So, currently, my Linux development system is a manually stripped down version of Ubuntu off of a default install of 11.10. It runs Unity and it runs beautifully. I am very pleased that I have ended up at this point and will be sticking with Unity for a while until either Canonical effs it up like Gnome did with Gnome 3 or until Gnome becomes more usable. Linux Mint Cinnamon is also around the corner and Elementary OS is maturing quickly, so only time will tell for them!

The title of this post sounds like something out a christian revival blog…

Ultimately, I’m disappointed more than anything. Panels worked really well and was never complained about. Fixing something that isn’t broken always has backlash like this and I wish that the Gnome team had thought this out from a usability standpoint more before just deprecating Gnome 2. For now, I’m going to have to waste time migrating to another desktop environment and then, maybe one day, I can come back to Gnome and create a desktop environment that I can actually get work done on again.

The worst part is that everyone is thinking what I just typed, very few members of the Gnome community are satisfied with these changes and are having to find other desktop environments. The world of desktop Linux users has literally been sent into a chaotic mess with this blunder by the Gnome project and I truly believe its a big setback for the entire community of desktop Linux users.

I know the main argument that you will here in response to this is “I’d rather spend time programming and making money than learning how to configure my system.” There is something called a capital cost people, spend a day or two (yes, it only takes a day or two to learn most things) and you will benefit from it forever. Then you don’t have to pay for your OS anymore, you will become a stronger programmer because you can use the terminal (IMO you’re not even a programmer if you can’t use the terminal, that’s just sad), and you can stop complaining and sounding ignorant on Google Plus posts .

Am I way off here in assuming these are the reasons why some developers hate on Linux? I can understand certain situations such as developing on a mac so you can make iPhone apps (even if it is evil). The issues that I was addressing are clearly different from that however.

I found an article by David Eisinger entitled “Backup your Database in Git” in which he suggested using mysqldump and simply adding the dump to the database. While I had hoped for a more elegant solution than simply dumping your database every time, it would certainly work. He also suggested using cron to schedule dumps. David wrote this article describing a way to backup a production site, which cron would work fine for, but for me it would have to be a little more complicated as I wanted an updated database on every commit.

To automate the process of dumping and restoring my database dumps I decided that git hooks would be able to do the job. I had recently worked with git hooks as a way to deploy website updates directly from a git database. Using the pre-commit hook and the post-merge hook, we can create a system  that will automatically dump and add your database to each commit and update your local database from each pull.

Let’s start with pre-commit. The pre-commit hook will run a script directly before a commit is executed. To edit your pre-commit hook:

[your editor] /path/to/your/repo/.git/hooks/pre-commit

Now, lets write the pre-commit script. We are going to tell the system to dump our MySQL database to our git repository and add it to be committed.

1
2
3
4

#!/bin/sh
mysqldump -u [mysql user] -p[mysql password] --skip-extended-insert [database] > /path/to/your/repo/[database].sql
cd /path/to/your/repo
git add [database].sql

Now, lets write the post-merge script. We are going to tell the system to restore the MySQL dump to the local database for the latest changes. Edit the post-merge hook with:

[your editor] /path/to/your/repo/.git/hooks/post-merge

And write:

1
2

#!/bin/sh
mysql -u [mysql user] -p[mysql password] [database] < /path/to/your/repo/[database].sql

Note that in both in the mysqldump and mysql commands, there is no space between the -p and the password.

That is it! Now your MySQL database will be pushed and pulled with the rest of the commit and the pre-commit and post-merge hooks will handle the importing and exporting of the dumps.

I propose something like an Eclipse style IDE. The IDE runs a  portable web server on the local computer and will test for you at the click of a button. Built in git support is also a must. That alone would save me the copying of Notepad++ Portable, Git Portable, and XAMPP portable. Instead it would be an easy to use and quick to get running. Another issue that would need to be addressed is the unnecessary networking restrictions often put in place by administrators. Currently XAMPP Portable has an issue running on it’s own and requires extensive modification to run. With a little bit of checking the IDE could figure out the required settings for the host computer and run without complaint.

I may attempt to develop this in the future because unfortunately I am far too busy with my senior year to be able to take the time out to do a project of this caliber any time soon. I am certain it would end up unfinished. Also, as I will be graduating and moving on to college next year, my need for this sort of thing will be gone soon. I am sure that I am not the only student out there with this problem however, so I decided it would be a good idea to outline the idea for someone else in the future or if I get really bored in college. No idea if that will happen or not, lots of things are set to change for me in the next 12 months.

When I began running a few VPS’s I experimented with a few different distributions. I started with Ubuntu server and soon realized that this was foolish as Ubuntu runs a lot of stuff by default that I didn’t need. I decided to move to a more minimal OS. I decided on ArchLinux. I still love Arch, I really do, and its a great system, but I missed the familiarity of a Debian based system. The subtle differences were getting annoying. I realized, why don’t I just use Debian? I began researching Debian and realized that I should have been doing this all along. Debian has all the familiarities of Ubuntu without all the bloat. Also, Debian is insanely stable, I’m pretty sure the most stable around. I can trust the updates on the Debian’s stable track not to break my system. I might not have the latest versions of software, but I don’t need it.

This weekend, I moved to Debian on my laptop computer, which is my development machine. I am not looking back. I really love the speed. It’s something that I never knew with Ubuntu. Debian feels no different from Ubuntu, but it preforms so much better. Debian still runs Gnome, uses apt-get, etc. but it does literally everything faster. It boots faster, it runs faster, it shuts down faster.

I still feel Ubuntu is an excellent first step for someone who is new to Linux. Debian certainly requires more command line use, you will get a little lost in Debian if you have not yet become comfortable in the command line, or at the very least you would be uncomfortable, but once you get to the point where you use the command line more in Ubuntu than it’s fancy tools, you’re ready. All in all, Debian is a logical next step for a Linux user who is ready to move on from Ubuntu and I would encourage you to try it, you’ll love it.

So, this website is now dedicated to containing the thoughts, discoveries, and opinions of Ben Kulbertis in my ever ongoing journey into the realm of technology. The content of this site has no particular topic and may range from a great and many different things.