phc-discussions - Re: [PHC] Specification of a modular crypt format (2)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Wed, 30 Sep 2015 07:12:04 +0300
From: Solar Designer <solar@...nwall.com>
To: discussions@...sword-hashing.net
Subject: Re: [PHC] Specification of a modular crypt format (2)

On Tue, Sep 29, 2015 at 02:55:48PM +0200, Thomas Pornin wrote:
> Strictly speaking, the decoder tries to be strict when strictness is
> cheap. If we want a strict decoder of numerical values, strtoul()
> won't be enough; we would need something like this:

That's true, and your code works, but I gave strtoul() a try anyway.
This code appears equivalent to yours, as per the attached program
(which also happened to test yours).

static const char *
decode_decimal_solar(const char *str, unsigned long *v)
{
	char *end;

	errno = 0;
	*v = strtoul(str, &end, 10);
	if (errno ||
	    str == end ||
	    end - str != strspn(str, "0123456789") ||
	    (*v == 0 && end - str != 1))
		return NULL;
	return end;
}

It's quite a bit shorter than yours, but it relies on correct behavior
of two libc functions, so it might be less reliable for that reason.

I think we should actually go with manual parsing, like you did.  I just
wanted to see how bad strtoul() is at this, and how many additional
checks are needed to repair it.  BTW, the comparison against strspn() is
needed to detect strtoul()'s ignoring of leading whitespace, without
having to enumerate all whitespace chars (' ', '\n', any others?); it
also happens to take care of '-', letting me omit that check.

Alexander

View attachment "decimal.c" of type "text/x-c" (2537 bytes)