Re: Run-time taint support proposal

On 15-Dec-06, at 6:11 PM, Stanislav Malyshev wrote:

All it means is extra work for developers with little or no tangible benefits. I also wonder how taint will work with the standard remove/add

Security is benefit. Of course, the developers that are sure they write secure code anyway need not be bothered by tainting and can leave it off forever.

So you claim that without taint mode it is not possible to write safe PHP code?

The job of a language is to provide tools, not arbitrary crippling limitation under the guise of security improvement.

I agree. Tainting is one of such tools, aimed at improving security.

Tainting is a false security it makes you feel secure, when you aren't. First its off in production and that's where all the hacks appear, it will have holes due to unforeseen function usage, dynamic variables, false untainting etc...

safe_mode sounded like a really reasonable idea too, I would've hoped some lessons from past mistakes could be made.

I do not see what exactly you propose to learn from safe mode mistakes - that we should never try to improve PHP security by providing language level tools? I do not see how this can be derived from whatever was wrong with safe mode. It may be that the tainting would not catch but I do not think safe mode problems should prevent us from even trying.

Good luck, I suppose on a base level it is entertaining seeing someone bang their head against the wall time and time again. Ilia Alshanetsky