[Written in September 2003, a couple of days after we made a software release]
The blue wheel of checkin was in progress. It would stop occasionally to indicate an error and the person sitting at the screen would passively press the "Enter" button, thinking, "we'll see this once the rest of them are over". After all, this data cannot be lost. The hungry souls around him (they were hungry inspite of the fact that they just gobbled up close to 6 Pizzas, salads and quite a bit of 'pesticides') are all looking at the "show". A few sleepily, a few with their hands to their chin/cheeks and elbows on the table, a few scrolling a lifeless browser window up and down with the mouse wheel, a few working away to glory on the few pages that would announce the hard work of many people to the others (incidentally, they call it the release notes), and a few of them thinking as to when all this would end. Welcome to the GGAS 3.2 Release scene :-). The earlier releases were no different. Except that the "screenplay" wasn't documented by yours truly.
The frustration of a few mistakes, the anger and then the dialogue: "They shouldn't do it this way", "How many more times?", "What's going on" and it all subsides a little later. It is late, already. A few look at their watches and at each others faces. They walk away. They drive back home. They are all tired of the many late hours they had put in since a few days and today (Oh! it is already tomorrow. Past 12:00 Isn't considered Today ;)). And the rest remain.
After a midnight chai, the rest get back to check the sanity of the release. It breaks! Very much in the spirit of what they say "All hell broke loose". A few very anxious moments. Many theories why it could have happened.
Was it the last moment checkins that we did? Why doesn't 3_1_11.pl campon? Hmmm, ARFCN 214 is in GSM 850 band and we just compiled it out. So yet another busy (partial) clean and build cycle. We are through with the camping phase. But the mother of all CS procedures, the MO call, fails. Is it because we did not do a complete clean-build?. So let us try that. As those compiler messages scroll past the dark cygwin window, a few are busy testing other features. Thankfully, they work. The build is ready but the scripts 3_1_11 and 3_2_11 fail, with a certain hazy consistency in the failure. Thoughts in the minds of those half awake were going like ... "Could it be because of an incorrect merge?" Why not try this on an older build?. The fate of the outcome of the test on older builds wasn't all that confidence boosting either. It was already 4:00 Am. What is gonna happen to this release? Could it be a problem with the boards? Were the boards overworked just like us? Let us try and hard-boot them once. And as we were on our way to do that, we realize it was a minor script issue. Roll that back, and we were back on track, and we were as confident about our release as we were a few hours ago. It was just a few more minutes before the sanity testing was through and the release was made. The ride back home on airport road with cool breeze striking hard against my face was equally satisfying.
While we were busily testing our stack, I got a mail in my Inbox which announced the release of the GNOME 2.4.0 Desktop & Developer Platform. The 2.4.0 release was named after Chengis Khan (whose real name is Temujin) and it said this about his army ...
... Often referred to as 'the Mongol hordes', and considered barbaric, his army was, in reality, a thoroughly modern military outfit in medieval times. They were swift, precise, equipped with superior technology, ingenious in their approach to war, keenly organized and consistently outnumbered - not that it detracted from their exploits on the battlefield.
Every word in there reminds me of all you wonderful people that I work with. Great job team. We are the best. Anybody to contest the claim? *me rolls me sleeve up*.
