[Seaside] Deployment question: Anyone using modSecurity (or
equiv) to ensure hackers keep out of Seaside?
philippe.marschall at gmail.com
Mon Feb 16 19:09:36 UTC 2009
2009/2/16 Randal L. Schwartz <merlyn at stonehenge.com>:
>>>>>> "Rick" == Rick Flower <rickf at ca-flower.com> writes:
> Rick> So.. In reviewing the issue involved with this it appears that something
> Rick> like ModSecurity (see links below) and rules from GotRoot.com might help
> Rick> prevent this sort of thing from happening and was curious if anyone
> Rick> running Seaside/Apache combinations has gone down this path to ensure
> Rick> naughty things don't get passed into Seaside if possible.. Obviously I
> Rick> realize that PHP != Smalltalk and that exploits could be different but
> Rick> I'd like to reduce the chances as best I can.
> Since Seaside doesn't have any of the things listed at
> Variable-length buffer injection
> Meta character injection
> SQL injection
> SOAP fault code disclosure
> I'm not sure why you think modsecurity would help.
> Assuming you're using GLORP for most of your sql, and follow best practices
> for the small stuff that you would handwrite, of course.
> The reason most PHP suffers from this stuff is that (a) PHP is written in C
> and (b) many PHP coders are notoriously bad at understanding metacharacters
> and (c) most PHP coders do not appear to understand SQL placeholders. And
> SOAP is just bad. :)
I have seen an arbitrary remote code excution vulnerability in Squeak
including SOAP. Except that there was nothing SOAP specific about it.
It included something as simple as converting a String to a Boolean.
Given the PostgreS Squeak driver doesn't support prepared statements I
would be very cautios about any database related statements especially
considering that various vulerabilities in PHP can only be solved by
passing the character set of the input to the driver or using prepared
statements. Glashaus und so.
More information about the seaside