Ping and Social Overload

By · Published · apple, ramblings

Two days ago Apple announced Ping: a social network geared towards music sharing. And a bunch of iPods too. Personally, I was more excited by the new AppleTV (I have two of them and absolutely love them) but more on that later. This is about Ping.

My thoughts on **Ping:**

Apple's first real attempt at social networking reminds me of Google's countless attempts to get into the social networking space: they're like that guy that shows up to the party really late - I mean beyond fashionably late - when the party is already over and everyone else is already drunk and thinking about stumbling across the street to IHOP or Taco Bell. They say they were at the library studying and now they want to go out and drink, but the keg has floated, the bars and liquor stores are already closed and all you want to do is eat a burrito supreme and find some sofa to pass out on.

Ping is a good first start, but it has some problems:

  • What is the target here? Am I supposed to follow people or artists or both or what? And what are they supposed to do? All this feels like is Twitter or Facebook + iTunes. The people I'm following can share messages and pictures? Yep. Twitter in iTunes. I can like and share and post comments? Yep. Facebook in iTunes.

  • Why not allow independent artists into the fold? Some of my favorite artists (such as Matthew Ebel - check him out if you love piano rock) are independent. Right now there are like 10 artists you can follow, and that Lady Gaga is one of them makes me want to break something. The only ones on there I'm remotely interested in following is Dave Matthews Band and maybe U2.

  • I can't access it in any way other than in iTunes. No web access. While this means I can fire it up myself on my computer and laptop, and (currently) on my iPhone via the iTunes application, I can't check Ping at a friend's house. I can't go to the Apple store and check Ping. Everything has to go through iTunes, and this absolutely cripples it. Think that's overkill? Go to the Apple store and watch for  15 minutes how many people walk in and use one of the computers to check Facebook.

  • I can only "like" and "share" content I purchased from iTunes. I have purchased 58 songs from iTunes over the years, out of 3,621 songs in my library. About 1% of my library is available.

If Apple fixes these (and other, more minor) problems, Ping could be really cool. The problem is that these aren't code fixes. They're not something they can test and roll out a change for. These are conceptual problems relating to what their idea of Ping is versus the what the rest of the world is going to use it as. The question is, will they be Google and throw this out here, not maintain it and mercifully kill it a year later (a la Google Wave and the impending death of Google Buzz), or will they adapt and change it to better suit the needs of the public? Because that's the thing about social networking: you have to embrace the users thoughts, opinions, and ideas. It's a lesson digg just learned the hard way and a lesson that frankly, given Apple's reputation as wanting to control everything, I don't see them embracing.

As a side note, I will however salute Apple for not giving into Facebook if the rumor is true. Facebook plays fast and loose with people's information, and I really don't like how it seems to have become the de facto standard for social network usage (and thus the reason you can comment with your Facebook login). That, and Zuckerberg. I hate that guy.

Still, Ping is yet another player in this social networking space. A space that is becoming increasingly full ...

Social Overload

I'm already Facebooked, Myspaced and Twittered. I'm LiveJournaled, Wordpressed, and Youtube'd. I'm Flickr'd, LinkedIn'd, Vimeo'd,'d and Gowalla'd. I'm on any number of dozens of message boards and mailing lists that predate "Web 2.0" and the social networking "revolution," and I follow nearly 100 various blogs and other feeds via RSS. They're on my desktop, on my laptop, on my tablet and in my phone.

And now, apparently, I'm Ping'd as well. Le sigh.

Now, to be fair, I don't check all these sites. I last logged into Myspace about 9 months ago. I last used Gowalla about a year ago. I usually only look at Youtube, Flickr or Vimeo when I need something, and haven't updated a LiveJournal in about 3 years. But at what point does all this interaction - this social networking - become social overload?

Are any of these services adding value to my life? And at what point does a social network - Ping, in this case - simply become yet another thing I have to think about and check? Or will it become yet another service I sign up for, try for awhile and ignore?

Diffing, flattening and expanding multidimensional arrays in PHP

By · Published · php

PHP has functions that can compute the difference between two arrays built in. The comments sections for those functions are filled with people trying to figure out the best way to do the same thing with multidimensional arrays, and almost all of them are recursive diffing functions that try to walk the tree and do a diff at each level.

The problem with this approach are 1) they are unreliable as they usually don't account for all data types at each level, and 2) they're slow, due to multiple calls to array_diff at each level of the tree.

A better approach, I think, is to flatten a multidimensional array into a single dimension, make a single call to array_diff, then (if needed) expand it back out if you really need the resulting diff to be multidimensional.

Lets look at some code. The following recursive function flattens a multidimensional array into a single dimension.

function flatten($arr, $base = "", $divider_char = "/") {
    $ret = array();
    if(is_array($arr)) {
        foreach($arr as $k = $v) {
            if(is_array($v)) {
                $tmp_array = flatten($v, $base.$k.$divider_char, $divider_char);
                $ret = array_merge($ret, $tmp_array);
            } else {
                $ret[$base.$k] = $v;
    return $ret;

The following function (based on this function found here) reinflates the array back up after it's been deflated.

function inflate($arr, $divider_char = "/") {
    if(!is_array($arr)) {
        return false;

    $split = '/' . preg_quote($divider_char, '/') . '/';

    $ret = array();
    foreach ($arr as $key = $val) {
        $parts = preg_split($split, $key, -1, PREG_SPLIT_NO_EMPTY);
        $leafpart = array_pop($parts);
        $parent = &$ret;
        foreach ($parts as $part) {
            if (!isset($parent[$part])) {
                $parent[$part] = array();
            } elseif (!is_array($parent[$part])) {
                $parent[$part] = array();
            $parent = &$parent[$part];

        if (empty($parent[$leafpart])) {
            $parent[$leafpart] = $val;
    return $ret;

Now, with arrays in flat form, it's easy to use the built-in functions to diff:

$arr1_flat = array();
$arr2_flat = array();

$arr1_flat = flatten($arr1);
$arr2_flat = flatten($arr2);

$ret = array_diff_assoc($arr1_flat, $arr2_flat);

$diff = inflate($ret);

Hard Drive Upgrade

By · Published · apple, mac, osx

So Sunday night, my iMac died.

been having strange problems the few months leading up to it. Mostly random freezes. I always notice when they happen because I leave running all the time to filter my messages, so when my iPhone would start going crazy, I'd know it had crashed again. It actually happened while I was out of town in Atlanta earlier this year, so all weekend my phone was constantly buzzing.

Well, Sunday while we were working in the yard, I had set up a DVD rip job - my current project is digitizing all my DVDs for the AppleTV - to run, and while we were working it randomly reset itself and got all sluggish. That night, I tried to boot of the Snow Leopard DVD to run Disk Utility, and it couldn't even mount the drive and refused to repair it. Couldn't reboot either. I tried DiskWarrior, and that fixed things up enough to boot it, but it was REALLY SLOW (it took 10 minutes to boot). It was good enough to get the last few remaining files that hadn't been backed up yet onto the external drive. Then, I tried reinstalling, and it never came back. My conclusion, since I could still boot fine from the DVD, was dead hard drive.

The original hard drive was 500GB, but I figured I'd upgrade while doing this. Ordered a new 1TB hard drive via a deal at work and had it overnighted. It arrived yesterday. And, after some interesting surgery (who says you can't work on Macs!), got it installed, formatted, and Snow Leopard reinstalled.

You know, I remember the first computer I owned that crossed the 1GB barrier, back in late 1999. I guess I'll have to remember this one, too.

Scripting iTerm with AppleScript

By · Published · apple, applescript, mac, osx

Every day, when I get to work, there are a number of tasks I do. Among the first thing I do is connect to a number of servers via SSH. These servers - our development testing, staging, and code rolling servers - are part of the development infrastructure at dealnews.

So every morning, I launch iTerm, make three sessions and log into the various servers. Over time, I've written some helper scripts to make this faster. My "go" script contains the SSH commands (using keys) to log into these machines so that all I have to do is type "go rpeck" to log into my development machine.

Still, this morning, the lunacy of every morning having to open iTerm and execute three commands, every day without fail, struck me. Why not script this so that, when my laptop is plugged into the network at work, it automatically launches iTerm and logs me into the relevant services?

Fortunately, iTerm exposes a pretty complete set of AppleScript commands, so with a little work, I was able to come up with this:

tell application "System Events"
    set appWasRunning to exists (processes where name is "iTerm")

    tell application "iTerm"

        if not appWasRunning then
            terminate the first session of the first terminal
        end if

        set myterm to (make new terminal)

        tell myterm
            set dev_session to (make new session at the end of sessions)
            tell dev_session
                exec command "/Volumes/iDisk/bin/go rpeck"
            end tell

            set staging_session to (make new session at the end of sessions)
            tell staging_session
                exec command "/Volumes/iDisk/bin/go staging2"
            end tell

            set nfs_session to (make new session at the end of sessions)
            tell nfs_session
                exec command "/Volumes/iDisk/bin/go nfs"
            end tell

            select dev_session
        end tell
    end tell
end tell

What this little script does is, when launched, checks to see if an instance of iTerm is already running. If it is, it just creates a new window, otherwise creates the first window, then connects to the relevant services using my "go" script (which is synchronized across all my Macs by MobileMe).

Then, with it saved, I wrap it in a shell script:

/usr/bin/osascript /Users/peckrob/Scripts/launch-iterm.scpt

And launch it with MarcoPolo using my "Work" rule that is executed when my computer arrives at Work. Works great!

DD-WRT Hacks, Part 2 - Setting up an OpenVPN Server

By · Published · dd-wrt, networking

In my previous entry, I wrote about how awesome DD-WRT is, and how it had replaced a number of network devices allowing me to reduce the number of machines at home I had to administer. I finished the article by talking about how I'd set up a VPN tunnel to the office so multiple machines - namely, my Macbook Pro and my iMac - could access company resources at the same time.

But at the end, I mentioned that PPTP was _not _what I was using to connect myself back to my home network when I'm on the road. But why?

Two words: broadcast packets.

PPTP, by default, does not support the relaying of broadcast packets across the VPN link.* For Mac users, this means Bonjour/Rendezvous based services such as easily shared computers on a network are not accessible as they rely on network broadcasts to advertise their services.

PPTP can support broadcast packets with the help of a program called bcrelay. This program is actually installed on DD-WRT routers even, but does not work even though the DD-WRT web GUI claims that they can support relaying broadcast packets. To verify, you can drop to shell and try yourself:

[email protected]:~# bcrelay
bcrelay: pptpd was compiled without support for bcrelay, exiting.
         run configure --with-bcrelay, make, and install.

The version of pptpd that ships with v24sp2 of DD-WRT lacks bcrelay support. It's important to note that this doesn't mean the services are completely inaccessible. You can still reach them if you know IP addresses. Good for people with and understanding of networking, but not good for people like my wife and definitely not the "Mac way."

So, what options are left, if no PPTP?

Enter OpenVPN

OpenVPN is a massively flexible (and therefore massively difficult to configure) open source VPN solution. DD-WRT ships with OpenVPN server available with support for broadcast packets, so that is what I decided to use.

A couple of notes before you begin. There are some tradeoffs to using OpenVPN. Perhaps the biggest is that it's not natively supported on any operating system (unlike PPTP). That means on Windows or Mac, you'll need a third-party client. And it's not compatible at all with iPhones, iPods or iPads (unless they're jailbroken). It is also much more difficult to configure that the relatively easy and reasonably well documented PPTP server setup. It was a worthwile tradeoff for me, but it may not be for you.

So, before you begin, you'll need the following:

  • You have already configured your router using DD-WRT and have the most recent release (as of this writing, v24-sp2), VPN version installed.

    • The version number should be in the upper right corner of the web admin. If it says “std” or “vpn,” you’re in good shape. If it says “micro,” you probably don’t have the necessary tools.
  • You possess some basic understanding of networking, and have the necessary settings to complete a VPN connection. If you’ve gotten as far as flashing with third-party firmware, you probably do.

  • You understand that there is the possibility, albeit remote, that you could brick your router. I am not responsible for that, which is why I suggest you purchase an additional router to get all this set up on first before sacrificing your primary router.

  • You're not scared of the shell.

  • You must sacrifice a goat to the networking Gods.

For reference, my network uses 192.168.1.x for addresses. This can cause problems as it's incredibly common for LANs. You may want to change your addresses to something less common. Not that big a deal for me, though. I also have mine set up in bridged, as opposed to routed, mode. I thing this is smarter (and easier), but if you're curious, the difference is explained here.

The first thing you need to do is install OpenVPN on your client machine. Even if you intend to use something different, you still need to install it so that you can generate all the certificates you'll need. On a Mac, I find the best way to do this is with MacPorts.

toruk:~ peckrob$ sudo port install openvpn2

It'll crank for awhile compiling and installing what it needs, so go get a snack. Then, once you have it installed, head over to /opt/local/share/doc/openvpn2/easy-rsa/2.0/ and run the following commands:

source ./vars
./build-key-server server
./build-key client1

At each stage, it will ask you questions. It is important to provide consistent answers or you will get errors. Importantly, don't add passwords to your certificates. Once you are finished, you will find all your keys in the keys/ directory.

Now, the fun part.

Head over to the keys directory (/opt/local/share/doc/openvpn2/easy-rsa/2.0/keys). There should be a bunch of files in there. In a browser, open up your router's web admin, and go to Services -> VPN.

  1. Under OpenVPN Daemon, next to "Start OpenVPN Daemon," select "Enable"

  2. "Start Type," set to "WAN Up"

  3. CA Cert. Go back to your shell and "cat ca.crt". Past everything between the "-----BEGIN CERTIFICATE-----" and "-----END CERTIFICATE-----" including those two lines. You must include the BEGIN and END for this to work on each one! (This was a major trip-up for me).

  4. "Public Client Cert," go back to shell and "cat server.crt". Past everything between the "-----BEGIN CERTIFICATE-----" and "-----END CERTIFICATE-----" as above.

  5. "Private Client Key," go back to shell and "cat server.key." You need everything between "-----BEGIN RSA PRIVATE KEY-----" and "-----END RSA PRIVATE KEY-----" as above.

  6. "DH PEM," go back to shell and "cat dh1024.pem". You need everything between "-----BEGIN DH PARAMETERS-----" and "-----END DH PARAMETERS-----" as above.

The important not above is to include the lines containing "----whatever----". Not doing this cost me about 3 hours of messing around until I figured this out.

With that all complete, it's now time for your server config. Here is my server config:

mode server
proto tcp
port 1194
dev tap0
 # Gateway (VPN Server)   Subnetmask   Start-IP   End-IP
keepalive 10 120
verb 6
dh /tmp/openvpn/dh.pem
ca /tmp/openvpn/ca.crt
cert /tmp/openvpn/cert.pem
key /tmp/openvpn/key.pem 

The important things here are "dev tap0", which creates an ethernet bridge and not a tunnel (as "dev tun0" would do), and the "server-bridge" line. The documentation for that line is below it. The start IP and end IP specifies an IP range that VPN clients will receive addresses from.

With all this complete, press "Save" and "Apply Settings" at the bottom of the screen. Wait patiently. Then, in the web admin, go to Administration -> Commands. If you already have a Startup script, edit it, otherwise, add this to the commands window:

openvpn --mktun --dev tap0
brctl addif br0 tap0
ifconfig tap0 promisc up

Press "Save Startup." Then, if you already have rules in "Firewall," edit those, otherwise add:

iptables -I INPUT 2 -p tcp --dport 1194 -j ACCEPT

Press "Save Firewall." Now, reboot your router. When it comes back up, you should have a running OpenVPN server. To check, go to Administration -> Commands, and type this into the command window:

ps | grep openvpn

If you see something that looks like:

11456 root      2720 S    openvpn --config /tmp/openvpn/openvpn.conf --route-up
17606 root       932 S    grep openvpn

Then it worked. Congratulations, you have a working OpenVPN instance. But how to connect to it?

If you use Mac, you really have two choices: Tunnelblick or Viscosity. Tunnelblick is a little on the ugly side and difficult to configure, but is free and open source. Viscosity is reasonably pretty to look at and easier to configure, but is a commercial product. I chose Viscosity, so that's what I'm demonstrating here.

Once you have Viscosity downloaded and installed, go to Preferences and Connections, and add a connection. Enter a name and server address. Set the protocol to TCP and the device to tap.

Now, before you continue, go back to your shell. Go back to the /opt/local/share/doc/openvpn2/easy-rsa/2.0/keys directory, and copy those keys someplace in your home (~) folder that you'll be able to access.

Back in Viscosity, go to the "Certificates" tab. You should see three lines labeled "CA," "Cert," and "Key." For "CA," select the "ca.crt" file you just moved. For "Cert," select "client1.crt". And, for "Key," select "client1.key".

Under the "Options" tab, disabled LZO compression. For some reason this was causing a problem for me, so I just disabled it.

Click "Save." If all is right in the Universe and the goat you sacrificed to the Gods (you did do the goat sacrifice step, right?) was pleasing, you should now be able to connect back to your home network. Broadcast packets will work, and everything will be wonderful.

DD-WRT Hacks, Part 1 - Setting up a PPTP VPN Endpoint

By · Published · dd-wrt, networking

To celebrate the re-launch of my "blog," I'm going to do a multi-part entry about DD-WRT. But, first, a little history.

For the first time in 10 years, I have no servers running in my house. At one point, I had three servers running in here doing various things. Then, I moved my public server offsite (it's in the rack at the office now).

That left two more Gentoo boxes running here in the house. Late last year I picked up a 1TB external hard drive, which I attached to my iMac and deactivated the file server. I will probably eventually replace this with a Drobo FS, but for now it's fine.

That just left a single Gentoo box that was running Asterisk and various network services. But I finally convinced my wife to let me drop the goofy VoIP line that I was paying $30 for and just add more minutes to her cellphone. With Asterisk out of the picture, the only thing left running on that box was network services.

Well, a few weeks ago I ordered a TP-Link TL-WR1043ND router, intending to use it as a testbed for DD-WRT. Well, my experiments worked so well that I pulled my old router out and replaced it with the DD-WRT one. The faster processor also afforded a nice speed bump of about 7 Mb/s. With it handling all the services, I pulled out the final server and deactivated it. And my office is blissfully quiet now.

DD-WRT is now handling all the minor network services (DHCP, NTP, etc).

But what is it about DD-WRT that makes it so awesome - awesome enough to rip out some of my network infrastructure to make way for it? A few things that I will cover in this post.

1. DHCP static address assignments

Believe it or not, the built-in firmware of the WRT-54G did not give you the ability to define a static address to be assigned by DHCP based on MAC address. This seems like a glaring oversight to me, but it was the reason I ran my own DHCP server rather than use the built-in ones.

In DD-WRT (v24-sp2) you can go to the Services tab and set as many as you'd like. In my case, these are a couple of devices (like printers) that are addressed via IP address by the various machines, as well as my laptop and iMac.

So that's one nice thing, but it's not nearly as cool as ...

2. VPN Support

The standard and VPN versions of DD-WRT support both PPTP and OpenVPN varieties of VPN ... and I'm actually using both at the same time. My router is both a VPN server and VPN client as well. How? Why?

Well, as to why, at dealnews, we run a PPTP-based VPN to allow us to work at home as needed. Once connected, we have access to our testing servers and all our development services. It's like being directly connected to the work network, but I'm sitting at my iMac at home in my pajamas.

I had been connecting directly from my Macs to the VPN for some time but, sitting at home the other day, I reflected on how silly it was that I was connecting two machines to the VPN and only when I needed them, rather than using DD-WRT to have a single tunnel up all the time that any computer on the home network could use if needed.

Setting up a PPTP VPN Endpoint using DD-WRT

So how did I set it up? Trial and error, as, frankly, the DD-WRT documentation is a bit lacking. So if you find yourself in my position of wanting to have a tunnel to your workplace VPN, hopefully this documentation will help you.

I'm making a few assumptions before we begin:

  • You have already configured your router using DD-WRT and have the most recent release (as of this writing, v24-sp2), VPN version installed.

    • The version number should be in the upper right corner of the web admin. If it says "std" or "vpn," you're in good shape. If it says "micro," you probably don't have the necessary tools.
  • You possess some basic understanding of networking, and have the necessary settings to complete a VPN connection. If you've gotten as far as flashing with third-party firmware, you probably do.

  • You understand that there is the possibility, albeit remote, that you could brick your router. I am not responsible for that, which is why I suggest you purchase an additional router to get all this set up on first before sacrificing your primary router.

With that out of the way, let's begin!

  1. Log into your router's DD-WRT web admin, and go to the Services -> VPN tab.

  2. Under PPTPD Client, click the radio button next to Enable.

  3. In the "Server IP or DNS Name" box, enter your VPN server.

  4. In the "Remote Subnet" box, enter the network address of the remote network. In my case, this was

  5. In the " Remote Subnet Mask" box, enter the remote subnet mask. In my case, this was

  6. In the "MPPE Encryption" box, I have "mppe required,no40,no56,stateless". This was required to get mine to work, but may not be necessary for you. Try first without it, then try with it if it won't work.

  7. Leave the MTU and MRU values alone unless you know what you're doing.

  8. Enable NAT.

  9. Username and password are self explanatory.

WIth that done, press "Save" and "Apply Settings" at the bottom the page. With any luck, you should now have a VPN tunnel up to your remote host.

To test it, go to Administration -> Commands, and in the command box, enter the following:

ping -c 1 <some remote address on VPN>

If you get a response back that looks like:

PING <remote service IP> (<remote service IP>): 56 data bytes
64 bytes from <remote service IP>: seq=0 ttl=64 time=281.288 ms
--- <remote service IP> ping statistics ---
1 packets transmitted, 1 packets received, 0% packet loss
round-trip min/avg/max = 281.288/281.288/281.288 ms

Then it's up and working. Now, try from your computer...

Probably didn't work, did it? This is because your router's firewall doesn't yet know about the remote network or to route packets to it appropriately. For some reason, the current version of DD-WRT does not add the appropriate configuration to the firewall automatically when the PPTP tunnel is established. So, we have to do it manually.

Go to Administration -> Commands, and enter the following:

iptables -I OUTPUT 1 --source --destination <remote network address>/16 --jump ACCEPT --out-interface ppp0
iptables -I INPUT 1 --source <remote network address>/16 --destination --jump ACCEPT --in-interface ppp0
iptables -I FORWARD 1 --source --destination <remote network address>/16 --jump ACCEPT --out-interface ppp0
iptables -I FORWARD 1 --source <remote network address>/16 --destination --jump ACCEPT
iptables --table nat --append POSTROUTING --out-interface ppp0 --jump MASQUERADE
iptables --append FORWARD --protocol tcp --tcp-flags SYN,RST SYN --jump TCPMSS --clamp-mss-to-pmtu

At the bottom, press "Run Commands" and wait. It shouldn't take long, and should produce no output. Then, enter that command again, and press "Save Firewall" at the bottom. Give your router a few seconds to restart the appropriate services, then try again from your computer.

Your machine, and all machines on your network, should now be able to access the VPN. In this configuration, only traffic matching the remote network will pass over the VPN - the rest of your traffic will be routed to the Internet in normal fashion.

Now, in my next entry, I'll tell you why I'm not using PPTP to connect myself back to my home network when I'm on the road.


By · Published · news

Welcome to the new home for the Code Lemur blog ...! I've sat on this domain for six years - I don't know why it took me so long to port my blog from over to here.

Nonetheless, it is done now. And hopefully I'll find time to update it more with musings about my life and adventures writing code in dot-com.

MySQL-based Apache HTTP Authentication for Trac and Subversion

By · Published · apache, mysql, php

In working on a side project with a few friendly developers, we decided to set up a Subversion repository and a Trac bug and issue tracker. Both of these, in normal setups, rely on HTTP authentication. So, being that we already had an authentication database as part of the project, my natural first thought was to find a way to authenticate Trac and Subversion of these against our existing MySQL authentication database rather than to rely on Apache passwd files that would have to be updated separately.

Surprisingly, this was more difficult than it sounded.

My first thought was to try modauthmysql. However, from the front page, it looks as if this project has not been updated since 2005 and is likely not being actively maintained. Nonetheless, I gave it a shot and, surprisingly, got it mostly working against Apache 2.2.14.

Notice I said "mostly." It would authenticate about 50% of the time, while filling the Apache error logs with fun things like:

[Sat Feb 13 11:11:27 2010] [error] [client -.-.-.-] MySQL ERROR: Lost connection to MySQL server at 'reading initial communication packet', system error: 0
[Sat Feb 13 11:11:28 2010] [notice] child pid 19074 exit signal Segmentation fault (11)
[Sat Feb 13 11:34:14 2010] [error] [client -.-.-.-] MySQL ERROR: Lost connection to MySQL server during query:
[Sat Feb 13 11:34:15 2010] [error] [client -.-.-.-] MySQL ERROR: MySQL server has gone away:`

Rather than tear into this and try to figure out why a 5-year-old auth module isn't working against far newer code, and with very little to actually go on, I just concluded that it wasn't compatible and looked for a different solution.

That's when I came across modauthnzexternal. If your'e not familiar with this module, what it allows you to do is auth against a program or script running on your system, therefore allowing you to auth against anything you want - a script talking to a database, PAM system logins, LDAP, pretty much anything you have access to. All you have to do is write the glue code.

In pipe mode, modauthnzexternal uses pwauth format, where it passes the username and password to stdin, each separated with a newline. It uses exit codes to return back to Apache whether or not the login was valid. Knowing that, it's pretty easy to write a little script to intercept the username/password, run a query, and return the login.


include "secure_prepend.php";
include "database.php";

$username = stream_get_line($fp,1024,"\n");
$password = stream_get_line($fp,1024,"\n");
$sql = "select user_id from users where username='%s' and password='%s' and disabled=0"; $sql = sprintf($sql, $db->escape_string($username), $db->escape_string($password));

$user = $db->get_row($sql); if(!empty($user)) { exit(0); } exit(1);


Then, you just hook this into your Apache config for Trac or Subversion:

AddExternalAuth auth /path/to/authenticator/script
SetExternalAuthMethod auth pipe

<Location />
    DAV svn
    SVNPath /path/to/svn
    AuthName "SVN"
    AuthType Basic
    AuthBasicProvider external
    AuthExternal auth
    require valid-user

Restart, and it should be all working.

Some may argue that the true "right" way to do this is LDAP. But with just three of us, LDAP is overkill, especially when we already have the rest of the database stuf in place. The big advantage to this, even over modauthmysql, is the amount of processing you can do on login. You basically can run any number of queries in your authenticator script - rather than just one. You can update with last login or last commit date, for instance. Or you can join tables for group checking; say you want someone to have access to Trac, but not Subversion. You can do that with this.

OSCON 2009 Summary

By · Published · conferences

Have to say that, everything that didn't involved air travel (I'll go ALL into that later) was awesome on this trip. Had a good time and learned some useful things at OSCON, enjoyed good company and had a good time exploring San Jose and the Bay Area in general.

OSCON was good this year but not as good as in years' past. This may be due to the new location, which doesn't seem as conducive as the Oregon Convention Center did to a conference like this. The OCC was round, and all the meeting rooms were clustered in a central area - there was never more than a short walk between panels. But the San Jose Convention Center is more of a traditional box design, with a single LONG hallway. This means that if you're in J3 and have to go to B2, good luck, because it's a 15 minute walk. For a conference like OSCON, this kind of sucks and absolutely kills the "community" feel of it.

Also, like many things, it suffers from diminishing returns. Because a lot of this is stuff I've seen before, every year that I come, I have to work harder and harder to find something new. Three years ago, I was doing well to decide what not to learn about. So this may be my last OSCON for a few years, though I'm thinking of attending Velocity (held down the road at the Fairmont) next year.

I did attend some interesting side panels, including one on home automation. I have some ideas that I'm sure will drive Sarah crazy.