On Jul 31, 2009, at 3:33 AM, olivier wrote: >> >> I'd prefer some less-predictable salt than the suggestion below, >> though. How about the old Unix passwd trick of choosing a some random >> salt, and appending the salt in plaintext to the hash? > > we could do that, without braking backward compatibility i think... > - store password in the following format: hashtype$salt > $passwordhash (with '$' separators, pretty standard stuff...) > > to authenticate user: > - add a special case: if no hashtype: passwordhash -> md5$ > $passwordhash > - use hashtype(password + salt), for most of the standard hashtypes > (md5, sha1, sha256, ...) > (- do other weird stuffs... like add a special hashtype bottiger, > bottiger$$passwordhash will use the function md5( f(password) ) , f > being a deterministic function ) ;-) > > we could add a special attribute for migration. if set to True, then > storage is migrated from the old format to the new format when user > logs in > the length of the salt can also be a parameter (if you really want to > play with some trade-off between storage/security, bad idea) > > i can work on that, if it makes sense to you guys
It makes sense to me. We might also want to require a strong password by default (if we're not already), which would also be backward compatible. I was writing up a separate message on tightening up password security generally; I'll send it shortly. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "web2py-users" group. To post to this group, send email to web2py@googlegroups.com To unsubscribe from this group, send email to web2py+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/web2py?hl=en -~----------~----~----~----~------~----~------~--~---
