The Official Web Extravaganza of Joshua John Sherman

Server Migration

Josh Sherman — Sat, 22 Aug 2009 01:02:15 -0400

After nearly a decade of hosting on a dedicated server, I've finally made the jump to a virtual private host. Partially to save money, partially for some personal reasons, but mostly because I wanted more ease in scalability. The host I ultimately settled with was Linode, mainly because of their operating system selection, specifically, I wanted to run Ubuntu and MediaTemple lost out because they aren't doing Ubuntu servers yet. Over the last week I went ahead and migrated all of my sites over to the new server as well as all of my subversion repositories to git. Thus far, I'm very impressed with both Linode's and git's speed. To finish this hectic week, I also did some redesign on this site to help bring it out of the 90s and into the now. Oh, and if you're interested in hosting with Linode, please use me as a referral, my code is 5f682793582e82ce686747c851b998dc1f86a55b.

Redesign to the Max

Josh Sherman — Thu, 13 Aug 2009 21:12:25 -0400

If you haven't noticed already, this site got a drastic face lift recently. For the first time in a while, I made sure to mock up the entire site before starting to mess with the mark up. Weapon of choice was the GIMP (as per usual) and the process was fairly painless. Just to restate this for all the hates, you can actually make graphics in the GIMP. Just because you don't know how to use the application, doesn't mean other people don't. The same holds true for any piece of software, but for whatever reason it comes up very often in the Photoshop versus the GIMP debate. Sadly, I've let my Photoshop skills fade over the years, but have also been able to acquire a decent skill set in the GIMP. To get back on topic, yeah, the site's been revamped and no longer looks like a cheap rip off of Jon Gruber's Daring Fireball website. There will be more updates in the coming weeks / months, and perhaps a decent amount of tweaking to the actual design just because I'm not 100% satisfied with the new header. I've been partially swayed because it was described as "gay" (twice) by Daniel Jabbour (not that there's anything wrong with that.. but I can see his point) as well as "very 90s" by Dean Jones (I can see that as well, scan lines in any capacity can do that). Oh, I didn't mentioned, the site looks like total ass in Internet Explorer 6... that was on purpose. Eventually I'll get around to doing something special for the IE6 users out there (like a redirect to the goatse man, or simply stripping all of the CSS from the page). Special thanks to my friend Geoff from Plan8 Studios for helping out with some CSS questions I had.

So I'm addicted to checking my email

Josh Sherman — Fri, 24 Jul 2009 09:41:42 -0400

It's true, and not only do I check it, but I like to address the messages as soon as they come in. Great customer service ends up meaning huge time sink and often times a loss of momentum for me. Interestingly enough, I spoke to my boss on the matter earlier in the year concerning only checking my email a few times a day (first thing, before/after lunch and late afternoon before calling it quits for the day). I explained to him that this isn't feasible at all because if you don't reply to someone's message in what the sender considers a timely fashion, they will end up calling or in your cube to ask if you got their email. His response was that I need to check emails as they come in and prioritize when I respond to them (i.e. if you get an email from your boss it's your #1 priority, no questions asked). He obviously didn't get my point on the matter, and I didn't pursue it further. The point is, checking the mail is part of the problem. To stop what I'm doing go up and click Outlook then process the message (who's it from, what's the subject) and then rationalize whether or not to read the message and further to respond. To add to the problem, our corporate policy is that everyone's Outlook silently auto-responds to read receipts. So let's say that I check the email and say "this isn't pressing, I can respond after lunch" and the sender added the return receipt option. Well now they know I read the message, and the cycle starts anew. Fast forward to last week. I decided I have had my fill. At home, I removed `checkgmail` from my notification area and started detoxing, checking my personal email in the morning, after work and sometimes before bed. This week I started treating my work email a bit differently as well. If I'm in the middle of a task and a new message comes in, I do not check it until I'm at a logical stopping point. Thus far it's worked out well and no one's called me or shown up in my cube just to further disturb me (not yet). At home without gmail taunting me constantly my productivity has gone up a bit, and it's afforded me the opportunity to get into using Twitter (@joshtronic) to communicate with people and not simply keep a log of witty statements and links to trite nonsense. It was yesterday that not constantly checking my email became more habit than a chore, which was nice. To add to it all, I've recently stopped checking my Google Reader constantly as well and the last 3 mornings I've started my day before 8am and with a morning jog. Definitely a reboot from the norm.

Websites are more than just buying a domain

Josh Sherman — Sun, 12 Jul 2009 13:14:05 -0400

Most people I come in contact with (still) don't have a grasp of how much goes into having a website. This isn't their fault, and I definitely don't blame them for it. Hell there's a lot of details about things that I don't know nor do I plan to find out. To give you a back story, I had been planning on writing up "the n00b's guide to a website" for a while now, and even wrote up a rough draft to a friend that was in need of a website for her business. Now that I'm finally putting the time into it, it should be a bit more comprehensive. To start, having a website consists of 3 major parts: domain name, web hosting and the website itself. The domain name is most likely to be the first thing you acquire in the process and prices vary greatly based on who you are registering (DirectNic - $15 [and has been for years], GoDaddy - $1.99 - ??? [based on which up sells you are scammed into pick] and NetworkSolutions - $9 [but used to be significantly higher years ago]) and what type of domain you want to purchase (.com, .org, .me, et cetera). This cost is a recurring cost and many companies allow you to purchase a domain for multiple years, sometimes at a discount. I would recommend purchasing a .com and avoid the others as people still have a tendency to type .com even though you tell them .net or more obscure .mobi. Granted, none of this really means much in a world driven by Google searches. After you have your domain name, you probably will either attempt to build a site yourself or pursue a developer (or younger relative) to help out. This is an important step but shouldn't be done until you decide on where you want to host. Hosting comes in many shapes and sizes, you can lease space on a server with other people (very cheap, sometimes even

Going full circle on Allman Style Indenting

Josh Sherman — Thu, 09 Jul 2009 20:57:23 -0400

I've been programming for quite a while now, and I'm very much the type of person that will change coding styles if another style or technique will benefit me more than my current way of doing things. That being said, I've web full circle on using Allman style indenting. The obvious question would be "why?" or potentially "ewww, why? that's so ugly", both of which are acceptable. To answer the former, because it satisfies an issue I've been having lately with vertical whitespace after declarations and conditionals. For instance, here is a non-formatted example:

class Module { function DoSomething() { if (true) { echo 'Done!'; } else { echo 'Not done'; } } }

Nothing really wrong with that, but as of late I would format it like this:

class Module { function DoSomething() { if (true) { echo 'Done!'; } else { echo 'Not done'; } } }

And sometimes even

class Module { function DoSomething() { if (true) { echo 'Done!'; } else { echo 'Not done'; } } }

So yeah, I have vertical whitespace issues when it comes to code, and incidentally, I was half-assing the Allman style on my else statements (you can thank my current employer for that). After reading up on indenting styles I came full circle on the Allman style in it's pure form. I say this because anyone from the way back machine would know that I used to code like this:

class Module { function DoSomething() { if (true) { echo 'Done!'; } else { echo 'Not done'; } } }

Yeah past Josh really should be kicked for that one. Just for reference, and perhaps I'm late to find this out, but this is valid is valid code in PHP:

class Module { function DoSomething() { // if (true) //

RMS makes me want to be a better person

Josh Sherman — Mon, 06 Jul 2009 20:12:21 -0400

So I've been following the saga of Tomboy Notes (and subsequently Mono) being added to the Debian install. On the surface, it seems fairly harmless, and part of that is because I not only use Tomboy, but I enjoy it. As part of the ongoing saga, Richard M. Stallman (RMS) has spoken out against the inclusion because .NET is not free software.Regardless of where this all falls, I have to come out and say that RMS makes me want to be a better person. So much so, that I even gave gNewSense a once over to perhaps alleviate my shackles towards non-free software. Well, as much as I want to be a better person, and I may even give up Tomboy / any other Mono apps for that matter, I don't see me using an operating system that doesn't seem to handle my wireless card when I boot up the Live CD. I'm hoping to give Gobuntu a try soon, I'm assuming it will work better than gNewSense out of the box. So yeah, RMS is definitely a better person than I... for now. Oh, and for the record, #fuckpatents.

Oops I tried KDE again... and back to GNOME

Josh Sherman — Wed, 24 Jun 2009 14:05:37 -0400

It had been a good 6+ months or so since my last attempt at using KDE and I had that bug up my ass yet again. I had originally installed KDE on my wife's laptop on whim and was very impressed at first. So much so that I decided to install 4.2 on my desktop computer. I was starry eyed for a day or so but then kept coming back to the fundamental issue I always have, KDE just isn't productive for me. I did successfully get my GTK apps (specifically Firefox and Pidgin) to look as they should in KDE. I even gave Kopete another chance, but did end up back using Pidgin. I made sure to give it the old college try at least for a little while, but then it happened. I came decided I wasn't going to make Kwin perform the way I wanted it to, so I switched back to Compiz Fusion. Compiz ran like a champ, but I was missing hotkeys. I went into the config and messed around with resetting them and all that and they just wouldn't work. I can do without the Desktop Cube, but I needed some sort of hotkey to flip between virtual desktops, clicking the panel widget or mouse wheeling on the wallpaper wasn't going to cut it. At this point I was a bit defeated in the lack of hotkey for desktop switching, but I was going to give it some more time. That's when I tried adding some more plasmoid desktop widgets. Some installed fine, but most threw errors. Error I wasn't planning on researching. Long story short, I'm back in GNOME. I really like the desktop widgets in KDE but that wasn't enough to keep me there. KDE, I'll see you again in 6 months.

/home is where the heart is... on the NAS

Josh Sherman — Tue, 16 Jun 2009 14:04:47 -0400

In an attempt to unify my desktop data on my newly acquired laptop I decided to try out sharing my /home directory from my desktop as a NFS and then mounting it as /home on my laptop. Originally this idea was flawed because the laptop was not connecting to the wireless router on boot (remedied by my previous post). Now that this wasn't an issue, the idea didn't seem half, sans the fact that I was bound to my home network. To get started, I added the appropriate packages to my client:

sudo apt-get install nfs-common

and on my server:

sudo apt-get install nfs-kernel-server

Next I went ahead and modified /etc/exports on the server to share out /home:

/home *(rw,sync,no_root_squash)

and started nfs-kernel-server:

sudo /etc/init.d/nfs-kernel-server start

Once the server was ready to go, I added an entry in /etc/fstab on my laptop (client):

192.168.1.4:/home /home nfs rw 0 0

I rebooted my laptop, logged in and was welcomed by a crippled version of my desktop machine in the other room. Crippled by the fact that I didn't have the same set of packages install on the laptop, which was remedied by running the package installation portion of my SASSY (see previous post) script. All in all the system seemed to have a smidge of lag when opening files across the network and such, but overall it was very usable.

Wireless Connection at Boot (pre-GUI)

Josh Sherman — Mon, 15 Jun 2009 10:18:33 -0400

A while back I got my wife a new laptop from Dell, got it set up with Ubuntu 8.10 and then proceeded to fight with YoVille! for a week or so prior. Since then, I still had a few action items on my list, specifically getting wireless working on boot / the network manager not prompting for a password on boot. See, my wife's laptop is set up to auto login (yeah I know, gaping security hole). Said auto login would be great, but then she has to enter a password in to unlock the keyring so that the network manager can connect to our wireless network. Kind of self defeating IMHO. So I finally set out to fix the situation, and started with my laptop which is currently running Alpha 2 of Ubuntu 9.10. To get wireless working on boot, I added an entry to /etc/network/interfaces

iface wlan0 inet dhcpwireless-essid secret-router-namewireless-key s:even-more-secret-passphraseauto wlan0

Superb! Except that the network-manager showed that I wasn't connected. That's when I met the wonderful network-manager replacement known as Wicd. In Ubuntu 9.10 (and perhaps in 9.04) Wicd is included in the repository so installation was as easy as pie:

sudo apt-get remove network-manager network-manager-gnomesudp apt-get install wicd

After a reboot I had a wireless connection and a widget to show the status. Keep in mind this was on the beater laptop and not my wife's. When I went to repeat the process on my wife's laptop I ran into a mess of issues. Specifically, the network interface was called eth1 and not wlan0. Also, Wicd was not in the repository for 8.10 (going to upgrade her soon, but that's another post) so I had to download it from the Wicd SourceForge page. All in all, it was a fairly minimal hiccup to get her laptop running as well, and with a small bit of hacking I was able to get the wireless auto connect working flawlessly when coming out of hibernation and suspends.

New Project: Function of the Day

Josh Sherman — Sat, 13 Jun 2009 15:13:09 -0400

So I've bought yet another domain, but actually got a site up in a timely fashion this time. The site is called Function of the Day and it's just that, a site that presents a single programming function every day. Right now it only supports PHP, but the plan is to add more languages as time progresses. Assuming the site gains a bit of interest from the community, the sooner new languages will be added. Hell, if you really want to see the ball get rolling faster, get in touch, and I'll put you to work.

Super Awesome Setup Script, Yummy!

Josh Sherman — Sun, 07 Jun 2009 15:01:18 -0400

The script better known simply as `sassy` was a long time coming. Every time I upgrade Ubuntu and then have crap fall apart to the point that I end up reinstalling I end up saying "Josh, if you had a script that would easily get a fresh install of Ubuntu up to speed for you, you'd be set". And set I shall be. I started building a script that includes a listing of all of my installed packages and updates all of the /etc conf files to get me there. This did include some changes in how my system was set up, specifically where I kept files. I've since moved everything into my /home directory, including my development web sites which are now symlinked back to /var/www to match my production server. The plan is to eventually move /home on to it's own partition so that I can reinstall the system and leave it intact. The script thus far is under 100 lines and is going to be a lifesaver on my next install (9.10 I will be performing a clean install). I won't be posting the source code as I feel the script is just a bit too customized towards my needs. I do recommend setting something similar up yourself, it's a very fun project and takes a minimal amount of Bash knowledge to get going. It will also help get you in the mindset of keeping backups. The only gotcha I've come across is that when restoring MySQL on Debian, if you restore the "mysql" database, you will end up hosing your install because the debian-sys-maint account's password changes. It's easily remedied, but I'd avoid doing a full import on that database if at all possible.

And finally, I upgraded my hard drive (part 2)

Josh Sherman — Fri, 29 May 2009 14:40:37 -0400

And so the story continues. After much difficulty installing / setting up my new hard drive, I decided to make matters worse by cloning my old Windows XP partition to one of the partitions on the new drive. After a small amount of research, I came across `ntfsclone` part of the ntfsprogs package. As quick as you can say "apt-get install" I had it installed and was toying around with it. My biggest concern was that I was coping a 400GB drive to a 500GB drive. I was able to clone my drive and boot it up successfully, albeit it was only reporting 400GB. Wait a second, where did my Vista bootloader go? Shit, lost it again. Before I addressed that issue, I went ahead and resolved my 400GB drive issue. Luckily, there was another application named `ntfsresize` that fixed me right up. Okay, so now I'm at the point where I have successfully cloned my XP drive, but Vista was lost because of the missing bootloader. I went through the motions of running the restore bootloader from the installation disc... 5 times. Yes, 5 times, because after Googling around a bit, I found that the restore doesn't work all of the time, in fact it's very common for it to take multiple attempts before it works. "WOW", is all I can say about that. Now the repair did finally work, only now I could only boot into Vista, and no option for XP. Fed up, I went searching for an application to fix the glitch. Said application was Vista Boot Pro, which was well worth the $9.95 to fix my situation. The program worked first time and allowed me to add Vista and XP to the bootloader with the labels that I wanted. I've yet to find an open source alternative so that may end up being a project of mine in the near future.

And finally, I upgraded my hard drive

Josh Sherman — Mon, 25 May 2009 22:10:52 -0400

Was a long time coming, seeing as I purchased the 1.5TB drive back in February. Pre-upgrade I had a 500GB drive partitioned into two partitions. The 100GB partition housed Linux (Ubuntu 8.04 upgraded to 8.10 upgraded to 9.04) and the 400GB partition was Windows XP. Wait, what? To clarify, I still use Windows XP for audio production. The larger partition size was to accomodate massive amounts of audio in raw formats. In addition to my internal storage, I have a 120GB external Seagate which housed music, audio samples, VST Plugins, software installers, and a mess of other things. Before I proceed with discussing the upgrade and issues I ran into let me tell you a little story. It's only since my latest computer purchase that I have had a computer that didn't use IDE drives. That being said, I don't have SATA cables lying around or even know much about them. The same day I made mention to a coworker that this would be my first SATA drive upgrade and that I would need to purchase a cable from the store, he offered to bring one to me the next day as he a whole box of them. So began my issues. Initially I went ahead and removed the old drive and installed the new drive in it's place. The old drive was moved to the second drive slot and I hooked up the power to both drives, the original SATA cable to the new drive, and my coworker's cable on my old drive. Rebooted, checked the BIOS, only the new drive was showing up. Weird. After a few reboots and reseating of the cables I ended up moving my coworker's cable to my DVD burner, the cable from the burner to my old drive and rebooted. Everything showed up fine so I p roceeded to partition the new drive into 3 500GB partitions. The act of partitioning ended up being a bear as Windows wasn't too fond of the partitions I had made with `fdisk`. Sadly, Windows XP saved the day and I set up my 2 Windows (XP and Vista) partitions and left the rest for the Ubuntu install to manage. After the partitions were set up I had to make a decision about what order I should install the operating systems. First pass I went ahead and decided on Vista, then XP then Ubuntu. This didn't seem like much of an issue until I learned that the Vista Bootloader was new and improved, and that XP killed it when I installed it after Vista. At this point, I couldn't boot into Vista and figured, "might as well give the Ubuntu 9.04 Live CD a shot". A previous post of mine was a rantfest concerning the Ubuntu upgrade process and how Compiz stopped working after my upgrade to 9.04 due to a faulty driver. Well guess what, Live CD worked flawlessly. I suppose I could appologize for what I directed at Mark Shuttleworth but in all actuality the Ubuntu upgrade process is still pretty immature IMHO. It was around this time that the fun really started. I was going to start reinstalling my two Windows partitions (XP and then Vista) when I started having drastic issu es reading my installation media. After multiple attempts with different [Dell certified] copies of Windows I decided to start messing with the SATA cables again. At this point my DVD-ROM was using my coworker's SATA cable. I said to my wife "I bet if I swap the cord back to the original cord Vista will install properly." Sure as shit that was how it went. Windows Vista installed, Windows XP installed, and now on to Ubuntu. Surprisingly the Windows installed went relatively smooth after the cord swap compared to the Ubuntu installation. Ubuntu would crap out at seeminly random places during the install bitching about how it was unable to read the disc. A few retries and it finally installed without issue. Wonderful! So now I'm sitting on a wonderful triple boot system. Unfortunately the story doesn't stop here. At this point I decided to test the DVD-ROM under different circumstances. I cycled through all 3 cords, different connections on the motherboard, different power cables I even completely restored the configuration back to the original. Just for reference, I was all sorts of pissed off at this point, specifically because the coworker that OFFERED me the cable without any solicitation had something insightful to say to me when I told him the situation. He said, and I quote, "beggars can't be choosers". To say the least I gave him a huge piece of my mind, and subsequently decided that I will never take anything from said person in the future. Long story short, the drive was unable to burn DVD's, CD's were sporadic (1 out of 3 successful) and eventually, the drive just flat out stopped showing up if I was using the coworkers cable. So $6.99 (plus tax) at the recently re-opened CompUSA later I owned a new cable, and all 3 drives were showing up correctly. Unfortunately, the DVD-ROM still won't burn discs correctly and I eventually gave up on it. Stay turned for Episode 2: Migrating Data to the New Drive or: How I Learned to Stop Worrying and Love NTFSProgs.

Mass `svn add` script

Josh Sherman — Sun, 17 May 2009 23:37:26 -0400

So I have a tendency to add a bunch of new files to a project before actually doing a commit in Subversion. It's usually not an issue, I go through an add each new file and then finally do my commit. Welp, I'm sick of it, so I wrote a little Bash script to help automate the process slightly. The script takes the directory and loops through detecting any files marked as "?". From there, it will prompt if you would like to add it or not. Easy as that. Unfortunately, I wrote it after I did a large commit earlier this evening. Next time, I suppose.

Source code for svn-add

Media Temple Grid-Service... ehhh, not so much

Josh Sherman — Wed, 13 May 2009 20:08:50 -0400

I'm starting to feel like I'm bitching and complaining a lot on here instead of documenting steps towards resolving issues. Today's subject, Media Temple's Grid-Service hosting. I personally don't have a Media Temple account (for reference, I am currently hosting on a dedicated box with ServerBeach) but have a few friends, and a few clients that use them. Now I've noticed their slowness and have heard a few back up horror stories, but never before have I worked through a Grid-Server crash. Before I continue, let me quote their web site:

"(gs) is a cluster-based, modern hosting service powered by hundreds of servers working in tandem to power your websites, applications and email with unrivaled power, burstability and reliability."

Sounds great, right? I thought so, in theory this is perfect; multiple servers, virtual machines, redundancy, zero change of a server crash? So I thought, I thought the point of having cloud hosting like this is that is anything went wrong, they could just bring up a copy of the VM that took a shit and you'd be good to go. Not so much, to say the least. A week or so ago, one of my client's site's went down, hard. Upon further investigation, their (gs) took a shit and was in the process of being brought back up. WHen I say brought back up, I don't mean a new image was brought up, I mean they had to restore backups and it took literally days. Okay, not bad, server's back up and all is well in the neighborhood, right? Wrong. Now their shopping cart is throwing errors. "Did it work before the crash?" I asked. "Sho'nuff" was the reply (paraphrasing of course). So I was later asked to troubleshoot and resolve the issue, which I did. So what was the problem? The shopping cart code had hardcoded paths (that's an issue in itself, but I wasn't here to update the code to use config files) that were all pointing to non-existing paths. I logged into the (mt) admin and got the system paths from there, updated the code and "voilà" all is well again. That struck me as very strange, if we're running on a cluster and on virtual machines, why wouldn't the server come back up exactly the same as before the crash? My conspiracy theory is that they didn't bring back up a copy of that server and restore an incremental backup, they built a new VM differently than the original was built, and that led to the path's being different. Now I could be wrong with all this, but the whole thing does smell fishy. Either way, I think it's great that Media Temple was willing and able to provide a credit for the suffered downtime and issue that was encountered after the "restore".

Super massive content update

Josh Sherman — Sun, 10 May 2009 22:10:32 -0400

Gutted just about every page on the site. There's more of a focus on trim information without a lot of BS. No more links to friend's sites and software and shit that didn't really matter. Rewrote / updated the text on some pages to update their accuracy and revamp based on my own writing skills progressing. Oh, I also wrote this bitchin' age calculator plugin for Smarty:

Source code for function.age.php

Intel Graphics Regressions in Jaunty, say wha?

Josh Sherman — Fri, 08 May 2009 10:33:28 -0400

A piece of me died after I upgraded from Ubuntu 8.10 to 9.04. Apparently, and serves me right for not reading up on it before the upgrade, the Intel Graphics drivers are completely crippled by a bug in the code. Wow, and this made it to the final release? I'm actually taken back by it a bit since Canonical Ltd. has earned my respect (for the most part) when it comes to being able to release very stable versions of Ubuntu every 6 months. What's this really mean? It means I no longer have Compiz running because my screen goes black and glitches out if I do. Oh, and I know that Compiz is just "slick eye candy" to most of you, but there's more to it than that. Window previews, real transparent terminals (huge for me), task switcher with live updates, and yeah, the cube rocks and is more for show than anything ;) Now I bet you're saying to yourself, "Josh, there's fixes all over the web right now for the issues". And I say to you, yes there are, there's a good chance I've tried the ones you're thinking of, and no they didn't work. Kernel switches, driver updates, config hacking, the closest I've gotten is Compiz works for like 10 seconds, and once Gnome is 100% loaded, it takes a shit. So now I'm faced with a big decision, reinstall Ubuntu 8.10 (I'm leaning towards this, since it will give me a chance to finally install my new hard drive), upgrade my desktop to a beefy Nvidia video card (high on the list since I need a new video card to play Portal, but doesn't help the fact that I have two other computers in the house with Intel video) or abandon Ubuntu all together and find a new Linux du jour (highly unlikely as I've had my fair share of issues with Ubuntu in the past and am not willing to completely abandon what I generally consider to be a great product). To Mark Shuttleworth I say, I am disappointed, a bug this large shouldn't have made it to a "stable" release, ever. And if anyone is wondering, `lspci | grep VGA` yields: 00:02.0 VGA compatible controller: Intel Corporation 82G33/G31 Express Integrated Graphics Controller (rev 02). Other than the Intel bug, the upgrade to Jaunty was generally easy going. Only other gripe was that /etc/hosts seems to have gotten wiped out, so my local development URLs didn't work and had to be set up again.

I wish YoVille! would just die

Josh Sherman — Sun, 05 Apr 2009 14:11:38 -0400

So my wife loves this ridiculous online / Facebook game called YoVille!. It's basically a silly point system tacked onto a chat room. Even worse, it's formal equivalent would be Animal Crossing, which I shelled out like 70$ for when it first came out. Sadly, the new Animal Crossing on the Nintendo Wii was a port of Animal Crossing on the DS which was basically an update to the original Animal Crossing on the GameCube (which we already own). My wife grew bored with it pretty quick, and thusly attached herself to YoVille!. Let me get back onto the topic at hand, my wife's laptop was fairly dated at this point, 1.2GHz Intel Celeron with a laughable 512MB of RAM. I was noticing she was rebooting the computer an awful lot due to freeze ups (something she could handle w/o a reboot if she'd be willing to embrace Linux and let me teach her how to solve her problems). So I decided to buy her a new computer, especially after the recent "YoVille! stopped working" scenarios. Oh I didn't mention, not only is the game stupid (take that with a grain of salt, I love me some MobWars) but it's always having issues. Zynga, Inc. knows how to market the hell out of disposable games, but they really don't know how to build solid user experiences. That all being said, YoVille! stopped working, and my wife is now the proud owner of a shiny new Dell Vostro 1510, with a Intel Core 2 Duo T5670, 1.8GHz proc, with 3GB of RAM and I even sprung for the larger battery for extra wireless time on that damn game. I'm actually a bit jealous, it's not quite as beefy as my desktop rig, but it's pretty sweet. So at this point, I've wiped Windows Vista off and installed Ubuntu 8.10 (actually, I contemplated a XP / Ubuntu dual boot scenario, but my wife's not really the dual boot type). Firefox is installed by default, so I made sure that Flash was installed, loaded up YoVille! and tried to log in. Son of a bitch! All this machine, and it won't load. Let's fast forward a few hours. I've tried Googling around, installed different builds of the Flash player that are "proven" to work. Tried different versions of Firefox, tried this, tried that, tired of trying. Sadly, I learned something very important. Apparently there's a lot of other husbands out there with wives that insist on playing that infernal game, and they too have been trying to fix a myriad of issues, most of which seem to be issues with the Flash movie that Zynga, Inc. has released. Sadder still, I finally gave up after many many many package installs, reinstalls and removals and did something that I normally wouldn't do to solve a "Linux" problem. I installed the WINDOWS version of Firefox, installed the latest Flash player. Sure as shit, it works, and my wife now has an icon on her lovely Gnome desktop labeled "Windows Firefox". Sickens me, but at least this brand new laptop wasn't purchased in vain.

New Feature, an RSS Feed!

Josh Sherman — Tue, 03 Mar 2009 20:22:51 -0500

Sadly, I failed to add an RSS feed to the site when it initially launched. Even sadder, the "system" I have been developing for the last year (it's called PICKLES and is about to make it to revision 100 in Subversion!) has a nice display module for RSS feeds, so all of the leg work was already done. Well I took a small bit of time for myself this evening before jumping back on a freelance project to add the feed capabilities to the site. Enjoy!

Nintendo Wii Won't Power On

Josh Sherman — Tue, 24 Feb 2009 20:26:29 -0500

This is the second time this has happened to me in the last 6 weeks, my Nintendo Wii simply will not power on. No lights, no anything. The last time it happened, I unplugged everything, unplugged the power adapter from the wall and proceeded to pop the side doors off before taking it apart. Before I got too deep, I went ahead and tried another outlet and the damned thing powered up just fine. Well it happened again tonight, and I decided to troubleshoot a bit better. I unplugged the power from the back of the Wii and left it unplugged for 10 minutes or so. No lights. So I repeated 2 more times. Nothing x2. Before giving up all hope, I said to myself "you're only unplugging it from the back of the console, how about unplugging it from the wall?" I proceeded with unplugging the cord from the outlet and let it sit for another 10 minutes. Plugged it in and it worked just fine. The issue only seems to come up after the power is abruptly cut from the unit (we have a not so secure 6 outlet adapter that occasionally gets knocked by the door to the entertainment center). If your Wii won't boot up, definitely give this a try.

Pidgin Issues Revisited

Josh Sherman — Sun, 22 Feb 2009 19:26:29 -0500

Welp, the mysterious Pidgin I was receiving a while back popped back up. For those that don't remember, it was "Oops, father has gone, wait for me, wait...!" and then a segfault. Well the issue magically corrected itself the last time, so I was at a loss for what to do this time around. Fortunately Firefox decided to give me a big error message when trying to watch a video. The gist of the message was "hey fool, you're out of disk space". I learned a valuable lesson that 100GB for my Linux partition is a gross underestimate of the space I needed. Fortunately, my new 1.5TB drive from Dell came in the mail, now if only I had time to install it!

File Usage Reporting Script

Josh Sherman — Wed, 18 Feb 2009 23:47:54 -0500

In retrospect on writing this Python script, I probably could have pulled it off in a short Bash script. Maybe another day. Anyway, this latest utility script of mine takes a variable number of arguments, the first being the "needle" directory. The needle directory are the files you want to check the usage on. Any additional directories listed will be considered "haystacks" and the files will be searched for any occurrence of the needle files. A report is generated so you can easily identify which files are no longer being used. I typically search my Smarty template files for occurrences of my images. Any images that aren't referenced are typically removed from the repository promptly. Helps a lot while developing / updating existing pages will new content.

Source code for file_usage.py

Smarty Function: State Drop Down

Josh Sherman — Thu, 05 Feb 2009 21:35:52 -0500

Today is the last day I will have to dig around for a drop down of states to use on a web site. I've finally stopped being lazy and added a Smarty function to PICKLES to allow for easy insertion of a state (50 + DC) select box. Later enhancements may come to support Puerto Rico and the Canadian provinces, but for right now, it's just the United States (as that's all I needed on there). PHP with PICKLES (soon to be released maybe?) automatically loads this function in for me, but if you're interested in using it and you're not using PICKLES, you may want to consult this page for how to register a custom Smarty function.

Source code for function.html_select_state.php

Using SQL's IN syntax to search multiple fields

Josh Sherman — Mon, 02 Feb 2009 18:12:03 -0500

This may not be something new for anyone, but figured I'd post it anyway. I needed to check for a certain date in multiple fields, instead of listing each one out, I used IN but used the value on the left side, and listed the fields in the parenthesis. Before:

SELECT id FROM table WHERE date1 = "1981-02-23", OR date2 = "1981-02-23", OR date3 = "1981-02-23", OR date4 = "1981-02-23", OR date5 = "1981-02-23";

After:

SELECT id FROM table WHERE "1981-02-23" IN (date1, date2, date3, date4, date5);

Nonogram Demo in PyGTK

Josh Sherman — Mon, 19 Jan 2009 22:26:53 -0500

So a while back I had gotten 3-in-1 Sodoku Garden on my cell phone because I like Sodoku even though I'm not that great at it. I figured, 3 games for the price of 1, wicked cool. Turns out one of the 3 games would later steal my heart. It went by the name "Tenpenki" and I couldn't find anything on the interweb about it that didn't include the game on my cell phone. I thought to myself, "self, this would be a great project to learn PyGTK and you'd be rewarded for your efforts with a verion of Tenpenki for the computer". And it was. Funny enough though, once I got into coding my version of the puzzle game, I did some further digging on the game and I came across the Wikipedia page that completely summed up the alleged "Tenpenki". It was called a Nonogram and it went by many names, none of which included "Tenpenki". At this point I figured, no reason to stop just my Nonogram project just because Simon Tatham's Portable Puzzle Collection has one. And so I present to you my Nonogram Demo. It's extremely hard due to the random puzzle generation with no regard for skill level. I also opted to use check boxes either checked or unchecked, this just adds to the confusion as you can't mark off any cells are "definitely blank". Expect a better version in the future.

Source code for nonogram.py

Oops, father has gone, wait for me, wait...!

Josh Sherman — Fri, 02 Jan 2009 11:14:34 -0500

So out of the blue last night, Pidgin (version 2.5.2) crashed on me and when I tried to relaunch it, it would only stay open for a brief moment. I hadn't done any updates to my system since my last reboot, so I couldn't imagine it was some new busted package. I tried launching it from the command line to find that there was a segmentation fault. Passing --debug ended up giving me a DNS error with a pretty nifty message, "Oops, father has gone, wait for me, wait...!" I went ahead and passed --nologin which got the program running, but as soon as I tried to connect with any of my services (AIM, Gtalk and IRC) it would bomb out, seemingly at random times. I systematically went through my accounts and found that a certain AIM account was the issue. It seemed whenever that account was connected, it would try to load the buddy information and then died on a specific person. I don't think that person was necessarily the issue, as this morning I tried to connect that account, and everything worked out fine. Might have to call unsolved mysteries on this one.

All Music Guide (AMG) Tagging Script

Josh Sherman — Wed, 24 Dec 2008 20:00:03 -0500

So recently I decided that I hated how my music collection was tagged. I try to break apart the genre's but ideally most of my music falls under multiple genres. Well All Music Guide (AMG, allmusic.com) does a pretty good job of classifying everything. They use broad genres, and then further define music based on the styles it falls into. Most of us consider those styles to be genres of their own, but ideally you can't really have more than one genre, and if you use iTunes (like I do) you know that comma separating the genres is a mess (each unique genre string is treated as such). My solution is to use the broad genre as the actual genre, and then plug the styles into the comment tag. This worked out great, except that it was a lot of leg work to look up those artists and update the tags manually. Unfortunately I did that with my entire library before writing this script. The script I wrote is in Python and looks up the artist / album on the AMG website, then loads in the genre and styles, and updates the files accordingly (currently only supports MP3 files). Just to warn you, this script is completely proof of concept and may be considered a malicious spider by AMG as it does mine their site for data. "Why not use one of those other services out there to update my library," you ask? Well it's because I don't really considered "this is shit" to be a valid tag for any of my music. AMG is for the most part the best repository of such data (albeit it's mostly mainstream music) and it's not biased the way other community driven databases are. Enjoy allmusictagger.py and if you end up expanding upon it, drop me a line and let me know.

Source code for allmusictagger.py

PHP Warning: Module 'mcrypt' already loaded

Josh Sherman — Sun, 14 Dec 2008 17:11:00 -0500

So the other day I realized that my concert spider stopped running. To my surprise I had a lot going wrong on my local machine (which runs the spider). First and foremost, MySQL wasn't running. Still not sure why, but I manually started it to get it going. It ran through a recovery routine to check for corrupt data, I suspect I had a bad shutdown at some point. At this point, I figured I was golden, so I tried to launch my spider. Blech... no dice. After a few attempts of starting it from the init script, I decided to go to the actual location of the script, and launch it manually. "PHP Warning: Module 'mcrypt' already loaded in Unknown on line 0" was presented to me. Yes, it's only a warning, but I work in E_STRICT mode in PHP so warnings are not an option. Well I dig around the internet and see that other folks have had the same issue, and it's apparently reported to Canonical as a bug in Ubuntu 8.10 Intrepid Ibex. From what I can tell, this has yet to be resolved. Keep in mind, this is on the CLI version of PHP and I'm having no issues with the Apache module. So at this point, I went ahead and did a grep for 'mcrypt' in /etc/php5. Surprisingly, I got more hits than I expected. The extension was being loaded by mcrypt.ini in conf.d, apache2 and cli directories as well as the php.ini under the cli directory. Now don't quote me on this, but it seems as if the package php5-mcrypt created mcrypt.ini files in all of those directories but somewhere along the line, the php.ini under the cli directory was also modified. This caused the double load of the extension but only for the command line version (I haven't seen an issue with the Apache module). Easy solution is to comment out of the extension from /etc/php5/cli/php.ini. This did in fact solve my issue.

Bash script to flatten a directory

Josh Sherman — Tue, 25 Nov 2008 12:08:48 -0500

I just came across an old CD of Fonts that I had purchased before I knew better. Well I didn't want to throw the disc out without copying all the fonts to my system. All the fonts were broken up into directories that corresponded to the first letter of the name. It seems like there is some Windows software interface to handle installing the fonts that would make those directories transparent. Since I didn't want to have those fonts split up by first letter, I wrote a little bash script to flatten the directories. All it does is loop through all the directories in the passed source directory, and copies every non-directory file to the passed target directory (or . if the argument is omitted).

Source code for flatten-dir

Note: It was pointed out to me that this script doesn't handle identically named files inside the directories you are flattening. This wasn't something I had to worry about with the fonts CD, but you may need to worry about if you intend to use this. A couple of solutions would be to prefix the files with the directory name they came from, or keep a list of the files you've already moved, and if you hit a duplicate append a number to the end, or even prompt for user input and let them rename it or something. Perhaps in the future I'll expand upon this script, but not today my friends.

Why I still use GNOME instead of KDE

Josh Sherman — Sun, 23 Nov 2008 16:12:48 -0500

When I first started using Linux a decade or so ago I had no real preference in desktop environment. I used KDE, I used GNOME, I even ran command line only depending on the distribution. Once settling on Slackware as my distro of choice, I found myself using GNOME on a full time basis. From time to time I would switch to KDE for a short stint and then switch back and acted like the whole thing never even happened. I even dabbled with XFCE, Enlightenment, and Blackbox and it's derivatives on a few occasions, still going back to GNOME. Just to get it out of the way, Enlightenment was too sluggish for me back in the day, and XFCE and Blackbox are fine, but I really haven't been able to get into the minimalistic menus and stuff. And then there was KDE. KDE is something I keep giving every major release and a lot of times on the minor releases a try, but none have ever swayed me. First and foremost, early on KDE was sluggish to me compared to GNOME and it reminded me too much of Microsoft Windows, which is what I was trying to get away from. In modern times, the sluggishness isn't too much of an issue for me. What does drag me down is the software and the GTK versus QT display issues. GTK apps look like shit in KDE without a lot of massaging and hackery. I don't have the time in my day to sit around trying to make those applications look pretty. Now that being said, I am aware there are alternatives to the GTK software I use in KDE. I don't want to use those applications, I want to use the applications that I'm used to and have fallen in love with. Specifically, Firefox is the web browser of Linux. I know WebKit is all the rage these days, but I've never liked Konqueror, and all of it's superior abilities when it comes to passing the Acid test. The other app that I can't stand in KDE is Kopete. I'm not sure what the hell that name is supposed to mean or what, but the application in general has given me more problems than it's worth. Pidgin (formerly GAIM) has been my choice for instant messaging for many years and will remain as such for many more years. Those two applications aside, I really dislike the Konsole and for whatever reason (perhaps I'm dumb) I have a hell of a time finding anything on the KDE menu, the new one (KDE 4.x) especially. So call me a creature of habit, but that's why I don't use KDE. I am infinitely more productive in GNOME, it still seems to be the de facto standard for a lot of distributions and quite frankly, I don't care that Linus Torvalds says I should use KDE over GNOME. The beauty of Linux is that we can pick and choose what we want to use, and I choose GNOME.

Upgrading from Ubuntu 8.04 LTS to Ubuntu 8.10

Josh Sherman — Sat, 01 Nov 2008 13:27:24 -0400

Well after a few days of not being able to run any upgrades on my installed copy of Ubuntu 8.04 LTS, last night things started moving. The upgrade process from Ubuntu 8.04 LTS to Ubuntu 8.10 took all of last night (I started it around 9PM) and on through this morning due to some interactive dialogs. The upgrade went surprisingly well compared to my previous Ubuntu upgrades (most of which ended up with me doing a clean install). This upgrade was not without it's complications though. First and foremost, the UbuntuStudio menu refused to upgrade at all. I know it's not a package maintained by Canonical, but it was a flat out failure when trying to upgrade it. I proceeded with the rest of the upgrade, and then went on to remove all of the UbuntuStudio packages I had installed previously, and then reinstall them. No issues with the UbuntuStudio menu on reinstall. Next came Compiz Fusion / Emerald Theme Manager. Compiz was running prior to the upgrade, and after the upgrade as well but Emerald Theme Manager was not running. After some Googling I found a nice forum post that outlined the uninstall and reinstall steps. Still no dice until after I started Emerald Theme Manage with the --replace flag. In hindsight, I should have just tried that first, I ran `emerald` from the command line, but didn't include the --replace flag, so that would have saved a bit of time. Long story short, Compiz Fusion with Emerald Theme Manager is back and I installed the Compiz Fusion Icon to help troubleshooting in the future. At this point, my desktop is looking good and my window decorations are solid, "I wonder if my local development environment is hosed," I thought. Sure as shit, it was, and kind of baffling to boot. Apache2 was up and running but all of my virtual hosts were pulling up the default document root. This was a surprisingly easy fix once I started digging into the Apache error log. Seems the way I had my virtual hosts set up, as <VirtualHost *> was not accurate compared to the NameVirtualHost directives in apache2.conf. I went ahead and added NameVirtualHost * to the end of apache2.conf and restarted. Now all my sites are up and running. On top of doing the upgrade, I made the decision to (yet again) remove KDE entirely from my system. I wasn't entirely sure of the entire list of packages to remove, so I used Synaptic's section list and removed everything from all three KDE Desktop Environment sections. There are a small list of packages still installed when I search for "KDE" but it's not enough to force me to put any more efforts into it. I'll elaborate more on why I continue to go back to GNOME after numerous attempts with KDE on another day. So that was my upgrade in a nutshell, most of it was just waiting, but the outcome was well worth it. I'm hoping that my next upgrade to Ubuntu 9.04 next year will be even easier as I really don't have the time in my life to spend 10 hours trying to troubleshoot something that should just work (my Gentoo and Slackware days are long over).

"My Computer" hangs when accessing it

Josh Sherman — Mon, 13 Oct 2008 10:56:53 -0400

This just started happening to my work PC (Windows XP SP3) about a week or so ago. It was very annoying, mainly because I rely heavily on the CTRL+E hotkey to access "My Computer" on the fly. Well today I had enough. First stop was Google. My searching resulted in many sites recommending to launch Explorer, go to Tools then Folder Options and click on the View tab. From there uncheck "Automatically search for network folders and printers" (the first option) and then click OK (or Apply then OK if you're that guy). Yeah, no dice, still slow. Next step was to start disabling software and running services, I was sure Trend Micro OfficeScan or Apple were behind the hang ups. Still nothing. Well after another 20 seconds of waiting for "My Computer" to load, I decided to start going through my mapped drives to see if perhaps one of those was bad. Four mapped drives, the first three were fine. Then came number four. I double-clicked it and sure as shit, it hung for a moment, and then displayed an Explorer window showing my C:\ drive. Not a good sign. Well I disconnected the drive and tried CTRL+E. Surprise, everything is loading up as it should. Funny enough, I would have expected the option I unchecked to have possibly resolved the issue. Guess not, but now I'm back in business.

Recursive grep (grepr) for Solaris

Josh Sherman — Thu, 09 Oct 2008 10:33:04 -0400

So if you're a Linux user that's ever used Solaris, you know that a lot of the commands don't translate exactly the same. One such issue is the lack of a recursive flag (-R) on the grep command. Not a problem though, bash is a powerful thing. Here's my solution to the problem:

Source code for grepr

Firefox 3.0.1 on Mac OS X weirdness

Josh Sherman — Fri, 03 Oct 2008 20:22:00 -0400

So I ran into an interesting problem today. A site I built (validates as XHTML 1.0 Strict and CSS level 2.1 and runs in Standards-mode instead of Quirks-mode in Internet Explorer 6.0) renders great in Firefox 3.0.3 (tested on both on Linux and Windows), Internet Explorer 6.x (tested in Wine and Windows) and Safari (tested in Mac OS X and Windows). Unfortunately, it wasn't rendering properly (a side bar was displaying way too wide and was breaking the layout) on the client's installation of Firefox (version 3.0.1). My resident Apple fanboy / guru (his term not mine) tried it and it seemed fine on his Mac running Firefox 3.0.3. Well I restarted my Firefox today for the first time in ages (yeah I'm that guy) and noticed that there was an update to Firefox to be installed, 3.0.3. Wait, the client had 3.0.1 didn't they? "Hey client, can you check if there's an update for Firefox on your computer, just for shits and giggles and all?" "Hey, there was one, weird because I have automatic updates on as well... oh and look at that, it's working now" "Are you shitting me?" "No, seriously it looks fine now, must have been a bug in that version." Must have been, probably will never know. Anyone else ran into any weirdness like that with Firefox 3.0.1?

Issue with SQLyog Job Agent on Ubuntu

Josh Sherman — Sat, 27 Sep 2008 18:28:00 -0400

No error message, it just says that the sync had started. What sucks about this issue, is that I ran into it with a previous version of SQLyog Job Agent (sja) but was able to troubleshoot it faster because the program would dump an error message. I have been messing with the issue very sporadically over the last few weeks and had not come to a conclusion. Some luckly Googling this afternoon let me to the same answer I had before, sja is looking for the MySQL socket in a different place than where Ubuntu thinks it belongs. Long story short, sja is now working again, and ShowsTonight.com's database can be in sync more often. An honorable mention should go to Maatkit: a toolkit of utilities and tools for MySQL as I tried to indirectly fix my sja issue by migrating to using Maatkit. Unfortunately, I found the comparisons to take an awful long time, and didn't feel it would suit my needs compared to sja. Good thing I got it working I suppose. It's a pretty quick fix, one command:

sudo ln -s /var/run/mysqld/mysqld.sock /var/lib/mysql/mysql.sock

For reference, I am working with Ubuntu 8.04 and SQLyog 7.02

Re-establishing my web presence

Josh Sherman — Fri, 26 Sep 2008 20:00:00 -0400

So it's been a while since I've had a dedicated web presence for myself (roughly August 2005 was the last time). That's not counting the short lived static content site I had up on Joshtronic.com prior to this site. Anyway, long story short is, I'm re-establishing my web presence under the Joshtronic moniker. Why Joshtronic? Well because it seems there are quite a few Josh Sherman's out there with different variations of the name (I personally own joshuasherman.net but plan on letting it lapse... dot com's are where it's at). Seems every time I have a blog on a site, I end up posting the infamous "sorry it's been a while" type posts every few months. This time around will hopefully be different. First and foremost, no apologies and no regrets. Second of all, I run into problems that need a solution, so I figure, why not post about some of that stuff, and potential solutions that I ended up trying and what ended up working? That would force me to post on the site and also keep it from being so "dear diary" type drivel. This would also give me a bit of documentation on the problems, because I usually don't write out the steps to resolve the problems I've solved. I usually end up six months down the road wondering WTF I did previously to get it working. So yeah, I am Joshtronic, this is my site. Also, I'm aware that the site looks a lot like Daring Fireball to quote Picasso (yes, Steve Jobs stole the line) "Good artists copy. Great artists steal."