In message <AT5FSeBoP0EOFwC4@xxxxxxxxxxxxxxxxxxx> David Pilling <flist@xxxxxxxxxxxxxxxxxxx> wrote: > A good old fashioned bug in DPScan may be the cause, say a dependency on > the contents of a memory location. Perhaps an illegal address or one > that is not being correctly initialised. > I'll do a test version that will explore these possibilities. > I think one interesting point, is that Martin seems to have located the > bug at the same point in the code as the original user. The point being > that Martin has a production version, whilst the user had a "debug" > version with function names in. That means the code is at different > addresses. Yes, that is interesting. > Do we know if save to another application crashes - this won't work with > all file types, but a save of a file in sprite format to Paint should > work. Yes, saving to Paint or Draw crashes in the same way. It is worth noting that in all cases the transfer has completed successfully at the time the crash happens, i.e., the sprite appears on disc or in the target application. Maybe the crash occurs when DPScan processes the acknowledgement message? There is another interesting thing: As Chris has reported already the crash goes away if other applications are loaded *before* DPScan, and that is perfectly repeatable here: 1) load DPScan 2) load Paint 3) load sprite into DPScan, save to Paint: crashes OR save to disc: crashes 1) load Paint 2) load DPScan 3) load sprite into DPScan, save to Paint: works OR save to disc: works What is more intriguing is that quitting both DPScan and Paint restores everything back to the starting position, i.e., you can once again try one of the above scenarios, and you get the same results once again. In the second scenario (Paint loaded before DPScan), you can quit Paint and DPScan still continues working, but if you quit DPScan, too and reload it, it crashes when saving again. And just to muddy the waters a bit, I had Zap running all along and its presence does not influence the above scenarios. Obviously, the thing to do is to run DPScan under a debugger to get more information and to find out whether the instruction at the point where the crash occurs has been corrupted. Unfortunately, both DDT and DeskDebug crash when trying to run DPScan. Martin -- --------------------------------------------------------------------- Martin Wuerthner MW Software lists@xxxxxxxxxxxxxxx --------------------------------------------------------------------- To unsubscribe or subscribe goto: //www.freelists.org/list/davidpilling