[cups-devel] [UNKN] STR #4620: Superflous printer errors in appsocket 9100 and ipp 631, works in lpr/lpd 515

[Chris Severance]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

DO NOT REPLY TO THIS MESSAGE.  INSTEAD, POST ANY RESPONSES TO THE LINK BELOW.

[STR New]

Arch Linux: 3.19.3-3-ARCH x86_64
cups 2.0.2-4, cups-filters 1.0.68-1, cups-pdf 2.6.1-2, libcups 2.0.2-4
hplip installed then removed
foomatic installed then removed
gutenprint installed then removed
Printers: HP LaserJet Pro 400 m401n, Lexmark T640, HP P2035 with JetDirect
170x, Okidata Oki B411dn

echo -e "\f" > x; lp -dPrinter x; rm x

This works especially good with my Oki B411 because it is programmed to not
print blank pages.

If I set up a printer with appsocket or ipp I see in the cups web
interface:
Idle - "Waiting for printer to finish." (appsocket)
Idle - "Waiting for job to complete." (ipp)

and netstat -an reports a connection to the printer which takes a minute or
so to disappear. 

If I tell cups to disconnect with "socket://0.0.0.0:9100/?waiteof=false", I
get:
Idle - "Connected to printer."

Except that cups is lying. If I'm fast I can see a connection in netstat
- -an for about 10 seconds.

If I use lpr I get:
Idle - "Data file sent successfully."

That sounds like what I'm looking for. netstat -an shows a connection for
about 10 seconds. I'd rather not use lpr for every printer. It's harder to
set up and only some printer manufacturers are forthcoming with their queue
names and function and the big IBM list doesn't have all brands. I don't
like strange printing problems because of strange queue names. 

Some have suggested that the printers are failing to close the appsocket.
Maybe they are but all brands are doing it and they can't all be broken. I
think it's a cups problem and there are two problems.

* cups is not updating the web display correctly. Each connection type is
pulling from a grab bag of random messages that hang around and scare
people into thinking that their test servers will break in production when
multiple servers start hitting the printers.

* appsocket is not expiring sockets in the same timely fashion that other
connections are.

Link: https://www.cups.org/str.php?L4620
Version: 2.0.2
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.14 (GNU/Linux)
Comment: GPGTools - https://gpgtools.org
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=hKQO
-----END PGP SIGNATURE-----