[sane-devel] Technical infos about reverse engineering

m. allan noah anoah@pfeiffer.edu
Fri, 13 Feb 2004 13:37:09 -0500 (EST)


i recommend that you use the sniffusb.exe from 
http://benoit.papillault.free.fr/usbsnoop/doc.en.php

i like it's output better than the other snoopers that are available.

i usually start the sniffer, plug in the device, count to 30, and then 
unplug. that gives you a good feel for the initialization sequence, and 
the sort of packets the scanner gets every few seconds to check it's 
status.

then it is a matter of making very small steps, like scanning small image 
in b&w, then change the x dimension and scan again it again, then change 
the y and scan again, then change the paper size and scan again. then try 
a different resolution, then grayscale, then color...

after you go thru every option in the windows twain software, (if you kept 
a list of what order you did things in) then you can spend some time 
looking at the packets, and see what is the same from scan to scan, and 
what changes...

allan

On Fri, 13 Feb 2004, Julien TIERNY wrote:

> Hi dear all,
> 
> 	I began to gather information in order to write a backend for the HP Scanjet 
> 2400 scanner. I contacted the concerned people over the mailing list and I 
> joined their effort.
> 
> 	I just wanted to have a piece of advise about reverse engineering. 
> 
> 	Particularly, how can I extract actions (I mean commands) from usbsniffs 
> performed under win32 ?
> 
> 	I gave a try to the 'usb-robot' utilities, but the configure step broke 
> anormally (can't find libusb headers while libsub is totally functionnal on 
> my system). Did anyone manage to have interesting results with this app ?
> 
> 	At last, does anyone know a tip to get a scanner's chipset's manufacturor 
> id ?
> 
> 	Thank you very much, julien
> 
> 
> 

-- 
"so don't tell us it can't be done, putting down what you don't know.
money isn't our god, integrity will free our souls" - Max Cavalera