[sane-devel] possible sane problem, not sure

[Gene Heskett]

On Sunday 20 October 2002 18:37, Karl Heinz Kremer wrote:
>On Sun, Oct 20, 2002 at 05:04:36PM -0400, Gene Heskett wrote:
>[ ... ]
>
>> >Unfortunately nothing in your original email suggested that you
>> > were using the IScan software.
>>
>> No, as I stated, the iscan-1.0.4 build failed on the pisa_*
>> stuff in the frontends directory.  The previous version 1.0.3,
>> also failed at the same place, so I've always used the xsane
>> frontend.  Below is my /usr/local/etc/sane.d/dll.conf:
>
>[ ... ]
>
>> showing only that 2 backends are enabled.  Am I mistaken in
>> assumeing that the GT-7200 is coming from the iscan libraries?
>
>The support for the GT-7200 comes either from the Plustek backend,
>or from the modified EPSON backend that is provided by IScan. When
> you installed only those parts of IScan that compiled, you did
> replace the original EPSON backend from the Sane distribution.
> Support for the Perfection 1250 and 1260 in the IScan version of
> the EPSON backend is what got EPSON KOWA in trouble with the GPL.
> They provide a binary library that adds support for these two
> scanners that is linked into the backend.
>
>> I did not have that option before building and installing the
>> iscan parts that *would* compile from the iscan-1.3.0.tar.gz
>> back when it was first announced on this list as being
>> available.  And no replaced with supposedly identical parts from
>> 1.4.0, however it now has a pretty obvious cyan cast to the
>> high-lights of an image.
>>
>> I've sent the epson-kowa pages a message about the compile
>> problems, but no reply since this is the weekend.
>
>Is it a compile error or a link error?

It looks like a compile error to me, of the sort one would get from 
using an obsolete, older header file thats missing some prototypes.  
I'd post the whole error train it outputs, but its probably 2 
megabytes from the first mention of src/frontands/pisa_whatever 
problem to the final error 1 exit.  I won't inflict this list with 
that, but I could snip the first and last instances if someone 
wants to look at it that closely.  It also reports several 
megabytes of a repeat of the contents of backwards.h, once for 
virtually every sourcefile it makes.  It also did this with 1.3.0 
and gcc-2.96-112 although the message wasn't quite so pointed.

> The package also contains
> another binary library that provides the image manipulation
> routines from EPSON's Mac and Windows drivers. This is a C++
> library and therefore requires you to use the same compiler as
> was used to create this library. That is, up until GCC 3.2. From
> here on, it should no longer matter (once EPSON KOWA delivers a
> library that was compiled with GCC 3.2).
>
>> FWIW, the *only* active line in epson.conf is:
>>
>> usb /dev/usb/scanner0
>>
>> Everything else is commented out.  So _I_ think I'm running the
>> epson iscan libraries from xsane when I choose the GT-7200
>> option from the xscan startup menu.
>
>The backend provided with IScan is a regular Sane backend, it can
>therefore work with all Sane frontends.
>
>Have you tried the RPM package of IScan? This should get you
>around the compile/link problem.

The Iscan rpms from epson require that sane and xsane also be 
installed from rpms, and I don't have anything that new on hand, 
even the ones with rh8.0 are no newer than 1.0.7 and xsane-0.84 
IIRC.

-- 
Cheers Karl, Gene
AMD K6-III at 500mhz 320M
Athlon1600XP at 1400mhz  512M
99.18% setiathome rank, not too shabby for a WV hillbilly