Abnormal Behavior on Disassembler Widget

[halsten]

Hi,

I currently was using the latest development branch on a 32-bit machine on Windows 10 Professional. And I was using the sample provided with medusa as input to see how is it going to behave. The sample in question is the "Semantics Loop" sample. Once you open it and medusa finishes analyzing clicking on 1 of the XREFS of any of the function should take you that function, however it took me to the bottom of the screen and everything else disappeared. As the lack of a vertical scrollbar disabled me from getting back where I was, I quickly used the function list to the output again. This is where shit hits the fan, and the editor starts to behave abnormally. Resetting such behavior was achieved when I went to settings and changed the font to probably force a complete paint on the viewport. This fixed the issue, however trying to click on the second line of the editor where "start" was located it would only reflect to the line above it (line number 1). and that line became totally immune to my several attempts to click it.

--halsten

[wisk]

Hi halsten,

I think this problem appears when qMedusa failed to retrieve cell information from the current address.
Could you tell me the exact executable name please?
BTW jumping on XREF by clicking on it is still not supported, so I guess qMedusa tried to jump on the first operand reference. If the operand wasn't a correct address, it could make it jump to an invalid address, thus when qMedusa tried to refresh the disassembly view, it failed and gave up, that's why the view is empty.

[halsten]

Well, if I remember correctly, it was either the sample file symbolic something executable or one of the MASM binary files. I have to check the VM and it's not with me at the moment. But, its within this range. I had an idea of why this might happen, but wanted to submit a bug anyway to be tracked and fixed. Let me see when I get back and upload the binary file in here.