[cups.general] http://cups.org/str.php?L2535

[Michael R Sweet]

Tim Waugh wrote:
> This bug:
>   http://cups.org/str.php?L2535
> is marked as closed but I would like to submit a patch.
> 
> This comment is incorrect:
>   "The ioctl() is supposed to work in current kernels"
> 
> There is no support in the lp driver for fetching the device ID, and has
> not been for a while.

Which IMHO is a bug.  The USB driver supports it, and all of the
commercial UNIX's support it (with variations), so why not the
parallel port drivers?

> This comment is puzzling:
>   "better to have the OS do its job"
> 
> It *was* and *is* doing its job, but CUPS no longer fetches the Device
> ID from the correct place (/sys).
 >
> I'm not quite clear on the reason for the working code being removed,
> but here is a patch to make it work again in a different way (in case
> the '/sys' bit is the objection?).

The problem we had with getting the device ID from /proc is that
it was usually wrong.  That file only reflects the device ID at
boot time, making it useless and generally inaccurate.  We had a
long string of bugs reported from customers on this in the past, and
concluded that using the autoprobe file was the wrong way to go.

Your patch looks like it should support "live" reading of the device
ID, which is really what we need.  If you can file a bug report with
the patch attached, I'll be happy to clean up the patch and get it
into 1.3 and 1.4.

Thanks!

-- 
______________________________________________________________________
Michael R Sweet                        Senior Printing System Engineer