denise: Image: Me, facing away from camera, on top of the Castel Sant'Angelo in Rome (Default)
Denise ([staff profile] denise) wrote2009-07-13 03:14 am

(no subject)

* [staff profile] mark continues to do the monday updates this month! let him know if you've got anything.

* I had 51 patches in this push. That's crazy. Okay, some of them were codemerges from LJ that I just packaged, but still. Mark is threatening to take away my suit.

* I do, however, get the "I Broke The Build" badge of shame: two of my patches required emergency bugfixes after being pushed live, no matter how much I tested them. I console myself with the fact that one of the two emergency bugfixes was something really small and fiddly that I was not, in fact, really expected to notice. (The other, I have no excuse for, but at least someone else came along after me and broke it just as badly. *G*)

* I'm sure there are things that are still broken, so comment to [site community profile] dw_maintenance :)

[personal profile] desh 2009-07-14 11:44 am (UTC)(link)
Ha, I like it!
owl: Charlie Eppes. Geek. (geeky)

[personal profile] owl 2009-07-15 09:47 pm (UTC)(link)
Oh, yes, nice :)

Anyone who has ever coded in a team environment has broken the build, probably in the first month. Our record in work is four times in a week, including both the two seniors. That was not a good week, we ended up having Apology Buns on the Friday :)
owl: Motherboard and CD (computer)

[personal profile] owl 2009-07-16 07:55 pm (UTC)(link)
Our team is great for buns and general bonding over food. One of the boys swims on Thursday mornings, so he does a Thursday coffee run. He'll phone up from the coffee shop and take orders (he can tell who isn't in work by whether the order's missing, for example, a vanilla latte or an americano, NO MILK), and then one of us will go to meet him and help carry them back.

The week of shame went something like this:
Monday: I check in changes that weren't as well tested as I thought they were, and which were giving null pointer exceptions. Easygoing Senior spots it, gets me to back it out and pats me on the head in comforting yet slightly patronising manner.
Monday evening: Easygoing Senior pulls the classic 'check in a partial patch' move and heads off home.
Tuesday morning: Everyone else syncs their code and find it no longer compiles. Easygoing Senior commits the rest of his changes, Neurotic Senior is snarky.
Wednesday: One of the junior devs accidentally commits every single diff he has in his workspace, including settings files and archive files. Project no longer compiles. Neurotic Senior loses patience and stands up at his desk to tell the team at large that "we do not commit jars to CVS. We certainly do not commit our $&%*ing classpath to CVS!"
Wednesday afternoon: By comparing the repository with the local copy of someone who hadn't sync'd yet, hapless junior manages to get the build unborked.
Thursday: Neurotic Senior checks in a large change. It compiles, for a change, but turns out to throw runtime errors under certain conditions. The fact that two other people had checked in changes that morning as well complicates matters. Neurotic Senior frantically tries to fix his code in a half-hour break in an afternoon of meetings. The three other people trying to help get a fix are met with snarls, and his keyboard and mouse take a hammering. The fix goes in, just about. The meetings don't go well either, at least for the people who aren't us. I hide behind Neurotic Senior and let him fight the team corner, given that he's clearly in the mood for it. The juniors huddle around the kettle and complain, and message each other with things like "storms on the island. keep ur head down."
Friday afternoon: Neurotic Senior produces buns, rather sheepishly. Someone remarks that this is becoming a habit :)
owl: Motherboard and CD (computer)

[personal profile] owl 2009-07-16 08:22 pm (UTC)(link)
It was, rather. I'm glad that the frantic fix wasn't my responsibility, as there's nothing worse than trying to debug when your whole team is looking over your shoulder and twiddling their thumbs because their code won't run right.

I think the root of the problem was that the previous week we had been landed with a load of requirements that weren't in the original scope for the project. Hence N.S's overfull calender and general distress. (Most of them have now gone back out scope, after we wasted at least a fortnight's worth of dev-hours in analysis and design, hours which we could really be doing with right now.)
owl: Stylized barn owl (Default)

[personal profile] owl 2009-07-16 10:59 pm (UTC)(link)
All software development processes suck, but some suck more than others? I had pretty much this conversation with one of the other juniors actually the other day. He said it would be nice if future releases were smaller and more frequent. I said "You mean move in a vaguely Agile-wards direction? Nah, can't see them going for that." I think the business's idea is to have Waterfall-sized requirements in Agile-like timeframes. Not going to happen, folks...
kareila: "PERL!" (perl)

[personal profile] kareila 2009-07-14 01:39 am (UTC)(link)
That's almost exactly twice as many as I had (I counted 26).

I don't what's more shocking, you doing so much code work or Mark doing so little! But it's great that he can let us take care of the minor stuff and can spend most of his available time working on major new features.