View allAll Photos Tagged debugging
Chasing a stubborn problem with CC2500 radios last night. Eventually it turned out to be a problem with compiler settings for the cross-compiler generating code for a slightly different platform. Because of the similarities the firmware image didn't fail outright and made the mistake that much harder to detect.
DeBug Computer Nevada Day Parade 2005
Visit us on the web www.debugcomputer.net/?utm_source=flickr&utm_medium=p...
It would be impossible for me to overstate how frustrating today was. Effectively I spent 9 hours debugging a problem that was (as far as I can tell) completely non-deterministic / irrational. I swapped out everything, software and hardware. I reset everything. I jiggled the thingy. I rebooted stuff. I rebuilt. And then, just as magically as the crash had shown up, at about 5pm it just disappeared. The most horrifying thing is that for every thing that I did in the 30 minutes leading up to "it works", I can pretty decisively argue that thing as being NOT involved in the problem.
There are days that really make me question how cold and rational computers really are. If I was just slightly more superstitious, I would point at today and say, "There was a demon sitting on my desk today."
This is take-out dessert from the local bakery. In shooting this, I decided to consciously try the opposite approach that I would ordinarily use. So instead of starting with a wide-open lens, I'd shoot as narrow as possible (this is f/36), and instead of bringing out a big diffuser like my favorite umbrella, I'd shoot with two bare bulbs and see if I could tame the light with reflectors and in post. I'm not wholly convinced it worked, but it was at least fun to try.
Victoria is flying out to visit her folks for a few days starting tomorrow. So even if long hours continue, I'll start being a proper flickr friend again tomorrow night.
This is my 8x8 array of leds and buttons. The LEDs are 2 color LEDs (which give a third color if you switch them on together) and as you can see, there were some errors in the matrix :)
The Z8-02 bondout is connected to a 28C16 EEPROM serving as the emulated ROM. (Typically a 2716 EPROM would have been used.) The 28C16 has been programmed with a copy of the masked ROM from the Z8671 Single Chip Interpreter BASIC/Debug. An async serial to USB adapter is connected to a laptop running Kermit.
A simple but inefficient BASIC program to find prime numbers has been entered, and the LIST command used to list it back out for verification. The program is RUN, then stopped after a few prime numbers have been found.
It's true, errors are the prettiest part of graphics programming.
This pretty garbage haunted my late nights for a while as I finished my (free) Photoshop PSD Extract/Recovery plugin, telegraphics.com.au/sw#psdrecover
DeBug Computer in the 2008 Nevada Day Parade
Visit us on the web www.debugcomputer.net/?utm_source=flickr&utm_medium=p...