Arbitrary Deranged Digressions

Never made it to lunch today. Again. Oops. Going for a nice dinner, traffic permitting, so it's ok.

Around 11:45 my Project Manager came by to report a problem with the search engine on the web site I work on. I checked it out and sure enough, it wasn't working. No real idea why.

Did some debugging, figured out what was wrong. Called the author of the search engine. They had no clue. "That shouldn't ever happen." Thanks guys.

Figured out that rebuilding the index would get us through the crisis; that was a no-brainer. Getting it reindexed is another story altogether...

Problem 1- When we cut over from using ASP to using JSP, no one thought to port the utility scripts which, among other things, rebuild the index. So I quickly re-wrote that one. (3 lines of code. Not a big deal.)

Problem 2- After about 30 minutes, the re-indexing failed. First thought was that it timed out, since the ASP had been specially treated for your anti-timeout protection. So I figured out how to run the old ASP file to do the reindexing.

Problem 3- The ASP file had approximately the same problem. This is good in a sense, because it means that now I can be pretty sure that the JSP file will work for the future. But it's bad because it means that ... SOMETHING ELSE IS BROKEN.

Problem 4- (and this is where I get really peeved...) The code that performs the re-indexing is somewhat brain-damaged. The java is just screwed. It doesn't work even as it's internally documented. Took me 5 minutes to fix those problems, but that doesn't let me use the "new build" until we push out a new release. NOT going to happen today... well, not for this problem. So I figure out how to fix that and get the code to at least RUN.....

Problem 5- So I manage to get the java to start re-building the index... I put it into 'verbose' mode and notice that it crashes - with the same error as the JSP - on a particular item... I look at the item and determine why it's failing, and at that point things are ok... I can understand why the indexing code is breaking on this item. SO I restart the indexing and fix the product.....

Note that I (a) restart the indexing and THEN (b) fix the product. This brings us to...

Problem 6- The first thing the indexing code does when it starts up is load ALL the products it's going to index into memory. So - 35ish minutes after starting - I find out that it's crashing at EXACTLY the same place...

So I made sure that there were no other problems of this nature, and restart the indexing. Which brings us to now. It's gotten WAY further than it had been, and *knock on wood* should be done within the hour. 6 hours after the problem was reported. Yay.

When this is done, I have to prepare for the actual new release that's supposed to go out to the web site tonight. And I'm <SARCASM> REALLY looking forward to it</SARCASM>. It's the first release that I've done all on my own. No one else to [share any] blame for problems. Not that I really think there will be problems. The project is well-managed, and I made only small changes. Should be fine. *shrug*

So it looks like I'll make it to dinner tonight, and should have no problem getting home and prepared for the release by 11pm. Sounds like a plan.