Developing a new software or hardware product can be a lonely endeavor: making new features, testing them out, improving them. Or sometimes, sitting for hours behind your screen and oscilloscope, wondering why something doesn't quite work as expected. This can be somewhat discouraging, but fortunately every so often you get a boost of energy when you finally finish a new feature, or you find the reason behind an issue you've been trying to debug for hours. The fix usually follows quickly after that.
After spending a lot of time on development, it can be very helpful to get another set of eyes to look at and test your product. They may notice things that you overlooked. So, I sent an AppleSqueezer GS to France a while ago, to Antoine Vignau, known from brutaldeluxe.fr (II GS software). He tested it in his ROM03 machine, and I'm glad to report it mostly worked very well and as expected. This is especially good news for me since I sent him an earlier version of the AppleSqueezer before, which worked in my IIGS but perhaps somewhat surprisingly (since we both used a ROM03) not in his. After that, I went through the code again and stared at oscilloscope waveforms for some time to figure out what could cause it to work in one machine, and not the other. It turned out that the timing I used for putting addresses on the IIGS bus had been a bit too slow. Apparently, this is ok for certain IIGS's, but not for others (this could be because of chips made with slightly different tolerances, from different factories, etc.) Fortunately I had a pretty good clue that this was the reason after putting the card into my freezer :). This also causes slight timing differences, and while in this cold state it wouldn't be reliable even in my own IIGS. So, I went to work to shave some nanoseconds from my cycles to get a comfortable margin, and this fully resolved the issue. It's now reliable on his machine as well, as well as on my ROM01 machine that I got donated to me by a generous member of the Apple II Enthusiasts group on Facebook, Golan Klinger. Antoine also tested it on another ROM01 machine, which also performed as expected, and he tested it with more hardware, including the VidHD and CFFA3000, and a 3.5" 800K disk drive. All worked well!