Tobias Schlitt has a great (long) post responding to a proposal made on the php.internals mailing list for the inclusion of "taint mode" in upcoming PHP versions.

Tobias starts with an overview of what "taint mode" is so that everyone's on the same page. His example involves the incoming and outgoing data usually involved in an application and how the incoming can be the most problematic when it comes to the values inside. This is where taint mode can come to the rescue. At its most basic level, it's a method for, assuming everything coming in is "tainted", creating a mechanism to automatically clean the data before it's even touched.

With the basics down, Tobias looks back to the proposal for a few additional comments. He considers the proposal a great way to introduce the functionality to the language. There are some draw backs he mentions, though, including the additional overhead of working through every inputted value.

Overall, I think this whole thing would be a great addition to PHP and I hope this could come for 6.0. What do you think?

You can also check out some other opinions on the matter:

• "Tainted mode" pour php on the Kamelot Blog

In an effort to get some thought going about ways to encourage security in PHP applications, Stas has posted an idea about a simplified php.ini setting - production=On.

His idea is that, with this setting on, the PHP installation would:

• disable display errors
• disable phpinfo()
• turn expose_php off
• make max_execution_time/memory_limit reasonable
• and possibly a few others that some developers forget to set correctly

On the PHP Security Blog, there's two new posts concerning their Suhosin patch for PHP - one talking about a remote code execution vulnerability found in the uploadprogressmeter extension (which as already been corrected) and the other dealing with the "stealth mode" of the patch to find compatibility problems.

As mentioned, the issue with the uploadprogress extension has been fixed and as been introduced into the PECL repository. The other post just mentions what the patch is doing to try to keep compatibility problems from happening due to back extensions and/or modules being loaded in an application.

More details about the Suhosin patch can be found here.

A new release of the Zip PECL package has been made according to this post on Pierre-Alain Joye's blog today. The main update in this release is to counteract a Windows bug that's interfering with binary file opens.

The issue is actually a windows bug. No matter if I give or not the "b" flag to fopen, the write operations are not binary safe. It seems to be a known issue as many projects use the same trick.

The problem comes up when PHP forces the binary mode in SAPI and CLI, making the binary writes to a file non-binary safe no matter what. Pierre is also asking for help from anyone out there with any information/bug reports/references about this issue that would yield something useful.