[kismac] Re: Intel port -almost- complete

  • From: Geoffrey Kruse <gkruse@xxxxxxxxxxxxxx>
  • To: kismac@xxxxxxxxxxxxx
  • Date: Thu, 23 Feb 2006 16:15:06 -0700

Thanks, I got this figured out.

On Feb 23, 2006, at 4:07 PM, Robin L Darroch wrote:

Yes - you need to download, build and install the AMRollOverButton palette from http://www.harmless.de/cocoa.html#rollover

(it just goes in /Developer/Extras/Palettes)

Then next time you open a NIB in Interface Builder, tell it you'll help it find the RollOverButton palette.

You only need to do that once.

Thanks for the quick response.

I have one other unrelated question:

I'm unable to open the .nib files in IB, I get the following error:

Palette (RollOverButtonPalette) was not found

Any ideas on how to fix this?


On Feb 22, 2006, at 10:02 PM, Michael Rossberg wrote:


next time you can write me directly =). swapping is not really something you want to do here. because it is slow and this function needs to be fast. it also does not matter if it returns a different value on intel machines, as long as it always returns the same on that machines (wont affect saved files). however currently the function will always return 0 or some in defined value. you will need to remove the +1's and +2's i belive...
the fast lookup is indeed not used currently because of its intense memory usage...
good luck


On 23.02.2006, at 05:14, Geoffrey Kruse wrote:

Just writing to let everyone know I have made a large amount of progress on the intel port. Saving and scanning is -mostly- working with both Airport Extreme (active) and prism 2 usb. The largest remaining issue is giving me some trouble though. The hashformac() function seems to return different values for the same input on the intel machines which causes the same net to show up multiple times. I'm going to continue to look into this, but I thought I'd ask the list to see if someone (mick?) can figure this out more quickly than I can. Here is the function:

inline UInt32 hashForMAC(const UInt8* val) {
    UInt32 l, j, k;

    //add to hash table
    memcpy(((char*)&l)+1, val, 3);
    memcpy(((char*)&j)+1, val+3, 3);
    l = (l ^ j) & 0x00FFFFFF; #else
    memcpy(((char*)&l)+2, val,   2);
    memcpy(((char*)&j)+2, val+2, 2);
    memcpy(((char*)&k)+2, val+4, 2);
    l = (l ^ j ^ k) & 0x0000FFFF;

    return l;

My guess is maybe they bytes in UInt8 * val are swapped or something but I'm not really sure how to fix this. maybe with swab()? FASTLOOKUP seems to not be defined on the code, so it is using the second method to create the hash.

Hope to get this finished up soon, so we can get a little more heavy testing on it. I'm also pretty sure that all cracking methods are still borked but I have no way to test this. We can fix that later when I get my hands on an intel mac.


p.s. Sorry to the non-code oriented people on the list, you can just ignore this

Attachment converted: Macintosh HD:smime 26.p7s ( / ) (00E19D73)


---------------------------------------------------------------------- ---
Robin L. Darroch - PO Box 2715, South Hedland WA 6722 - +61 421 503 966
robin@xxxxxxxxxxxxx - robin@xxxxxxxxxxx - robin@xxxxxxxxxxxxx

Other related posts: