hpoj-devel — HPOJ development/support list

Re: [hpoj-devel] ptal-mlcd changes 03-Apr-2001

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

Joe Piolunek wrote:
> I added -lcrypto in three different ways, each of which allows the build to 
> complete without errors: 
> 
> Add "-lcrypto" to the LIBSNMP_CMDLINE="-lsnmp" string in configure.in 
> (running autoconf afterward), or to 'CFLAGS' in apps/cmdline/Makefile, or to 
> 'CFLAGS' in ptal/Makefile.
Hi, Joe.  Thanks for the information.  I've put it on my TODO list.

For now the easiest workaround is to run the configure script with the
"--without-snmp" switch.  At the moment SNMP isn't actually used for anything
in libptal.  The intention eventually is to use it in the JetDirect PTAL
provider (ptal-hpjd.c) to retrieve the device ID string and encapsulate PML.
(Speaking of PML, I finally managed to get PML gets and sets working through
PTAL late last night, although there's still some more development and
testing I need to do before it's ready to check in.)

> ptal-mlcd apparently doesn't use libcrypto.
Correct, because it doesn't use SNMP.  Of course, libptal doesn't
intentionally use libcrypto, but in this case it will need to be linked
against it anyway because libsnmp uses it but for some reason doesn't have
the dynamic linkage to it.

David

Re: [hpoj-devel] g85 under suse 7.0 (linux 2.2.16)

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

Manfred Gehrke wrote:
> thanks to all for the suggestions so far. now the g85 works very well with one
> exception:
> 
> i had to start ptal-printd manually. since suse has a different boot concept
> there is no rc.local. /etc/rc.d is linked to /sbin/init.d, where there are a
> lot of start and stop scripts for different run levels. 
> 
> has someone already made these scripts?
> 
> if not, i'll try it, though i'm not very experienced in that stuff.

Hi, Manfred.  Last December or January some init scripts floated around on
the mailing list, although I think they were for RedHat.  There's supposed
to be one in the hpoj-0.7 SRPM I recently posted from Timothy Lee, although
again that SRPM was for RedHat.  Since I don't use SuSE I can't help you
directly here.  My suggestion would be to look under /sbin/init.d and find
an init script that looks fairly simple, copy it to another name, and modify
it to start ptal-printd instead of whatever it was originally written for.

Eventually I do intend to provide a generic init script that can hopefully
be easily integrated into any distribution's init-script system using a
small distribution-specific wrapper script.

David

Re: [hpoj-devel] Problems

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

Micha Kersloot wrote:
> Here is the file. Thanks for your time sofar.
...
> ptal-mlcd: INFO at ExMgr.cpp:2149, dev=<usb:0>, pid=1306, errno=11
> llioService: sent 1022 bytes.
> 1B 25 2D 31 32 33 34 35 58 25 21 50 53 2D 41 64 ...    (.%-12345X%!PS-Ad)
...
> ptal-mlcd: WARNING at ExMgr.cpp:2284, dev=<usb:0>, pid=1306, errno=11
> llioRead(datalen=6,noblock=0): retrying (r0=0).
...
> ptal-mlcd: ERROR at ExMgr.cpp:2045, dev=<usb:0>, pid=1306, errno=11
> llioService: llioRead returns 5, expected=6!
>  
> ptal-mlcd: ERROR at ExMgr.cpp:841, dev=<usb:0>, pid=1306, errno=11
> exClose(reason=0x0010)
...

Hi, Micha.  That "%!PS-Ad" leads me to believe that you're trying to send
a PostScript file directly to the printer.  You need to configure
ghostscript to use a driver such as cdj550 (DeskJet 550C), or better yet,
DJ9xx (DeskJet 9xx from http://hpinkjet.sourceforge.net), to convert the
PostScript print job into PCL that the printer understands.  For some reason,
the rest of the print packets all seem to contain the 5-byte string "!s8W-",
at least according to the first 16 bytes of each packet.  What sort of file
were you trying to print?

The warning message above appeared only once, about halfway through the
file.  It didn't seem to indicate a particular problem, because it retried
the operation and continued on its merry way.  I will probably turn off
most or all of the runtime warnings eventually.

The error messages above do concern me, however.  They indicate that a
partial packet was read from the printer.  Since you didn't mention them
originally, perhaps they were due to some sort of timeout on the printer
due to slowdowns on the computer caused by the extra debug output, although
I've never seen this happen myself.

In any case, please first try setting up PCL printing (see PRINT-HOWTO).  For
now, I think it's safe to ignore the warning message, and you can normally
operate with debug messages turned off.  Of course, please let me know if you
still see "ERROR" messages, such as the ones above or any others, or if you
still have difficulties.

Have you tried setting up scanning yet (see SCAN-HOWTO)?  Any luck there?
You may see the same "llioRead retrying" warning above as well as warnings
about a caught SIGALRM signal.  In my experience I think scanning is the only
time I've ever seen either of these particular warnings.

David

[hpoj-devel] g85 under suse 7.0 (linux 2.2.16)

[<Bar...@t-...>]

hi,

thanks to all for the suggestions so far. now the g85 works very well with one
exception:

i had to start ptal-printd manually. since suse has a different boot concept
there is no rc.local. /etc/rc.d is linked to /sbin/init.d, where there are a
lot of start and stop scripts for different run levels. 

has someone already made these scripts?

if not, i'll try it, though i'm not very experienced in that stuff.

ciao
manfred

Re: [hpoj-devel] Problems

[Micha K. <mi...@ko...>]

Hi David,

Here is the file. Thanks for your time sofar.

-- 
Met vriendelijke groet,

Micha Kersloot
KovoKs Automatiseringspartner

Re: [hpoj-devel] Problems

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

Micha Kersloot wrote:
> As soon as I do that, it keeps spitting out the following message:
> 
> ptal-mlcd: ENTRY at ExMgr.cpp:724, dev=<usb:0>, pid=1167, errno=0
> fdRegister(fd=6,r=0,w=-1,x=0)
> 
> ptal-mlcd: ENTRY at ExMgr.cpp:724, dev=<usb:0>, pid=1167, errno=0
> fdRegister(fd=6,r=0,w=-1,x=0)
> 
> I've got several megabytes filled up now. I don't think that's the right
> thing is it?
My mistake.  Please do another "cvs update" and recompile/install and try it
again.

David

Re: [hpoj-devel] Problems

[Micha K. <mi...@ko...>]

On 05 Apr 2001 01:15:10 -0700, David Paschal wrote:
> 
> - Invoke "killall ptal-mlcd ptal-printd" to make sure old instances are
>   really killed.
> - Invoke "script" to capture the following output to a file.
> - Invoke ptal-mlcd as usual but with the "-nofork" switch added.
> - At the ptal-mlcd debug console prompt, enter the command "log".
> - Enter the command "activate".


As soon as I do that, it keeps spitting out the following message:

ptal-mlcd: ENTRY at ExMgr.cpp:724, dev=<usb:0>, pid=1167, errno=0
fdRegister(fd=6,r=0,w=-1,x=0)

ptal-mlcd: ENTRY at ExMgr.cpp:724, dev=<usb:0>, pid=1167, errno=0
fdRegister(fd=6,r=0,w=-1,x=0)

I've got several megabytes filled up now. I don't think that's the right
thing is it?


-- 
Met vriendelijke groet,

Micha Kersloot
KovoKs Automatiseringspartner

[hpoj-devel] TurboPrint

[Allen B. <ba...@lo...>]

I saw an announcement today for TurboPrint for Linux 1.0,
http://www.turboprint.de/turboprint.html. Has anyone tried their inkjet
drivers for the OfficeJet machines?

Thanks,
Allen

Re: [hpoj-devel] ptal-mlcd changes 03-Apr-2001

[Joe P. <joe...@sn...>]

On Thursday 05 April 2001 03:17 am, David Paschal wrote:
> Joe Piolunek wrote:
> > Updating the ucd-snmp package(s) by installing ucd-snmp-4.1.1-3.i386.rpm
> > or building from ucd-snmp-4.1.1-3.src.rpm causes the link errors to
> > appear in hpoj-0.7 and all of the CVS builds I tried (maybe earlier
> > versions as well).
> >
> > hpoj-0.7 and packages from CVS build with the new ucd-snmp if libcrypto
> > is linked in. The libcrypto I have is from the openssl-0.9.5a-2.6.x RPM.
>
> Hi, Joe.  When you say "linked in", do you mean you changed the Makefile in
> the ptal directory to add "-lcrypto" to CFLAGS?

I added -lcrypto in three different ways, each of which allows the build to 
complete without errors: 

Add "-lcrypto" to the LIBSNMP_CMDLINE="-lsnmp" string in configure.in 
(running autoconf afterward), or to 'CFLAGS' in apps/cmdline/Makefile, or to 
'CFLAGS' in ptal/Makefile . 

>
> Does the new ucd-snmp package indicate a dependency with either libcrypto
> specifically or the openssl package in general (i.e. "rpm -qR ucd-snmp")?

libcrypto specifically. 

]$ rpm -qR ucd-snmp
rpm >= 4.0.2
db3 >= 3.1.17
/sbin/chkconfig
/bin/sh
/bin/sh
rpmlib(PayloadFilesHavePrefix) <= 4.0-1
ld-linux.so.2
libcrypto.so.0
libc.so.6
libdb.so.2
libm.so.6
libnsl.so.1
libsnmp.so.0
libucdagent.so.0
libucdmibs.so.0
libz.so.1
/bin/bash
/bin/sh
/usr/bin/perl
libc.so.6(GLIBC_2.0)
libc.so.6(GLIBC_2.1)
libnsl.so.1(GLIBC_2.0)
rpmlib(CompressedFileNames) <= 3.0.4-1


> What is the output of "ldd /usr/bin/snmpset" with the new ucd-snmp? 
> snmpset is probably actually in the ucd-snmp-utils package.

]$ ldd /usr/bin/snmpset
        libsnmp.so.0 => /usr/lib/libsnmp.so.0 (0x40015000)
        libcrypto.so.0 => /usr/lib/libcrypto.so.0 (0x40056000)
        libdb.so.2 => /lib/libdb.so.2 (0x40131000)
        libz.so.1 => /usr/lib/libz.so.1 (0x4013f000)
        libnsl.so.1 => /lib/libnsl.so.1 (0x4014f000)
        libm.so.6 => /lib/libm.so.6 (0x40165000)
        libc.so.6 => /lib/libc.so.6 (0x40182000)
        /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)

Running 'ldd' on ptal-printd after building with "-lcrypto" added to the var 
LIBSNMP_CMDLINE in configure.in (and running autoconf after):

]$ ldd /usr/local/sbin/ptal-printd
        libptal.so.0 => /usr/local/lib/libptal.so.0 (0x4001f000)
        libc.so.6 => /lib/libc.so.6 (0x40029000)
        libsnmp.so.0 => /usr/lib/libsnmp.so.0 (0x4011e000)
        libcrypto.so.0 => /usr/lib/libcrypto.so.0 (0x4015f000)
        /lib/ld-linux.so.2 => /lib/ld-linux.so.2 (0x40000000)

ptal-mlcd apparently doesn't use libcrypto.

-- 
Joe

Re: [hpoj-devel] Starting PTAL Tools in a Script

[Allen B. <ba...@lo...>]

Hi David,

Sorry. I had a (incredibly egregious) typo in my script [kicks self].

With respect to ptal-mlcd logging to /var/log/messages, I was explicitly
redirecting the output from ptal-mlcd to /var/log/messages. Again,
apologies
for the confusion.

Anyway, to lay out the whole thing: I'd like the usermode drivers to
kick off when the printer is plugged in. As you know, the main problem
with just putting the commands in a startup script is that USB device
numbers can be juggled around. The USB printer module also selects a
minor device number for its character device dynamically and there is
currently no way to find out what this number is (short of grepping
/var/log/messages).

As suggested, and mostly worked out, by David Brownell on the usb-devel
mailing list, I added an ioctl to the USB printer module to extract the
major and minor device numbers of a USB printer through the USB device
filesystem. If anyone is interested, the patch is attached. A short
user-space program can be used to query this data (also attached).

The hotplug architecture (http://linux-hotplug.sourceforge.net/) allows
a per module script to be executed when a device is plugged in. It also
passes the USB device filesystem filename of the newly plugged in device
as $DEVICE and its product string as $PRODUCT. So I put this script in
/etc/hotplug/usb/printer:

#!/bin/bash
# Initialize the PTAL usermode drivers when the MFC is plugged in
if [ $(expr $PRODUCT : 3f0) = 3 ] ; then # Only for HP's (?)

    # Get device node data: creates variables DEVICE{TYPE|MAJOR|MINOR}
    export $(/root/printer/usbdevnode $DEVICE)

    /usr/local/sbin/ptal-mlcd usb:$DEVICEMINOR -device
/dev/usb/lp$DEVICEMINOR >> /var/log/messages

    /usr/local/sbin/ptal-printd mlc:usb:$DEVICEMINOR -like /dev/lp0

    # XXX Should also now enable a print queue for this device

fi

This could still use some help. Enabling print queues depends on which
printing system you use (LPD, CUPS, etc.). There's no support yet for
unplugging devices in hotplug (as near as I can tell, anyway); so, for
now, this is really only works as a one shot when the system boots. And
there
still seems to be some debate about the best way to describe dynamically
attached devices and their representation to user space; so, this
approach is probably short-lived.

Anyway, it works for me now. I'm happy :-)

Thanks,
Allen


David Paschal wrote:
> 
> Allen Barnett wrote:
> > I put the start up of ptal-mlcd and ptal-printd in a script. For some
> > strange reason, if I start them this way, when I try to print I get this
> > message:
> >
> > ptalChannelOpen(chan=0x0804AF00): provider failed open!
> > ptal-printd(mlc:usb:0): ptalChannelOpen failed!  Will delay and retry.
> >
> > and this is logged to /var/log/messages:
> >
> > ptal-mlcd: ERROR at ExMgr.cpp:1961, dev=<usb:0>, pid=1225, errno=2
> > llioOpen: open failed!
> >
> > I tried putting a 1 second sleep between the ptal-mlcd and ptal-printd
> > commands, but it didn't seem to help.
> >
> > It works fine if I type the commands in at the shell prompt, though.
> 
> Hi, Allen.  A few questions:
> - At what point in your bootup sequence are you invoking this script?
> - What are the exact commands you are using to start ptal-mlcd and
>   ptal-printd?  You could just e-mail the whole script.
> - Given that it works if you type the commands at the shell prompt, what
>   happens if you invoke the script at the shell prompt instead of typing
>   the commands manually?
> - When you attempt to print, is the USB "printer" module always loaded?
> 
> I'm also pleasantly surprised that ptal-mlcd messages go to /var/log/messages
> when it's started at boot time.
> 
> David

Re: [hpoj-devel] Starting PTAL Tools in a Script

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

Allen Barnett wrote:
...
> ptal-mlcd: ERROR at ExMgr.cpp:1961, dev=<usb:0>, pid=1225, errno=2
> llioOpen: open failed!

A clarification and another question regarding the "llioOpen" error message:
- errno 2 means "No such file or directory".
- Does /dev/usb/lp0 or whatever actually exist?  I'm sure it does since you
  said you can run ptal-mlcd manually from the command line, but it wouldn't
  hurt to check, since this is a rather strange problem.

David

Re: [hpoj-devel] Problems

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

Micha Kersloot wrote:
> I'm trying to get my Officejet G55 running for some time now. With
> simple(=small text) jobs, it seems working, but if the job grows I get the
> following errors:
> 
> ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=<usb:0>, pid=4481, errno=11
> llioRead(datalen=4096,noblock=0): retrying.
> 
> ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=<usb:0>, pid=4520, errno=11
> llioRead(datalen=6,noblock=0): retrying.
> 
> Then the printer stops. I've got to shut down the printer, kill ptal-mlcd
> and ptal-printd, restart it to get it working again. As I don't know what
> the errors mean, I don't have a clue where to start from here.
...
> I'm running Linux 2.4.3 on a Mac G4 and
> this machine does not contain any parallel ports.

Hi, Micha.  I just checked in a small change to this log message to give a
little more information.  Could you please run "cvs update" in the hpoj
directory, recompile, reinstall, and report back the new messages?

Given that the pid values are different in the two messages above, I assume
that these are from two separate tries, correct?  If that's not correct,
then you may still have an old ptal-mlcd process lying around and messing
things up.  "ps aux|grep ptal" will verify this.

If all you're doing is printing, I'm wouldn't expect that ptal-mlcd would ever
receive packets containing a 4096-byte data payload.  When I saw your first
message with no indication of what kernel you were running, I wondered if
maybe you were running an older kernel (earlier than 2.2.19 or 2.4.2) without
updating printer.c, which previously had issues such as reading bogus packets.
But then you said you were running 2.4.3, which "should" work.  I wonder if
there are any issues in 2.4.x and/or the PowerPC architecture?  Have you
successfully used any other USB devices on your system?  Are you using any
hubs?

Please try the following additional test:

- Invoke "killall ptal-mlcd ptal-printd" to make sure old instances are
  really killed.
- Invoke "script" to capture the following output to a file.
- Invoke ptal-mlcd as usual but with the "-nofork" switch added.
- At the ptal-mlcd debug console prompt, enter the command "log".
- Enter the command "activate".
- After the log messages stop, enter the command "dump".
- In a separate console or terminal window, invoke ptal-printd as usual.
  In this case I don't think it's necessary to add the "-nofork" option.
- Try printing a file that you expect to cause the failure.
- After you get the failure, press <Enter> a few times on the ptal-mlcd
  debug console to verify it's still executing its normal processing loop.
- Enter the command "dump" again.
- Control-C out of ptal-mlcd, and exit out of the subshell spawned by "script".
- E-mail me the file "typescript".  Since it may be rather large, you can
  send it directly to me, to both dav...@hp... and pa...@rc....
  Other than that, I'd like to keep this conversation on the mailing list,
  however.

David

Re: [hpoj-devel] Starting PTAL Tools in a Script

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

Allen Barnett wrote:
> I put the start up of ptal-mlcd and ptal-printd in a script. For some
> strange reason, if I start them this way, when I try to print I get this
> message:
> 
> ptalChannelOpen(chan=0x0804AF00): provider failed open!
> ptal-printd(mlc:usb:0): ptalChannelOpen failed!  Will delay and retry.
> 
> and this is logged to /var/log/messages:
> 
> ptal-mlcd: ERROR at ExMgr.cpp:1961, dev=<usb:0>, pid=1225, errno=2
> llioOpen: open failed!
> 
> I tried putting a 1 second sleep between the ptal-mlcd and ptal-printd
> commands, but it didn't seem to help.
> 
> It works fine if I type the commands in at the shell prompt, though.

Hi, Allen.  A few questions:
- At what point in your bootup sequence are you invoking this script?
- What are the exact commands you are using to start ptal-mlcd and
  ptal-printd?  You could just e-mail the whole script.
- Given that it works if you type the commands at the shell prompt, what
  happens if you invoke the script at the shell prompt instead of typing
  the commands manually?
- When you attempt to print, is the USB "printer" module always loaded?

I'm also pleasantly surprised that ptal-mlcd messages go to /var/log/messages
when it's started at boot time.

David

Re: [hpoj-devel] ptal-mlcd changes 03-Apr-2001

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

Joe Piolunek wrote:
> Updating the ucd-snmp package(s) by installing ucd-snmp-4.1.1-3.i386.rpm or 
> building from ucd-snmp-4.1.1-3.src.rpm causes the link errors to appear in 
> hpoj-0.7 and all of the CVS builds I tried (maybe earlier versions as well).
> 
> hpoj-0.7 and packages from CVS build with the new ucd-snmp if libcrypto is 
> linked in. The libcrypto I have is from the openssl-0.9.5a-2.6.x RPM.
Hi, Joe.  When you say "linked in", do you mean you changed the Makefile in
the ptal directory to add "-lcrypto" to CFLAGS?

Does the new ucd-snmp package indicate a dependency with either libcrypto
specifically or the openssl package in general (i.e. "rpm -qR ucd-snmp")?

What is the output of "ldd /usr/bin/snmpset" with the new ucd-snmp?  snmpset
is probably actually in the ucd-snmp-utils package.

David

[hpoj-devel] Starting PTAL Tools in a Script

[Allen B. <ba...@lo...>]

I put the start up of ptal-mlcd and ptal-printd in a script. For some
strange reason, if I start them this way, when I try to print I get this
message:

ptalChannelOpen(chan=0x0804AF00): provider failed open!
ptal-printd(mlc:usb:0): ptalChannelOpen failed!  Will delay and retry.

and this is logged to /var/log/messages:

ptal-mlcd: ERROR at ExMgr.cpp:1961, dev=<usb:0>, pid=1225, errno=2
llioOpen: open failed!

I tried putting a 1 second sleep between the ptal-mlcd and ptal-printd
commands, but it didn't seem to help.

It works fine if I type the commands in at the shell prompt, though.

Thanks,
Allen

Re: [hpoj-devel] Problems

[Micha K. <mi...@ko...>]

Hi Edwin,

I'm sorry, but that's no option. I'm running Linux 2.4.3 on a Mac G4 and
this machine does not contain any parallel ports.

On 2001.04.04 09:23 Edwin van Opstal wrote:
> Goedemorgen Micha,
> 
> My G55 works fine, but I connected it to the parallel port. Perhaps
> that's an 
> idea?
> 
> Groeten,
> Edwin
> 
> On Wednesday 04 April 2001 00:01, you wrote:
> > Hi there,
> >
> > I'm trying to get my Officejet G55 running for some time now. With
> > simple(=small text) jobs, it seems working, but if the job grows I get
> the
> > following errors:
> >
> > ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=<usb:0>, pid=4481, errno=11
> > llioRead(datalen=4096,noblock=0): retrying.
> >
> > ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=<usb:0>, pid=4520, errno=11
> > llioRead(datalen=6,noblock=0): retrying.
> >
> >
> > Then the printer stops. I've got to shut down the printer, kill
> ptal-mlcd
> > and ptal-printd, restart it to get it working again. As I don't know
> what
> > the errors mean, I don't have a clue where to start from here.
> 
> -- 
> 
> Edwin van Opstal
> S.E. Technology
> 
> _______________________________________________
> hpoj-devel mailing list
> hpo...@li...
> http://lists.sourceforge.net/lists/listinfo/hpoj-devel
> 
> 
-- 
Met vriendelijke groet,

Micha Kersloot
KovoKs Automatiseringspartner

Re: [hpoj-devel] Problems

[Edwin v. O. <edw...@se...>]

Goedemorgen Micha,

My G55 works fine, but I connected it to the parallel port. Perhaps that'=
s an=20
idea?

Groeten,
Edwin

On Wednesday 04 April 2001 00:01, you wrote:
> Hi there,
>
> I'm trying to get my Officejet G55 running for some time now. With
> simple(=3Dsmall text) jobs, it seems working, but if the job grows I ge=
t the
> following errors:
>
> ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=3D<usb:0>, pid=3D4481, errno=3D=
11
> llioRead(datalen=3D4096,noblock=3D0): retrying.
>
> ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=3D<usb:0>, pid=3D4520, errno=3D=
11
> llioRead(datalen=3D6,noblock=3D0): retrying.
>
>
> Then the printer stops. I've got to shut down the printer, kill ptal-ml=
cd
> and ptal-printd, restart it to get it working again. As I don't know wh=
at
> the errors mean, I don't have a clue where to start from here.

--=20

Edwin van Opstal
S.E. Technology

Re: [hpoj-devel] ptal-mlcd changes 03-Apr-2001

[Joe P. <joe...@sn...>]

David:

Updating the ucd-snmp package(s) by installing ucd-snmp-4.1.1-3.i386.rpm or 
building from ucd-snmp-4.1.1-3.src.rpm causes the link errors to appear in 
hpoj-0.7 and all of the CVS builds I tried (maybe earlier versions as well).

hpoj-0.7 and packages from CVS build with the new ucd-snmp if libcrypto is 
linked in. The libcrypto I have is from the openssl-0.9.5a-2.6.x RPM.

-- 
Joe

[hpoj-devel] Problems

[Micha K. <mi...@ko...>]

Hi there,

I'm trying to get my Officejet G55 running for some time now. With
simple(=small text) jobs, it seems working, but if the job grows I get the
following errors:

ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=<usb:0>, pid=4481, errno=11
llioRead(datalen=4096,noblock=0): retrying.

ptal-mlcd: WARNING at ExMgr.cpp:2281, dev=<usb:0>, pid=4520, errno=11
llioRead(datalen=6,noblock=0): retrying.


Then the printer stops. I've got to shut down the printer, kill ptal-mlcd
and ptal-printd, restart it to get it working again. As I don't know what
the errors mean, I don't have a clue where to start from here.

-- 
Met vriendelijke groet,

Micha Kersloot
KovoKs Automatiseringspartner

Re: [hpoj-devel] ptal-mlcd changes 03-Apr-2001

[Joe P. <joe...@sn...>]

On Tuesday 03 April 2001 07:19 am, David Paschal wrote:
> Hi,
>
> I just checked in some more changes for ptal-mlcd.  As always, it would be
> helpful if people did a "cvs update" in their sandboxes to pull down the
> latest code and help make sure I didn't break anything.  :-)

It still works, David.

I did find something that you can check into, though. After updating the 
ucd-snmp* rpms from 4.0.1-4 to 4.1.1-3 on my RH 6.1 system, link errors 
occurred when attempting to build the latest hpoj package from CVS:

make[1]: Entering directory 
`/home/joe/hpoj_proj/CVS/hpoj_CVS_04-03-2001/apps/cmdline'
gcc -O -Wall -g  -DHAVE_SNMP 
-I/home/joe/hpoj_proj/CVS/hpoj_CVS_04-03-2001/include -I/usr/include/ucd-snmp 
-I/usr/local/qt/include -L/home/joe/hpoj_proj/CVS/hpoj_CVS_04-03-2001/ptal 
-L/usr/local/qt/lib -lptal -o ptal-connect ptal-connect.c
/usr/bin/../lib/libsnmp.so.0: undefined reference to `EVP_DigestInit'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `EVP_DigestFinal'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `EVP_md5'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `HMAC'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `EVP_sha1'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `des_cbc_encrypt'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `des_key_sched'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `EVP_DigestUpdate'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `RAND_bytes'
/usr/bin/../lib/libsnmp.so.0: undefined reference to `des_ncbc_encrypt'
collect2: ld returned 1 exit status
make[1]: *** [ptal-connect] Error 1
make[1]: Leaving directory 
`/home/joe/hpoj_proj/CVS/hpoj_CVS_04-03-2001/apps/cmdline'
make: *** [all] Error 2

The latest CVS does build with the older ucd-snmp* rpms from redhat. The 
problem seems to be a difference in the 'ucd-snmp' rpm rather than 
'ucd-snmp-devel' or 'ucd-snmp-utils'.

Linking in libcrypto apparently fixes the problem. A reinstall + restart 
shows that printing still works.

-- 
Joe

[hpoj-devel] ptal-mlcd changes 03-Apr-2001

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

Hi,

I just checked in some more changes for ptal-mlcd.  As always, it would be
helpful if people did a "cvs update" in their sandboxes to pull down the
latest code and help make sure I didn't break anything.  :-)

Some of the more significant changes are:

- I partially fixed the "scanimage --test" problem that Daniel reported.
I haven't completely fixed it yet, because it has turned out to be harder
to fix reliably than I originally thought.  I know it worked with the old
kernel-mode drivers, but I suspect it's a difference in timing or buffer
counts, which is what's making it so difficult.  Hang in there...  :-)

- Fixed a bug where resources weren't getting reclaimed after failure to
open a channel on the peripheral.

- Added an "-activate" switch so ptal-mlcd will attempt to start
communicating with the peripheral at startup.  Of course, if it fails
(i.e. if the peripheral is powered off), then it doesn't mind, unlike the
old ieee12844.o kernel module.

- Added support for passing a "-device /dev/lpX" switch when using a
peripheral connected to a parallel port.  ptal-mlcd keeps it open while
it's activated with the peripheral to force the kernel modules to load and
get their problematic initialization over with.  It also prevents someone
from opening and writing to /dev/lpX later and breaking communication.
Joe, this is what we talked about recently.

I think I will take the "-device /dev/lpX" change one step further by
adding periodic sanity checking on the parallel port control registers,
to detect the case where the "-device" switch was omitted or incorrectly
specified and the parallel port was interfered with.

I made a note of most of these things on the "Bugs and TODO" page.
I'll update the documentation as appropriate later, for both this and to
give information on the new hpinkjet print drivers in PRINT-HOWTO.

David

Re: [hpoj-devel] Double-sided printing

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

Bill Wohler wrote:
>   I just received the double-sided printing attachment for my G85. The
>   instructions indicate that you have to click the Double-sided option
>   in the Windows software.
...
>   I just figured everything would come out double-sided by default
>   once I added the attachment. If this is not the case, I'd appreciate
>   any tips on enabling it. 

Hi, Bill.  Here's what one of the hpinkjet developers had to say when I
asked if there was a quick-and-dirty way to use the duplexor:

> I looked into this so I could understand what the issues might be,
> but I am open for more discussion on this subject.
>
> Duplexing is on our current list to add to the APDK. I just don't know
> when it will be done.
>
> Ok, what can be done now? There is some code in broadway.cpp (DJ9xx
> driver) that does enable quick-and-dirty duplexing. The code is
> commented out, and I don't know how well it works. If it did work it
> would only work for tablet mode duplexing.
>
> Book mode duplexing would require alternate pages to be reversed
> (rotated 180 degrees) so the bottom of the page prints first.  Also,
> the top and bottom margins would have to be reversed. I am  not aware
> if gs would handle any of this work for us or the driver  would have
> to do this. 

It sounds like it's not trivial, but it might be doable if you or anybody
else has a burning desire to do a little hacking to make it work.  :-)

David

Re: [hpoj-devel] hpinkjet testing results

[Joe P. <joe...@sn...>]

On Sunday 01 April 2001 07:52 am, David Paschal wrote:

<....>
> When you get a chance,
> would you try out this patch (to Header.cpp in the hpijs directory) and
> verify that you can now print through ptal-printd?

It does work now.

I originally tested the printing system with the drivers installed from the 
0.93 rpms, but have since updated the ghostscript-hpijs rpm to ver. 0.94, 
built from the SRPM. I now have hpijs installed from the (patched) tarball 
instead of the hpijs rpm.

'rpm' complains about an unsatisfied dependency due to the hpijs rpm not 
being installed, but since hpijs actually is installed, there is no real 
problem. To avoid some future confusion, I'll reinstall the hpijs rpm if it 
becomes available in a patched version.
 
-- 
Joe

Re: [hpoj-devel] Double-sided printing

[Bill W. <wo...@ne...>]

  Thanks for the feedback guys.

Judd Montgomery <ju...@jp...> writes:
> I'm using gs 5.50a.  The cdj970 is an add on that slackware included. 
> Sorry about that.  The filename is called gdevdj9.c.  I don't know where
> slackware got it from.  I would be happy to send you that file, or a
> compiled gs.

  Could you send me both please? That might tide me over until the
  hpinkjet folks provide for duplex. I'll look at the links David sent
  in a moment...

> #include <time.h>, it was probably included in some other header on
> someone elses system, but not yours.

  Sounds reasonable.

> Ah, so you got it.  Good.  So, the photos look better?

  Yes! But I do miss the razor-sharp laser printer text quality.

-- 
Bill Wohler <wo...@ne...>  http://www.newt.com/wohler/  GnuPG ID:610BD9AD
Maintainer of comp.mail.mh FAQ and mh-e. Vote Libertarian!
If you're passed on the right, you're in the wrong lane.

Re: [hpoj-devel] Double-sided printing

[Judd M. <ju...@jp...>]

Bill Wohler wrote:
> 
> Judd Montgomery <ju...@jp...> writes:
> > I'm using the cdj970 ghostscript driver and use the -dDuplex=2 option.
> 
>   Where is this cdj970 driver you speak of? I had my head in gs 5.10,
>   5.50 and 6.50 today (long day). Didn't see it anywhere.
> 
I'm using gs 5.50a.  The cdj970 is an add on that slackware included. 
Sorry about that.  The filename is called gdevdj9.c.  I don't know where
slackware got it from.  I would be happy to send you that file, or a
compiled gs.

>   Meanwhile, I tried to compile the new inkjet drivers (DJ9XX). I say
>   compile since I'm running a Debian system and while I was able to
>   create .debs from the RPMs with alien, the package name was
>   different and thus didn't replace gs nor its dependencies. The
>   install therefore failed because it wanted to overwrite existing
>   files which dpkg won't let you do.
> 
>   So, I downloaded all the sources on the sources in
>   ftp://mirror.cs.wisc.edu/pub/mirrors/ghost/gnu/gs550/ and
>   http://prdownloads.sourceforge.net/hpinkjet/hpijs0.94.tar.gz and got
>   this while compiling gs5.50:
> 
>     [wohler@gbr:803]$ make -f unix-gcc.mak prefix=/usr/local
>     gcc `cat ./obj/cc.tr`  -O2 -Wall -Wcast-align -Wstrict-prototypes
>         -Wwrite-strings -fno-common  -I. -I./obj -I. -I./obj  -o
>         ./obj/zdevcal.o -c ./zdevcal.c
>     ./zdevcal.c: In function `calendar_get_params':
>     ./zdevcal.c:53: storage size of `ltime' isn't known
>     ./zdevcal.c:55: dereferencing pointer to incomplete type
>     ./zdevcal.c:55: initializer element is not constant
>     ./zdevcal.c:55: (near initialization for `items[0].offset')
>     ...
> 

That looks like someone didn't include <time.h>, easy fix.

>   I probably won't pursue that path, but am curious why struct tm
>   didn't get defined. After a few years of programming in Java, I have
>   happily forgotten how to follow the spaghetti in /usr/include.
> 

#include <time.h>, it was probably included in some other header on
someone elses system, but not yours.

>   I finally broke down and ran:
> 
>     [root@gbr:942]# dpkg --install --force-overwrite ghostscript-hpijs_5.50-2_i386.deb
> 
>   gs --help shows the DJ9xx driver and it appears to work although I
>   haven't noticed any difference in print quality for text. Ah, just
>   printed a photo--there's the difference! (Possible to get my old 300
>   dpi laser printer quality on the text on this 600 dpi OJ G85
>   printer?)
> 

Ah, so you got it.  Good.  So, the photos look better?

>   Another strange thing is that lprngtool doesn't show the new
>   drivers... Any thoughts on why not?
> 

No, I just edit files by hand.

Judd