Alessandro Zummo <azummo-lists at towertech.it> writes:
> On Wed, 14 Jan 2009 19:33:54 -0600
> Chris Bagwell <chris at cnpbagwell.com> wrote:
>
>> But google turned up a PDF of some other random Epson scanner that shows
>> the length of identity2 response. The patch at the end of this email
>> specifies the response length and also fixes the issue. As I've spent a
>> grand total of 1 hour on researching this so far, I do not know enough
>> yet to decide if the length is the real fix (and safe for all scanners)
>> or disabling the call like you did is right (esci_request_identity()
>> also uses a zero length but is wrapped by a != SANE_EPSON_NET to prevent
>> similar error).
>
> Hi,
>
> iirc the length can be different for each scanner model, but I might
> be wrong. I've been a little busy but will try to research the issue.
Correct. The identity2 command (ESC i) uses the same handshake as the
identity command (ESC I).
- send 2 byte command
- read 4 byte info block
- compute size of upcoming data block from info block
(n = info[2] | info[3] << 8)
- read n byte data block
The size of the data block is model specific.
There are more commands that use this handshake, e.g. ESC f.
Hope this helps,
--
Olaf Meeuwissen, LPIC-2 FLOSS Engineer -- AVASYS Corporation
FSF Associate Member #1962 Help support software freedom
http://www.fsf.org/jf?referrer=1962
