[sane-devel] patch for sane-desc.c / support for descfiles / modifications

[m. allan noah]

On Mon, Jan 12, 2009 at 9:37 AM, Julien BLACHE <jb at jblache.org> wrote:
> Dieter Jurzitza <dieter.jurzitza at t-online.de> wrote:
>
> Hi,
>
>> +  printf ("\nLABEL=\"libsane_usb_rules_end\"\n");
>> +  printf ("\nSUBSYSTEMS!=\"scsi\", GOTO=\"libsane_scsi_rules_end\"\n");
>> +  printf ("LABEL=\"libsane_scsi_rules_begin\"\n");
>> +  printf ("# anything with device-id \"6\" in the SCSI-system is a scanner.\n");
>> +  printf ("KERNEL==\"sg[0-9]*\", NAME=\"%%k\", ATTRS{type}==\"6\", MODE=\"%s\", GROUP=\"%s\"\n", DEVMODE, DEVGROUP);
>> +  printf ("# anything from HP with device-id \"3\" is treated accordingly (hal needs to make sure, finally).\n");
>> +  printf ("KERNEL==\"sg[0-9]*\", NAME=\"%%k\", ATTRS{vendor}==\"HP\", ATTRS{type}==\"3\", MODE=\"%s\", GROUP=\"%s\"\n", DEVMODE, DEVGROUP);
>> +  printf ("LABEL=\"libsane_scsi_rules_end\"\n");
>> +  printf ("LABEL=\"libsane_rules_end\"\n");
>
> Err, that sounds a little bit too generic. Could you generate rules to
> match the SCSI scanners we know about, and only those, like is done
> for USB scanners?
>
> There's no way generic rules like that are going in, because it's a
> given that they'll end up breaking something on users' systems down
> the road.

Julien- I think I have to side with Dieter at least on the device type
6 being always allowed. We don't know the exact vendor or model
strings for any scsi device in the .desc files. Many times, the
backends don't even know. I'm indifferent toward the 'processor'
models, since there are so few and it should be possible to get a list
of those.

allan
-- 
"The truth is an offense, but not a sin"