Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:61934
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Received-SPF: pass (pb1.pair.com: domain gmail.com designates 209.85.215.170 as permitted sender)
Message-ID: <50197136.7040401@gmail.com>
Date: Wed, 01 Aug 2012 20:11:02 +0200
User-Agent: Thunderbird
MIME-Version: 1.0
To: Peter Lind <peter.e.lind@gmail.com>
CC: Anthony Ferrara <ircmaxell@gmail.com>, internals@lists.php.net
References: <CAAyV7nFRT24rksh6iDWxstfQwJ4+KdEC7Zs6EC8wBtJ8Zorf7w@mail.gmail.com> <4FFF1831.8070902@sugarcrm.com> <CAAyV7nHi0k6S-zbi5=xCE8EmXpaEds9RoZTB8RPOeO8K_ZLY=w@mail.gmail.com> <005101cd6f18$9da38510$d8ea8f30$@com> <CAAyV7nGOiVD0=z+6nJKAhr6qu96GuU4+nBTTQiizcP9DxNRUUg@mail.gmail.com> <009401cd6f28$b71c69c0$25553d40$@com> <CAAyV7nH8w5j+AQYQqh0O5eA78tjrvEzAAvKkPNPNrRFhBvDxvQ@mail.gmail.com> <00b701cd6f35$b2d621a0$188264e0$@com> <CAAyV7nHr532Bi-K0k99wSSg5RgJYLGAdusMqCYt9pRRr6LqscA@mail.gmail.com> <CAEU6NAeQsnDr0uDKMLU3F9SE=dopaiELGozkPX8zOj9CsGFagQ@mail.gmail.com> <CAAyV7nGPnFtfDQY4UNdon0bbEZwRcBsejygvASDek15SjiPYvg@mail.gmail.com> <CAEU6NAcbB+wV=2YMotFQL-E0Jhx-S4MGQ5sgc64gmNo9VJxhMw@mail.gmail.com>
In-Reply-To: <CAEU6NAcbB+wV=2YMotFQL-E0Jhx-S4MGQ5sgc64gmNo9VJxhMw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] [PROPOSED] password_hash RFC - Implementing simplified
 password hashing functions
From: keisial@gmail.com (=?UTF-8?B?w4FuZ2VsIEdvbnrDoWxleg==?=)

On 31/07/12 22:20, Peter Lind wrote:
> On 31 July 2012 22:02, Anthony Ferrara <ircmaxell@gmail.com> wrote:
>> Peter,
>>
>> On Tue, Jul 31, 2012 at 3:46 PM, Peter Lind <peter.e.lind@gmail.com> wrote:
>>> On 31 July 2012 18:21, Anthony Ferrara <ircmaxell@gmail.com> wrote:
>>>
>>> *snip*
>>>
>>>> Also, be aware that BCrypt only uses the first 72 characters of the
>>>> password field. So if you use a hex encoded sha512 output, a good deal
>>>> of
>>>> entropy would be lost (almost half of it)...
>>>>
>>> Seeing as the hashing function will default (at first, at least) to
>>> bcrypt, would it be possible to add a warning if it's given an input
>>> longer than 72 chars? Preferably make the function context-aware so
>>> you don't get the same warning if using sha512. Otherwise I predict
>>> that someone will do:
>>>
>>> $hash = password_hash($my_128_char_pepper . $password, PASSWORD_DEFAULT);
>>>
>>> Which obviously renders the hashing useless, as you'll be hashing the
>>> same 72 chars over and over again. Which, currently, crypt() let's you
>>> get away with without as much as a hiccup.
>> That's actually a very good idea.
>>
>> I'm curious though. Should we warning? Or should we sha512 hash (to bring
>> down to 64 chars)...  That's something I think would be worth reaching out
>> to the crypt() maintainers for advice...
> I'd be wary of not doing what people actually expect - i.e. hash the
> provided string with the specified algo. Better to educate the users,
> I'd say, through a warning. Anyway, as you said, would be good to get
> other opinions on this.
>
> Regards
> Peter
I'd go with the hashing. I'd would however produce a slightly different
prefix
than with bare bcrypt.
$password is user-provided. The library client shouldn't need to know
the hash
internals and -depending on that- truncate the output.
There may be a $option to disable this additional hashing, though.