use Perl

It seems the list of things I need (or want) to learn keeps growing. I know this is a normal state for me, and probably for most of you out there. Here's a small list of what I'm looking at right now:

Poe - For multiple work projects and because it looks cool.

Class::DBI - Having used SQL a fair bit, trying not to think "Well with SQL I could do this by ...." and using Class::DBI as intended has been... interesting.

eXtreme Programming - I've been asked at work what development methodolgy we use. Adhoc doesn't seem to be the answer they wanted. As XP appears closest to the method we're using I answered that we're moving in that direction. Now to learn about it, and teach the others.

Template Toolkit - Our application is currently developed with a PHP web front end. We've found that maintaining the PHP code in a multi-developer environment difficult. As such a migration to an MVC design is now in progress and TT will be our templating system. I'm reading the Badger book now, it's quite good.

mod_perl - Exact same reason as TT.

PAR - We have a number of servers that are missing Perl modules that our application relies on. If PAR works the way I think it does, it'll solve this issue without impacting our customers.

Above all of this is increasing my Perl skills. I've learned a lot in the last 6 months, but I know there is much more for me to learn and improve on.

There are many others, like I said this is the small list.

Today I spent entirely too much time trying to install modules on RedHat boxes.

My new notebook for work (IBM T41) is now running Fedora Core 2. I was attempting to update Test::Harness and kept running into a problem with one of the tests (test-harness.t test #40). Which is a known issue but for some reason I decided to find the cause. I didn't find it, but spent too long looking. Hopefully I'll be able to look again.

On a RedHat Enterprise Server v3, I've been trying to install CPANPLUS. There were a number of different errors that I looked through but wound up looking at the Makefile. The Makefile is a mess. Unmatched quotes, half a directive on one line, the rest as the value for another. Really weird. I have more to look into there too but it might be pointless as we might rebuild with Fedora Core 2 as we are not happy with RHES.

I'm writing this very quickly (I hope) while I have a couple of seconds. I keep meaning to post here more often but I never seem to have the time.

I was finally at a point in my personal life where I was starting to get time back for my projects. Having 2 kids (5 and 2) takes up a lot of time, and even more so with me working during the day and my wife working at night. The kids are just getting to the point where they don't require constant supervision and doing most things for themselves. Something strange must have come over me because I made the decision that we're going to get a puppy.

On August 19th we picked up our 8 week old beagle puppy. She's more of a handful than the kids have been (so far). There goes any free time I had.

It seems that everyone is preparing to leave for OSCON or is already there. Not me, yet. My work only approved for me to go for the 3 days of sessions. So my flight doesn't leave until early Tuesday morning, putting me in Portland around noon.

This doesn't mean that the anxiety hasn't started already though. First for leaving my family, then for flying, and then finding my hotel, then meeting people, then not fitting in with the people I meet, and on and on. It never really ends....

The other day, I needed a feature in DBD::Pg that didn't exist. I'd asked on the mailing list to be sure that I hadn't missed anything, and sure enough what I was looking for hadn't been done. I started writing a patch that would allow me to do what I wanted. I sent off the patch to one of the authors that I'd been introduced to for comments, to which he quickly replied. I was wrapped up with my day job, and after a couple of days I got back to following the recommendations, however just as I started, a posting to the mailing list indicated that my feature request had been incorporated into the latest CVS. My patch no longer required.

Today Andy [use.perl.org] announced Test::WWW::Mechanize (yay!). I was also working on a module that would provide similar features. Once again, moving too slow.

I've received permission from my manager at work to attend OSCON (they're paying for it). Now I just need to work out a couple of details before I register and book my flight (pager and vacation coverage).

I've never been to a conference of this type and I'm really looking forward to it. Previous conferences that I attended were put on by my work, which had a slight marketing feel to them. This time, I'm hoping to learn something other than marketing information, but more importantly, meet some people.

Part of my current project at work is to take a CSV log file created by an application server, parse it for entries that meet the criteria, and insert them into a Postgres table. In the past I'd read in each line, split it into 30 different variables (only 5 of which I care about) run it through multiple if statements to make sure it's the data I need to store.

I've been reading Programming the Perl DBI [oreilly.com] and had skimmed through the section on DBD::CSV. I quickly wrote up a proof of concept using DBD::CSV and was impressed. Getting the exact data that I wanted out of the log file was much simpler and straightforward with a SQL select compared to all the if statements.

Now to roll this into production.

I've started writing modules for my project at work. We're in the midst of replacing all our PHP with Perl and moving toward a MVC based design.

My co-workers have asked for an example of how to use the modules I've written. As such I'm going to rewrite one of my existing scripts using the new modules. Digging through the program I was trying to figure out what I was thinking when I was writing this one.

I had split a delimited line of text into a list (not so bad yet), then I'd used the list one element at a time to do various things. This is where the problem lies. I can not remember what each element of the list represents without looking at the input file. So much for easily maintainable code. Needless to say, this script is going to be rewritten.

If I've learned anything from reading Perl Medic it's the value of writing tests. This is probably old hat to the rest of you, but I just started (well 5 weeks ago) to use them, and I can't believe the impact they've had on my coding.

Previously my testing would be done by printing variables and return values of functions, relying on my eye to catch any issues. While serving me well it wasn't very good for my productivity. Since writing tests (using Test::More and Test::Exception) I've been able to write more code, with increased confidence, and better results

Now I need to create test suites for our web application using WWW::Mechanize and get the other developers on my team to buy into them. Wish me luck (although I shouldn't need it).

is posting to them. I get an idea of what I want to write about, then I think it over, and I ponder it some more. By the time I get to writing about it (or close to it) the event or thought has long passed. This entry for example, has been thought about for the past 6 days.

My previous idea for a posting (a project for work that I'm doing) was 2 weeks in this thought process and was never written. I still know what I was going to write, but now I have more ideas and the list builds. I'm now at a point where there's just too much to say and I end up saying nothing at all.