Re: HP C5100A Photscanner support

Rex Vaughn (rvaughn@nova.org)
Mon, 09 Feb 1998 04:08:52 +0000

Pkirchg wrote:
>
> Hi,
>
> I just got the information that the HP C5100 Photosmart Photoscanner
> does really support HP-SCL. So I did some tests on Linux kernel 2.0.21
> with the new sane 0.69 and it seems to work. With these tests I got
> the following:
>
> Scanning negatives: supported from 100 to 2400 dpi
> Problems:
> - When scanning strips with 5 negatives of size 24x36mm,
> the 5th image shows wrong colours (scanner firmware ?)

I have noticed a problem while previewing a negative, however when the
actual
scan is done the last image seems fine. I started to dig in to the code
to
find out what was going on, and while scanning all but the last
"stripe" (block of 8192) of the negative, everything seems fine.
However,
when the last block is scanned in preview mode, the buffer returned from
sane_read() in input_available() (frontend/preview.c) does not change
from the
second to last block. Nor does sane_read return SANE_STATUS_EOF. After
six or so calls (each returning the same block slightly offset from the
previous "stripe") sane_read finally returns SANE_STATUS_EOF (with still
the same block in buf). (BTW I am assuming each block is slitly offset
from
each other because a block of 8192 bytes is not evenly divisible by the
width
of the negative)

Any ideas on where to attack the problem next?

I am using sane-0.70, gimp-0.99.18, gtk+-0.99.3 an aha152x scsi card,
which
came with an Iomega Bernoulli 150. (IE the scsi card does not have an
id
code that matches the scsi driver in the kernel, but it does work when
given
parameters on bootup.) and the linux kernel is 2.0.31.

-- 
rvaughn@nova.org (RBV)

--
Source code, list archive, and docs: http://www.mostang.com/sane/
To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com