We’ve uncovered a slew of things to work on. The first obvious place was to have the static content served by Apache rather than the mongrels. This has been on the todo-list for a long time, but has always fallen to the wayside. We knocked that off and looked towards the other items that were slowing us down. That’s when I uncovered some TCPSocket weirdness within memcache-client 1.7.4 that comes with activesupport 2.3.5.

During some tests we noticed a severe lag which we narrowed down to the fact that we were pointing to a list of memcached servers that contained one that didn’t exist. Turns out that when the memcached hostname resolves, but is not pingable, memcache-client 1.7.4 waits 3 seconds before responding with an error message (in addition it doesn’t mark the server as “down”, which I think is also a bug). This 3 second delay happens on RHEL 4, and in some brief tests on Ubuntu 10.9, it was even worse, taking over 30 seconds to respond. My guess is that there is some OS level setting that affects this, but I have yet to locate it. The fun part however, is that this problem does not exist in our old environment where we run acivesupport 2.1.2 which uses memcache-client 1.5.0.

Turns out, in 1.5.0, the memcache-client uses a 250ms timeout when calling TCPSocket.new. Something that was lost on the way to 1.7.4. Some initial tests of simply adding this CONNECT_TIMEOUT back in have been promising. It’s currently not throwing the right exception or marking the server as “down”, but once I do that, I will see about posting the source somewhere.

There was talk of iChat and sharing screens and some other not-so-simple-to-setup environment stuff, but in the end we coordinated what we were working on via chat and agreed we’d pair up on Thursday to review changes/updates we made or wanted to make. Even though I ended up doing the same on Thursday (no plow on my street all day, I couldn’t leave the house!), I was able to be productive and reviewed my work today with some team members.

It was weird working from home again. It was definitely different – I had to use my home laptop and my office space was a bit of a wreck. It took a little while to get totally setup with the VPN etc, but I ended up cranking out some good work.

My final conclusion is that I really like pairing, and working in an open space. Maybe I have great teammates, maybe I’m still in the honeymoon phase of pair programming, who knows, but for now, I’m loving it.

1. No more personal bathroom
2. Coordinating my attire
3. Not being able to pick through leftovers for lunch
4. I can’t (easily) hit the gym at say 11am.
5. No laptop, no cell phone
6. No one cared if I had bad breath when working from home!

Best parts of my new job:

1. The people – incredibly smart, and unbelievable willing to take time out of their day to explain how something works or why something was done a certain way. The openness is infectious.
2. Pair programming with really smart people, like David Black
3. No email. OK not ZERO email, but with everyone working in the same place, email is rarely used.
4. No conference calls. Literally none. I had one on my first day which was a “Welcome to Boeing” deal, but zero since.
5. Ruby & Rails – No more Java, no more WebSphere Portal
6. No laptop, no cell phone. My day is now done when I walk out the door. I can VPN in to check my mail, but see #3

It comes as no surprise to me that the toughest things am dealing with so far are all related to no longer working from home. I’m not completely over the fact that I now get up at 6 & get into the office at 7:30, but I think I am getting used to it. I feel like an old man sometimes going to bed before 10, but I could be doing worse.

Of course the title of this post was written to make you want to read it. I’m really enjoying the new job but need gimmicks to get people to come read my stuff 

At IBM I did a lot of development, in what I now realize was in a non-traditional environment, for better or worse at times. Skaven is hardcore into Agile. It’s actually a fairly thin layer on top of what I was most doing at IBM (other people are worrying more about the semantics of Agile right now), but it adds enough structure & reporting to make me feel like management is aware of what’s going on. It’s kinda cool to be learning about Agile at the same time as a Sprint is already in progress – it’s a great way to pick up on things quickly.

In my new position I’m going to be doing Ruby on Rails development, which means I have to actually learn it! I’m currently going through lotsa of learning: the system, the development environment, Agile, scrum… and the roads to get there! I’m really psyched, and I can’t wait to contribute some real work. That should be coming in the next few weeks after I “pass the Qualifications Board”.

I’m going to start trying to blog more as I figure out exactly what I’m doing 

Here’s the thing…. it sucks!

I’m no more lazy than the next person, but come on! I mean, I can turn the console on & off with the wireless remote. I change the turner on my TV with my remote & I’m off & playing .. as long as the game is in the console.

So here’s what we do:  I usually keep my games close by my console anyway (for easy access – cuz they need to be in to play) so why not install some kind of RFID or wireless thing-a-ma-bob that is part of the game case. The console can just check if it’s in range – and preseto – I don’t have to get off my lazy butt to play!

After much searching that yielded a lot of results about storing executable flags in SVN using propset. This didn’t help as I didn’t want it to be executable, just world readable. Turns out the problem was in my .profile, my umask setting was 077. Changing it to 012 solved it for me. Note: 012 makes files you create group writable.

Snippet from my .profile

#umask 077
umask 012
export SVN_SSH="ssh -l olore"
export EDITOR="/usr/bin/vi"

Read more about umask on Wikipedia

This is a step towards running multiple versions withing WebSphere Portal, which I’ll likely be dealing with in the near future.

Our portal is pretty open, we want people to come in & customize it as much as they’d like. One feature that we allow is for the user to come in & create new pages. We are running our own theme policy which is a copy of the DoubleTopNav policy. Our theme displays these 2 levels of navigation at the top, and we have no side navigation.

The hurdle I ran into was – how do I prevent the “New Page” option from appearing in the portlet menu for pages in the second level of navigation? The way the out-of-the-box Portal theme works is that when you create a new page, it is created as a child of the page that you are on. This works fine for top level pages, but there no policy or rule that can be set to prevent the “New Page” option from appearing in the portlet menu for second level pages.

The solution, proposed by a member of the Portal team was to create an alternate pageContextMenu.jsp which simply didn’t have a “New Page” option. Since the call the that JSP is done via AJAX, prior to the call we can determine what navigation level we are on and call the appropriate version of pageContextMenu.jsp.

Piece of cake, right? Eh .. maybe not cake, but not all that difficult. After creating the new pageContextMenu2.jsp & removing the “New Page” option, the next place to go is head_inlineJS.jspf. Below is the code I have added/modified to the head_inlineJS.jspf.

Now that you’ve got that taken care of, touch head.jspf && touch Default.jsp, reload your page and you should no longer see “New Page” as an option in the page drop down for pages in the 2nd level of your navigation.

That worked and was a thing of beauty .. well, almost. I wasn’t a big fan of looping through the entire navigation model, but I didn’t know any other way. I ran the code past my colleague from the Portal team and he pointed me at the navigationSelectionModel. The code below is smaller & faster. Best of all – it works

With tweets like

• silver jeep, FLB 272. thanks for cutting me off not once but TWICE when you were going straight in turn only lanes prick.
• To the girl in the Green Explorer outside Muldoon Fred Meyers this afternoon: Sorry my U-turn scared you. My bad
• white Nissan pu FEL 365 hang up ur cell phone pay attention 2 ur driving. ur lane ended & you nearly hit me merging 2 mine.

… I wanted a piece of that action!

So I started @NJBadDrivers and @CTBadDrivers. Then I realized I better grab @NYBadDrivers to complete the tri-state trifecta. Imitation is the sincerest form, right ? After talking to my buddy who moved up to Alaska, he said that people really are crazy drivers up there, but come on, can they really be worse than New York, Connecticut or New Jersey drivers? (Admittedly, I’ve had licenses in all 3 states.)

I’m running with Christopher Finke’s slick little python application called retweet on my Dreamhost account. If Chrstopher’s name sounds familiar (it should!) it’s cuz he’s the mastermind behind ScribeFire and the TwitterBar Firefox add-on. For those interested in running retweet on Dreamhost, be sure to use python2.5 as it already has the sqlite3 library configured. Once I figured that out, I cron’d it and it’s been working like a charm.

So how’s it work? All you need to do is send a message to one of the accounts via Twitter and it will get retweeted by the bot. Just say “@NJBadDrivers does anyone around here know how to function at a four way stop???” or “@NYBadDrivers srsly people, traffic circles are not that difficult”.

I wonder… is reserving Twitter names the new domain squatting? If so, would it be called twatting?