@daya Yes, but you still need to put the variables somewhere. As I said, enter does more than mov ebp, esp.
And you might think to just put the struct on the heap and return the pointer, like any sane person would do it, because then you don't need to memcpy() 2MB every time and can just return a pointer via register.
PHP 7.4 is nearing release... a few weeks to (at most) a couple of months... New feature is Foreign Function Interface, which gives PHP direct access to .so (.dll) files... I foresee a lot of segfault questions on SO... but I also see the potential for exploiting buffer overflows and underruns in PHP code; something that hasn't been much of an issue yet.
I know my peers who write PHP code haven't been known for being security conscious... I'm honestly hoping that FFI will prove to be so newbie unfriendly that it never really catches on.
@Ghedipunk I forsee a future where exploiting PHP applications will be a lot easier. I'm glad they make that change
You know, I have been known to rant on and on about how much I dislike PHP, but at the end of the day, I love it for just how much of a false sense of security it gives you. As a pentester, I love nothing more than for some company to tell me that their code is sooooo secure, and it breaks apart as soon as I do ?foo[]=a&foo[]=b everything breaks apart and their server commits sudoku.
Personally terrifies me... but honestly, it makes me glad that python has never reached the level of usage for networked applications that PHP has, since python has had FFI for several years...
You know what I truly think is the reason why so much PHP code is so broken?
Because the official PHP docs are an absolute nightmare
So everyone just goes do some random blog, written by people who also barely know PHP
It's like the blind leading the blind, but instead of banging their toes against the cupboard repeatedly, they build terrible webapplications.
And if you look at it from a security perspective...well...there is none
The few times you actually see anyone talk about security you get copy/pastes with code that was considered broken 2005, where people tried to manually escape SQL code
Good point. I've been railing against those random blogs that use mysql_* and "INSERT INTO $_POST['foo']"... but if php.net actually taught the language AND web development... well...
Documentation needs to be split, both into hands-on documentation and related high-level usage guides.
But the way it is now is that function X tells you "My return types are either false or a string", and then function Y states that some other function may return NULL in case of something weird happening
People love saying that PHP is so weird because it developed over time. Well, the same happened to their docs, and it's time to fix it
And all this talk about how some information is technically available on some other page doesn't mean shit if the developers who actually read the docs don't know this!
I'm about 4 hours past when I should be unconscious... Honestly, realizing that it isn't the random ad-laden blogs teaching obvious security vulnerabilities are the symptoms of the problem, rather than the root, is a revelation that in my sleep deprived state, I feel I need to address, but know I'm not capable of doing right now.
I am sometimes overcome with incredible sadness that I will probably not be part of the generation that has permanent habitation outside of earth, but none of this outweighs the tragedy of not being able to be in the moon pool.
I suppose the first step to updating PHP.net's documentation would be to subscribe to the developer mailing lists, then lurk for a while to figure out how to get changes in place. In the meantime, start a wiki or git repo to write articles.
Even if the articles are never adopted by The PHP Group, it'll still be good to have a collaborative effort to document PHP in a sane way, where people can be pointed to first.
