I agree with Jonathan on this one but I would not oppose something like IS_EMAIL(strict=True)?
Massimo On May 30, 9:47 pm, blackthorne <[email protected]> wrote: > My point is... why do you use IS_EMAIL() validator? because you want > to know if that email address is valid or not. why? probably because > you want your app to send the user an email, so to prevent it not to > arrive to its destiny it becomes important to make sure, that email > address is valid. > If the validator considers 'a...@d_-f.us' invalid, I would call it > success because that is not a valid email address, probably a typo. > > Yes, we can never be sure that the inserted address is correct but in > many cases, we can determine with certainty, that the address is wrong > and prevent further errors. You probably agree with me that we should > do what we can to improve even if we can't solve a problem in all its > extent. If this is isn't obvious, I give you an example: all passwords > can be guessed and its hashes cracked, so what do you about it? do you > choose a short and simple password just because the system is not > perfect and all passwords can be guessed/all hashing mechanisms > cracked? don't think so, you choose a strong password, you do what you > can... > > how about making a IS_REAL_EMAIL()? > > On May 30, 7:16 am, mdipierro <[email protected]> wrote: > > > +1 > > > On May 30, 1:12 am, Jonathan Lundell <[email protected]> wrote: > > > > On May 29, 2010, at 9:39 PM, mdipierro wrote: > > > > > The one your propose fails two of our tests: > > > > Failed example: > > > > IS_EMAIL()('a...@d_-f.us') > > > > Expected: > > > > ('a...@d_-f.us', None) > > > > Got: > > > > ('a...@d_-f.us', 'enter a valid email address') > > > > Underscores are not allowed, but we accept them because they've been used > > > historically and there's no compelling reason (in a validator) to reject > > > them. If we were actually allocating domain names, there'd be a reason do > > > so, but we're not. > > > > We should also keep in mind that there's no way we can guarantee that the > > > address that gets typed in is correct (in the sense of not having a typo, > > > even if it's syntactically correct). Mainly, we want to check that it > > > "looks like" an email address, and err on the side of permissiveness. > > > > > ********************************************************************** > > > > File "gluon/validators.py", line 859, in __main__.IS_EMAIL > > > > Failed example: > > > > IS_EMAIL()('local...@localhost') # localhost as domain > > > > Expected: > > > > ('local...@localhost', None) > > > > Got: > > > > ('local...@localhost', 'enter a valid email address') > > > > ********************************************************************** > > > > > On May 29, 9:36 pm, blackthorne <[email protected]> wrote: > > > >>http://fightingforalostcause.net/misc/2006/compare-email-regex.php > > > > >> Take it as a suggestion for a better email regex validator...
