Here's some excerpts from a conversation about making SANE work on 
my olde worlde PPC.


>> > Setting the correct data length is another problem: sane-find-scanner
>> > uses a different length depending on wether the command line option "-v"
>> > is set or not set -- I suspect a compiler bug; removing "-O2" from
>> > CFLAGS helped.
> 
>> 
>> What compiler makes trouble with "-O2"?
> 
> 
> [...-backends/work/sane-backends-1.0.7/sanei]# gcc -v
> Using builtin specs.
> gcc version egcs-2.91.66 19990314 (egcs-1.1.2 release)> (PPC evrsion)
> 
> It is really weird error. The following statements from
> scanner_do_inquiry in sane-find-scanner.c do not work properly.
> 
>   size = 36; /* patched; was 5 */
>   set_inquiry_return_size (inquiry.cmd, size);
>   sanei_scsi_cmd (sfd, inquiry.cmd, inquiry.size, buffer, &size);
> 
> Some fprintf(stderr,...) statements in sanei_scsi_cmd2 show that byte 4
> of the cdb and the variable "size" have the _same_, but wrong, value,
> sometimes 5, sometimes 8, if the program is compiled with -O2. My
> conclusion is that "size = 36;" does not work properly -- for whichever
> reason... This may sound weird, but I don't see how else inquiry.cmd[4]
> and size could have the same wrong value. 



-- 
Paul Beard
8040 27th Ave NE
Seattle WA 98115
206 529 8400

Murphy's Law is recursive.  Washing your car to make it rain 
doesn't work.