderful step!
Best of luck and thanks for the great project!
Thanks for all!