Paul Sokolovsky

> Ok, I added a generator for the idc Python scripts: http://www2.futureware.at/%7Ephilipp/ssd/disasm.html now how do I run it? Analysis -> Run plugin crashes when I enter init.py there. The easiest...

> I always have to undefine all that stuff first Yes, that's by design, doing it otherwise would lead to (more) mistakes. > undefine always only works a single value....

I have 2 news: good and bad. Good: with your latest init.py (size: 1244310, md5sum: 157a4c80860932f3d5f1691d46baa654), from-scratch analysis result can be saved and disasm-written without exceptions. Bad: this latest init.py...

> Ok, now I hopefully implemented the BLX handling correctly. Please try again. Better, much better. Still some fine details to take care of. > SetRegEx(0x00015C04,"T",0,2) #0x00015a96->0x00015c04 Thumb Supervisor 0x00015a96...

> and easily switch between those options You seem to miss that "easy switch" also means "easily switch by mistake". Let me again point that the act of disassembly is...

> Ok, perhaps I got it right now, feel free to try again. Getting there. Reality just gives examples of what I wrote above ;-) : ~~~ MakeFunction(0x0000BC72,4294967295) MakeNameEx(0x0000BC72,"fcn_0000bc72",1) SetRegEx(0x0000bc70,"T",0,2)...

> It is jumping away from Thumb mode, uses BLX to a direct address, ... but since it's an odd address, it must be Thumb code. I assume you means...

> Ok, please try again. Great, I don't see any artifacts caused by init.py. We can take the disassembly produced as a baseline, and now start to look how it...

> this is a closed one anyway. D'oh, I mixed up with something else ;-).

Anyway, continuing here re: issues with init.py, this time, more "serious", like possible problems with your tracing (or maybe openocd issues again): ~~~ SetRegEx(0x000146ba,"T",1,2) #0x000146ba->0x0001e5e0 Thumb Supervisor 0x000146ba 0x4788 BLX...