Hacker News Reader

Comment by VMG | Hacker News Reader

Putting passwords in the php file? That's not a seasoned developer in my mind.

I'm going to disagree with you. The problem with PHP is that there's no easy way of sharing state between requests.

If you have a separate (ini-style) configuration file, every time you get a new request, the file will have to be read in from disk and parsed. On heavily-loaded web servers this can be a significant performance issue.

Configuration stored in a PHP file will be cached by your opcode cache and so doesn't incur any per-request parsing/reading overhead.

The problem here is not that Tumblr stored their configuration in PHP. The problem is their lack of testing their changes.

(As I mentioned elsewhere in the thread, this particular nasty issue can be solved by simply enforcing that every .php file begins in '<?'.)

> If you have a separate (ini-style) configuration file, every time you get a new request...

it will just pull it from the cache, not needing to be parsed or read from disk.

Configuration stored in a PHP file is bad. Changing working code simply because you want to add a new slave is a horrible idea. It would be like having to recompile Apache from source every time you want to add a new vhost because all the entries are written in C.

Interpreted languages like PHP, Python, Perl, etc all make it easy to put config options directly into the language. When your app is small, it's fine. But as it gets larger, move it out.

But yes, that they didn't even catch this is in testing is the real problem.

> it will just pull it from the cache, not needing to be parsed or read from disk.

I don't believe PHP's ini_read caches anything. Sure, your OS cache is going to have that file in, but PHP still has to do the fopen(), fread(), parse, fclose() dance on every page hit. This does not happen with an opcode-cached source file.

I have actually benchmarked this. Amortized across billions of page hits per month it produces a notable saving.

Kudos points out that you could use APC - this is absolutely true but IMO it basically equates to doing the same thing in a more complex way.

> Configuration stored in a PHP file is bad. Changing working code simply because you want to add a new slave is a horrible idea. It would be like having to recompile Apache from source every time you want to add a new vhost because all the entries are written in C.

This is a false comparison between interpreted and compiled languages. Not to mention that plenty of C programs use #defines for configuration. Varnish even translates its configuration language into C and dynamically loads it as a module, which provides significant flexibility.

In a dynamic language, and where you trust the person doing the configuration, there's no reason why your configuration shouldn't be in a source file. Both Django and Rails do it this way.

"In a dynamic language, and where you trust the person doing the configuration, there's no reason why your configuration shouldn't be in a source file. Both Django and Rails do it this way."

Not true. In Rails, sensitive information like database passwords or AWS keys are not stored in the source, you store them in configuration files, YAML files by default. It's also widely recommended that these not be checked into version control.

My bad - it's been a while since I touched Rails and my memory was rusty.

> I don't believe PHP's ini_read caches anything.

I wasn't referring to straight PHP. There are multiple ways to cache the config file's contents so you don't have to hit the disk to read.

> This is a false comparison between interpreted and compiled languages. Not to mention that plenty of C programs use #defines for configuration.

Your misunderstanding me here. I'm not saying #defines are bad. However, they are a different level of configuration from vhosts. You aren't #defining your virtual hosts. Same with php. You have the C code with specific options being set, and then php.ini for the user facing stuff.

> In a dynamic language, and where you trust the person doing the configuration, there's no reason why your configuration shouldn't be in a source file.

Except we saw at least one reason today.

> I don't believe PHP's ini_read caches anything.

so check if the ini file key is in memcache, if so get it from there, if not, ini_read it and populate the cache. Not hard.

Sure there is, APC. But that aside, Linux won't read that file every time, it will end up living in the disk buffers that it maintains.

The disk buffers make the operation cheap but not free, i.e. keeping the information cached is the right thing to do.

If that is a big problem, pass secrets as environment variables. These are getting passed and read anyway.

A huge mistake, but Tumblr's staff are probably quite experienced, nonetheless.

How about that gotcha though? A single character is mistakenly changed from "<" to "i" and that exposes the source code to the browser. Think about that.

I wonder if the 'i' has anything to do with vim

Clearly they should ban vim from their development process! ;-)

No, joking aside, the normal way to handle this is to have the perimeter server (nginx or varnish) catch any 5* responses and turn them into a user-friendly error-page. That way you never expose sensitive stack traces to your users.

So, this is standard stuff and easy to fix. However who of us hasn't screwed up on a similarly trivial issue before? I wouldn't judge them too hard on this one, happens to the best of us.

Clearly they should make vim mandatory on every machine;-) I guess the 'i' comes from someone who is used to use vim but in this case using another editor.

That wouldn't have caught this though, this code wasn't even executed, just sent as plain text.

Nginx/varnish can also catch Content-type: text/plain (or text/php?) and treat that like a 5xx response. Yes, that's a bit of a hack, but unless tumblr needs to serve text-files (which I doubt) it's a legit safeguard.

I know you were kidding, removing all text editors from your production servers might be a good idea

This is my guess. Anyone who's used VIM for years knows that the first couple of times using a different editor, your text is littered with "i" and ":w".

It could simply come from forgetting to exit from insert mode, and then coming back thinking the editor is in normal mode.

That was my first thought as well, as a long time vim user I kind of cringed.

Well, a single character change + a developer not testing their code locally + not running high level unit tests + deploying untested code to production.

Tumblr has at least one full time sysadmin. I think it's more reasonable to suppose he made the change to a production config file, rather than a developer. A definite failing in their deployment process, but it really says nothing about the quality of the application code or developer testing.

It's needed to connect to databases with authentication. What would you suggest?

I believe the way to do it "correctly" would be to put the actual file containing the credentials in a non-public location and just do an include where you need to access it. At least that's how I do it. I could be wrong...

If one would made a similar typo in that file, its contents will be displayed too.

A solution would be to have a .ini-like (or some other simple-to-parse format) config file and PHP code to read its contents. PHP code could be leaked, but config file contents wouldn't.

Yeah seen a few client libraries that are doing that. Even putting them in a config only file lessens the chance of something like this happening as it is modified a lot less often.

Deleted comment

I believe you are wrong. include()/require() would equally happily display any file's contents outside of `<?php ... ?>` scope (the case with "i?php"), within document root or not.

Edit: I've tested this:

    $ php -v
    PHP 5.2.6-3ubuntu4.6 with Suhosin-Patch 0.9.6.2 (cli) (built: Sep 16 2010 19:51:25) 
    Copyright (c) 1997-2008 The PHP Group
    Zend Engine v2.2.0, Copyright (c) 1998-2008 Zend Technologies

    $ cat test.php
    <?php
        require "/tmp/test2.php";
    ?>
    $ cat /tmp/test2.php
    i?php
        define("TEST", "test");
    ?>
    $ GET http://localhost/test.php
    i?php
        define("TEST", "test");
    ?>

Off course, but in this case, passwords would only be exposed if the config file had a miss-typed opening PHP tag. If "test.php" had it, you wouldn't be able to see the contents of "test2.php".

Yes, you are right. And in this exact case they (mis-)edited the file, that contained passwords (i.e. test2.php in my improvised example).

Sure, I was replying for this hypothetical situation that you guys ware discussing, where they would store passwords in a different file outside of webroot ...

it's always included to the document root (the bootstrap file). it doesn't matter where you're including from, a broken open tag would cause errors like this one.

If you're doing an include, wouldn't the passwords still be in a PHP file?

if the php file is printed rather than executed, the include will not be followed. You'd see the "include /path/to/inaccessible/file" but you wouldn't see the passwords within the include.

I think. I haven't seriously used php since 2003 or so.

Yes, but included files still need a <?php tag at the top, or PHP just prints them out.

That wouldn't help in this case, since a typo in the config file is making the file not be parsed as PHP.

If you want to avoid that possibility, you can use a non-PHP format (eg: YAML) and parse it.

Client SSL certificates? I don't know what DB they're using, but at least Postgres and MySQL support certificate-based auth.

How is a client certificate different from a long random password? Both have to be stored somewhere. The correct solution is to keep this (and everything else) out of your web root.

If that i?php blunder happened to me, my users would see 10 lines of code. One include for the framework (which lives outside the web root), plus three calls to get the framework to handle the current request.

Those 10 lines would be totally unproblematic

.pem file won't be leaked in this ("i?php") manner. Only file's location would be revealed.

Of course, you are completely right, it should still be inaccessible from the web.

Typically, these files are not web accessible. Problem solved.

This bug was nothing to do with the file being web accessible. If you miss the opening <?php tag out in any file, regardless of whether it's in the web root, it will get printed straight out to the browser.

But you would never `include()` a PEM file.

No, better words would be: "Putting passwords in php file under doc root, why not keep it outside doc root?"

Good point. Minus the keys/passwords the "harm" would be limited.

Reply on news.ycombinator.com