[sane-devel] HAL and scanners.

[m. allan noah]

On Wed, Mar 19, 2008 at 9:38 AM, Étienne Bersac <bersace03 at gmail.com> wrote:
> Hi,
>
>  Very interesting, thanks goes to SUSE for working for long time on
>  scanner integration with udev and HAL.
>
>  The HAL callout goal would be simply to compute the SANE device name put
>  it to HAL device property scanner.sane.device_name . This does not need
>  any SANE call as long as all backends supports some well known naming
>  scheme along their own specific naming scheme.
>
>  Then, for scanner specific events (paper-in, adf-opened, "mailto" button
>  pushed, etc.), a HAL addon would fit, but then linking to SANE and
>  requiring 'scanner' privileges.
>
>  That's how it see it for now. Please comments.
>

so, just to be clear, are we suggesting:

1. a small callout, distributed with hal, which replicates the
sanei_{usb,scsi} device naming logic, and saves this device_name
string back into the hal data structure

2. a bigger callout, distributed with ??, which links to the dll
backend, and takes that device_name string, and does sane_open on it,
then does button monitoring, etc

3. modifying all existing sane backends to accept default device names

can hal run two callouts, one of which does not exit?

i have two small concerns about this plan: replication of naming code
makes that convention a de facto API, and what about external
backends?

other issues like odd-ball interface's device names and making sure
all backends can handle the default names are not that scary, since
there are few of those.

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