Lovely big packets, full of data. Now I add different kinds of data. Easy enough on the Hamster end: just generate the data and command the Hamsters to write it out to the socket. The real difficulty comes when I try to parse it out of the data stream on the PC side. I add some code that detects the new Type B data and sends it to a set of files separate from those that collect the old Type A data that we've been messing with all along. This works, except that it leaves a hole in the Type A data stream, which is graphed on the display.
I'm almost ashamed to say how long it took to figure out that the blip in the graph was a gap in the data stream, and not a result of junk data or a bug in the Type B data extraction routine.
For cosmetic reasons, the gap in the display must be filled (no gap appears in the files, which are actually more important). The logic involved is getting really ugly. It's what I have to look forward to tomorrow morning. Perhaps I'll work it all out in my sleep tonight.
200 words | July 20, 2004 07:33 PM | Rocket science