Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:69768
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain adamharvey.name designates 209.85.223.174 as permitted sender)
MIME-Version: 1.0
Sender: adam@adamharvey.name
In-Reply-To: <52664C58.3020901@ajf.me>
References: <CAGa2bXaRkv2asU1bAwXTe=9s7o2njsqxG8W8gE9OtcG=8G4qqg@mail.gmail.com>
 <8C33E1D9-8689-4E81-A79B-644CB690DB64@gmail.com> <CAGa2bXbkcn+hN5k=MeeOR3MaOn2kW_M_sKkcwfnOxhng8=gc8Q@mail.gmail.com>
 <CAGa2bXYONaoxFBbmDx_vXQCpWSxPfLUsQJDKoQ3OCo1gDTcBwQ@mail.gmail.com>
 <CAGa2bXYjLhaPUkDYWH+iRGeQuNvtCEwSYDn74uDNtyKzBWMs3w@mail.gmail.com> <52664C58.3020901@ajf.me>
Date: Tue, 22 Oct 2013 10:11:00 -0700
Message-ID: <CADajX4ACagw=F_sZnesP3uRj6s=+8kL0xahffMYBemckX4GMEg@mail.gmail.com>
To: Andrea Faulds <ajf@ajf.me>, Yasuo Ohgaki <yohgaki@ohgaki.net>
Cc: PHP internals <internals@lists.php.net>
Content-Type: text/plain; charset=UTF-8
Subject: Re: [PHP-DEV] [VOTE] Change crypt() behavior w/o salt
From: aharvey@php.net (Adam Harvey)

On 22 October 2013 02:58, Andrea Faulds <ajf@ajf.me> wrote:
> On 22/10/2013 07:10, Yasuo Ohgaki wrote:
>>
>> Any comments patch for this RFC?
>> Better E_NOTICE message is welcome.
>
> I'm a native English speaker, how about "Calling crypt() without giving a
> salt will not produce strong password hashes."?
>
> It doesn't necessarily say you will produce a strong hash with it (other
> factors are at play), but it does say that you can't without it.
>
> Perhaps "secure" might be better than "strong".

I think I'd prefer the wording to be a little stronger, since this is
going to be shown when the user has actually done that. How about:

"Generating an insecure weak hash as no salt was given: please ensure
the salt parameter is specified and uses a strong hash type in order
to generate a cryptographically secure hash"

On the bright side, at least php_error_docref() will ensure there's a
link to the crypt() manual page in most setups. Rereading that, we may
actually want to be slightly more opinionated there about which hash
types are good and which are bad (it's not at all obvious that the DES
and MD5 types shouldn't generally be used).

Adam