neszt
Are you sure, the crash is because of taking the picture? Switch on the camera and then wait 10-20 seconds.
Yes I am sure. A strange thing is that the camera always starts in playback mode no matter which button I start it with. With the original firmware one of the buttons starts it in capture mode. But now it always starts in playback mode. When I press the shutter button now, it tries to switch mode to take the picture, extends the lens then crashes.
acid2000
Did you go through every line of ASM in capt_sec.c? You have to check it against your firmware.
No, and I was already thinking that this might be the problem, and now you kind of confirmed it. I was hoping to be able to take a shortcut for this beta build, so I did not went through every line of firmware code. I was hoping to only need to do that once the few bugs in the beta of 1.02b have been ironed out. What I did was to search for all memory addresses and update them according to the new firmware. I guess this is just not enough and it really is a pain as I am no ASM hacker. By the way, thanks for the usefull debugging tips from both of you neszt and acid2000. But as things now stand I might wait so that I only have to do it once, if at all. Since I do not have a experience with ASM code, I get kind of discouraged by the fact that I have to update thousands of lines of instructions. I truely respect all the work you two have done for the community, now even better.
asthro
Stange isn't it ? UltimA, what(s your point on this thing ? Did you update so much code for the 102d ?
Furtunately, I didn't have great problems finding the corresponding code parts between the two firmware versions. This was the first time I saw ARM code and that I used a disassembler at all, so it took a few hours to get used to it, but in general I found each code piece easily. Try calculating the memory offsets between known points in both firmwares, that helps a lot for orientation. In my case, I had almost everything above 0xff86f2xx shifted by 0x50 bytes, and at higher areas there was a shift of 444 bytes. So when I couldn't find something, I just added the usual 80 bytes offset, jumped to that address in the new firmware, and voila, there it was in almost all cases, or just a few lines away from it.
