Here is a picture I made showing the differences between the osos header in the 19.1 version and the 19.2 version.
http://img523.imageshack.us/img523/709/graphdiffsec0.png Red areas shows differences between 19.1 and 19.2. The new osos is 112640 bytes tinier than the old : 19.1 : osos.fw is 0x615800 bytes long. 19.2 : osos.fw is 0x5FA000 bytes long. And that's why the bloc located at 0x14 is different, it represent the size of the data area. A quick diff off the data area shows that the area between 0x800 and 0xE90 are the same. The rest of the data area appears to differ BUT the real problem is that the two areas differs in their size : For example a long bloc insertion in the area is a problem for a lot of binairy diff programs : If the length difference is cause by a 112640 bytes long insertion in our binairy the diff program can't tell us where that bloc is (every bindiff program I found works byte/byte) and just tell that every byte after the beginning of the new bloc differs. So, let's sum up : - The areas between 0x800 and 0xE90 are in both versions. - Both differs in size by 112640 bytes. - The last bytes of both differs. (So it's not a simple bloc insertion.) I first think the bytes located from 0x18 to 0x2F are bloc hashs of the data area because the first 4 bytes bloc are the same in both 19.1 and 19.2 and moreover data from 0x800 and 0xE90 are the same too. But it's probably wrong because there is not enough hashs for the data area. Does anybody got fresh ideas ? JD. _______________________________________________ Linux4nano-dev mailing list [email protected] https://mail.gna.org/listinfo/linux4nano-dev http://www.linux4nano.org
