Re: Don: Compare Directories and Files
- Posted by Pete Lomax <petelomax at blueyonder.?o.uk> Jul 10, 2007
- 596 views
don cole wrote: > diff.zip by R. M. Forno. My main goal is to compare > different versions of Arwen to to see haw to place color tabs. > > Forno's diff works fine on small test programs but when I tried to compare > 2 versions of Arwen it took too long. So I left it running all night. > Next day it was still running. Maybe what I am tring to do is an imposible > task. ... >Hello Alan, > > http://palacebuilders.pwp.blueyonder.co.uk/edita.htm#download > > I'm only interested in Arwen at this time. Compare it with any old/ > download 'Arwen win32 library' by Mike. Don, what is the problem? It /is/ a tough compare, and while I am sure a windows port of the unix command-line diff runs much quicker, even on this crappy box, Edita solves this in 17 mins, at 2006 differences [see PPPS] and a complexity of ~5.8 million (meaning that in those 17 mins it logged 5.8 million decision points before finding the optimal answer) but it got there. Is there something wrong with the Edita file compare output? Is there something wrong with coloured tabs in Edita? If you failed to understand the source mods I posted what makes you think you will understand a source diff anyway? Is there something in the new Arwen you need in the one I ship with Edita? Confused, Pete PS to Alan Oxley: I would be interested in your results but I have to warn you I spent much time investigating source comparision and would be (pleasantly) surprised to see my code beat, on the same kit that is. PPS to Al Getz: I fell off my chair laughing at your "Comparing files really isnt hard at all." - obviously "are these files binary identical?" is what you probably meant, but "shortest edit difference" has kept some very bright acedemics awake at night long time now and still will for some time yet. PPPS If you can prove it is even one less than 2006, please let me know. PPPPS I just want to say that 17 mins is directly related to the 2006 differences; were there only 100 it should get it done in under a second.