[hpoj-devel] Re: checkpc

[<pa...@rc...>]

Alexander Zimmermann wrote:
> This is new in 7.0. From the man-page:
...
> I'm also not familiar with it, but it is called from
>   /etc/init.d/lpd start
> with the  -f  parameter and sometimes blocks the script and therewith
> the whole boot process. I don't know when it's actually blocking, maybe
> if ptal-printd is not running, or not ready. Therefore I moved the
> lpd start  to the very "last" link in  /etc/rc?.d/S*lpd  .
> 
> At least it performs the following actions on the printer device (from
> strace output)
...

Hi, Alexander.  Thanks for the info.  I ran some tests on the pipe node
(such as /dev/ptal-printd/mlc_usb_0).  If the node is stale (i.e.
ptal-printd isn't running), then "cat >/dev/ptal-printd/mlc_usb_0) seems
to block on the open, because it doesn't exit with ^D, which normally
signals end of file.  However, when I start ptal-printd, ^D works.

I wonder what checkpc does if /dev/ptal-printd/mlc_whatever doesn't exist
at all (i.e. if ptal-printd deleted the pipe node before exiting).  Would
it delete the corresponding queue entry in /etc/printcap?

I guess this would explain why the permissions on the old wrapfilter script
were getting messed up on RedHat 7.0 every time lpd got restarted, which was
one reason why I decided to shove ptal-printd into 0.7 at the last minute.

David