Post Two - 22 May 2007

I have written some more tools, but have not got much further with the decoding. Tonight's example was to create a file with a single text box, called C:\a.pub. (The filename is relevant, as it is stored as part of the file). I then moved the text box down by 1mm. The first file has the text box at 5.4cm down, 6.2 right; the second file (C:\b.pub) is 5.5cm down.
Yes, that's the only difference.
The changes in the files, once I've allowed for trivia, is 787 changes.
With these files, I've done a bit of engineering:
ab.txt - my notes on the diffs
phtml.c - the code which creates the diffs, and blanks out the stuff which needs to be blanked. It also splits its output into multiple files, because web browsers aren't terribly good at dealing with huge tables, which is what we are creating here. Maybe CSV would be a better format.
outfile1.html - The output - file 1
outfile2, outfile3, have no diffs at all.
outfile4.html - The output - file 4 - this has no real diffs, other than the 127/255 (losing the high bit) for some reason
outfile5.html - The output - file 5 - Strange temp filenames occur in one file, not the other.
outfile6.html - The output - file 6
outfile7.html - The output - file 7