Re: Secret URLs

2010-01-05 Thread tekomp
Maybe this will help: http://bakery.cakephp.org/articles/view/encrypting-urls On Jan 1, 11:54 am, anka wrote: > Hi @all, > > I have a (almost) simple question! Is it possible to make all cake php > REST URLs unreadable? What I mean is, that a given REST URL like ".../ > user/delete/10" is very si

Re: Secret URLs

2010-01-05 Thread j0n4s.h4rtm...@googlemail.com
Hi again, I am open to learn, so please tell me, whats the benefit of url obscurity if not protecting against CSRF. A large german social network uses that to protect against CSRF - though they do not use UUIDs. Just links like domain.tld/long_random_looking_stuff_here, it works against mining too

Re: Secret URLs

2010-01-04 Thread euromark
well the UUIDs do help for his problem of "being able to guess" the url i like UUIDs as you dont tell how many items have been added (this can be important in some cases) but of course it has nothing to do with CSRF delele should NEVER be done by a GET-Request anyway GET never alters the database

Re: Secret URLs

2010-01-03 Thread j0n4s.h4rtm...@googlemail.com
Well using UUIDs would result into /users/view/LONG-UU-ID-HERE Which would not resolve Cross Site Request Forgery. Use SecurityComponent and use form buttons that use method = post. That way "destructive" links get caught if no form security token was found. I do it like this: http://github.com/i

Re: Secret URLs

2010-01-03 Thread euromark
just use UUIDs that "feature" comes with it automatically On 3 Jan., 08:41, Walther wrote: > Not really possible afaik. > > But, security through obscurity is not the solution. You should rather > spend the time writing code that will make sure that the user/referer > has permissions to perform t

Re: Secret URLs

2010-01-02 Thread Walther
Not really possible afaik. But, security through obscurity is not the solution. You should rather spend the time writing code that will make sure that the user/referer has permissions to perform that action. On Jan 1, 9:54 pm, anka wrote: > Hi @all, > > I have a (almost) simple question! Is it p