Print errors and hangs using HPIJS with HP DJ4160

[Suffield, David]

I also link it with libusb 0.1.12 and other things.

If you are using hpijs only, no printer status or bi-di communication.
Then with configure option --enable-hpijs-only-build, you shouldn't need
libusb.

- sometimes (never the first time after power up) the Job is
created but the construction_error = SYSTEM_ERROR. It seems
that the printer is not responding and HPIJS times out.
Retrying works and printing can continue.

Not sure what this error is. If you use hpijs-only-build, then you
shouldn't get this error.

- sometimes HPIJS hangs on what seems like the last pass
(bottom) of the raster. Sometimes it does not hang and
NewPage() ejects the page but the bottom of the raster is
incomplete (unfinished part of a line)

This error sounds like you are closing usb io down before all the data
has be received by the printer.

-dave

-----Original Message-----
Of Martin Kajdas
Sent: Tuesday, August 28, 2007 10:15 AM
Subject: [HPLIP-Devel] Print errors and hangs using HPIJS
with HP DJ4160
Hello,
I am developing an embedded Linux application that uses HP
DJ4160 printer. (no desktop, no HPLIP, no CUPS, no lp, no
spooling) I am using HPIJS code from HPLIP 2.7.7 as is, which
I compile and create a static library and link with my application.
I also link it with libusb 0.1.12 and other things.
Basically, the printing works, I create a job, send some
rasters and close job.
- sometimes (never the first time after power up) the Job is
created but the construction_error = SYSTEM_ERROR. It seems
that the printer is not responding and HPIJS times out.
Retrying works and printing can continue.
- sometimes HPIJS hangs on what seems like the last pass
(bottom) of the raster. Sometimes it does not hang and
NewPage() ejects the page but the bottom of the raster is
incomplete (unfinished part of a line)
Does anyone has any idea what to try, what to report
(additional info?), how to fix it or have seen it before?
I think this is some kind of HPIJS driver problem that needs
to be resolved.
I used to have other problems (i.e. blinking panel lights)
that were fixed in this version however I was able to cause
it once with this version also but I cannot easily reproduce it.
Above problems occur regularily and I can provide debugging
info if requested.
Thank you,
Martin
--------------------------------------------------------------
-----------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and
a browser.
Download your FREE copy of Splunk now >>
http://get.splunk.com/ _______________________________________________
HPLIP-Devel mailing list
https://lists.sourceforge.net/lists/listinfo/hplip-devel

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

Yes, I am using --enable-hpijs-only-build, but also printer status and
bi-di comm are used by default by HPIJS for this printer.
Then, I link all of the HPIJS object files into a static library and
then with my app.
BTW, I link libusb also for other devices that the application uses
(i.e. USB drives).

For HPIJS, I first call usb_find_busses, ubs_find_devices, usb_open,
usb_claim_interface, usb_set_altinterface, and after done, I call
usb_release_interface, usb_close.
Are you saying I do not need to do all that?

The SYSTEM_ERROR is returned by constructor of Job class and is defined
in APDK documentation.
I think you maybe thinking about HPIJS server which I do not use.
I use hplip-2.7.7/prnt/hpijs code by itself.

I guess it is possible that I call usb_close while the buffered data is
still being transmitted.
I would expect usblib to take care of this and hpijs to work properly
with it.
I thought that when SendRasters returns, it is OK call NewPage() and to
close usb.

Martin



-----Original Message-----
From: Suffield, David [mailto:***@hp.com]
Sent: Wednesday, August 29, 2007 11:28 AM
To: Martin Kajdas; hplip-***@lists.sourceforge.net
Subject: RE: [HPLIP-Devel] Print errors and hangs using HPIJS with HP
DJ4160

I also link it with libusb 0.1.12 and other things.

If you are using hpijs only, no printer status or bi-di communication.
Then with configure option --enable-hpijs-only-build, you shouldn't need
libusb.

- sometimes (never the first time after power up) the Job is created
but the construction_error = SYSTEM_ERROR. It seems that the printer
is not responding and HPIJS times out.
Retrying works and printing can continue.

Not sure what this error is. If you use hpijs-only-build, then you
shouldn't get this error.

- sometimes HPIJS hangs on what seems like the last pass
(bottom) of the raster. Sometimes it does not hang and
NewPage() ejects the page but the bottom of the raster is incomplete
(unfinished part of a line)

This error sounds like you are closing usb io down before all the data
has be received by the printer.

-dave

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

Dave,
Thanks for the excellent tip.
I added similar code to my program and I believe it fixed the problem
with the bottom of the page being corrupted.
I wait for the printer to go idle before I close USB.

However, the problem with SYSTEM_ERROR still occurs and I will give you
some more info:
1. Never happens after power up (always works OK)
2. Seems to happen when I start a new job right after another print job.
After the previous job, the printer is still doing some head movement
and cleaning (?) for a while but its status says it is idle.
3. Querying ID works without a problem, I can get status and ink levels,
the error happens only when I try to write bulk (print) data.
4. The error happens because printer is not responding (to bulk data)
and timeout occurs (see #3, it responds to querying)
5. Pressing X button on the front panel fixes the problem
6. Sometimes, retrying the print job also makes it work

It seems that the printer goes into some kind of error mode and does not
accept print data.
Querying always works OK.

Any ideas,

Martin

-----Original Message-----
From: Suffield, David [mailto:***@hp.com]
Sent: Wednesday, August 29, 2007 3:04 PM
To: Martin Kajdas
Subject: RE: [HPLIP-Devel] Print errors and hangs using HPIJS with HP
DJ4160

Post by Martin Kajdas
Yes, I am using --enable-hpijs-only-build, but also printer status
and bi-di comm are used by default by HPIJS for this printer.
Then, I link all of the HPIJS object files into a static library and
then with my app.
BTW, I link libusb also for other devices that the application uses
(i.e. USB drives).

ok

Post by Martin Kajdas
For HPIJS, I first call usb_find_busses, ubs_find_devices, usb_open,
usb_claim_interface, usb_set_altinterface, and after done, I call
usb_release_interface, usb_close.
Are you saying I do not need to do all that?

ok, so you are doing your own device discovery.

Post by Martin Kajdas
The SYSTEM_ERROR is returned by constructor of Job class and is
defined in APDK documentation.
I think you maybe thinking about HPIJS server which I do not use.
I use hplip-2.7.7/prnt/hpijs code by itself.

ok, so you are not using IJS interface. If you look at the hpijs code,
there is code to force the apdk run in uni-di only. Meaning no device-id
will be queried by the apdk. Which may have something to do with this
error during printer instantiation.

Post by Martin Kajdas
I guess it is possible that I call usb_close while the buffered data
is still being transmitted.
I would expect usblib to take care of this and hpijs to work properly
with it.
I thought that when SendRasters returns, it is OK call
NewPage() and to close usb.

It's been my experience that libusb will not take care of flushing the
data pipe line for you. If you look at my hp.c code I wait for the
printer to go idle by querying device-id, then close the usb device.

-dave

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

I have more details after tracing the communication:
The error comes from dj3320.cpp, line 2928 (bCreditInitialized == 0)
because UpdateState(TRUE) above never returns TRUE.
I traced UpdateState and the error comes from GetPackets(...) == 0
I do not know the LDL protocol but there is something wrong with the
printer sometimes.
What can I do?
Martin


-----Original Message-----
From: Martin Kajdas
Sent: Friday, August 31, 2007 8:41 AM
To: 'Suffield, David'
Cc: hplip-***@lists.sourceforge.net
Subject: RE: [HPLIP-Devel] Print errors and hangs using HPIJS with HP
DJ4160

Dave,
Thanks for the excellent tip.
I added similar code to my program and I believe it fixed the problem
with the bottom of the page being corrupted.
I wait for the printer to go idle before I close USB.

However, the problem with SYSTEM_ERROR still occurs and I will give you
some more info:
1. Never happens after power up (always works OK) 2. Seems to happen
when I start a new job right after another print job. After the previous
job, the printer is still doing some head movement and cleaning (?) for
a while but its status says it is idle.
3. Querying ID works without a problem, I can get status and ink levels,
the error happens only when I try to write bulk (print) data.
4. The error happens because printer is not responding (to bulk data)
and timeout occurs (see #3, it responds to querying) 5. Pressing X
button on the front panel fixes the problem 6. Sometimes, retrying the
print job also makes it work

It seems that the printer goes into some kind of error mode and does not
accept print data.
Querying always works OK.

Any ideas,

Martin

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Suffield, David]

Credit pacing should not be enable for LDL. You are not running with
bi-di disabled. You need to look at the hpijs glue code and see how
bi-di is disabled (ie: see RestIOMode) before the printer job
instantiation.

-dave

-----Original Message-----
Sent: Friday, August 31, 2007 11:55 AM
To: Suffield, David
Subject: RE: [HPLIP-Devel] Print errors and hangs using HPIJS
with HP DJ4160
The error comes from dj3320.cpp, line 2928
(bCreditInitialized == 0) because UpdateState(TRUE) above
never returns TRUE.
I traced UpdateState and the error comes from GetPackets(...)
== 0 I do not know the LDL protocol but there is something
wrong with the printer sometimes.
What can I do?
Martin
-----Original Message-----
From: Martin Kajdas
Sent: Friday, August 31, 2007 8:41 AM
To: 'Suffield, David'
Subject: RE: [HPLIP-Devel] Print errors and hangs using HPIJS
with HP DJ4160
Dave,
Thanks for the excellent tip.
I added similar code to my program and I believe it fixed the
problem with the bottom of the page being corrupted.
I wait for the printer to go idle before I close USB.
However, the problem with SYSTEM_ERROR still occurs and I
1. Never happens after power up (always works OK) 2. Seems to
happen when I start a new job right after another print job.
After the previous job, the printer is still doing some head
movement and cleaning (?) for a while but its status says it is idle.
3. Querying ID works without a problem, I can get status and
ink levels, the error happens only when I try to write bulk
(print) data.
4. The error happens because printer is not responding (to
bulk data) and timeout occurs (see #3, it responds to
querying) 5. Pressing X button on the front panel fixes the
problem 6. Sometimes, retrying the print job also makes it work
It seems that the printer goes into some kind of error mode
and does not accept print data.
Querying always works OK.
Any ideas,
Martin

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

That was it.
It was my bug; I already did call ResetIOMode but because of my bug it
did not execute properly.
After I fixed the bug, it skips the credit pacing like it suppose to and
it seems to work better and faster.

Amazingly, it worked 90% of the time and I did not suspect something
this critical being broken.

The problem is of course is understanding LDL and as far as I know the
spec is not released.
Is this so and why?

I will do some more testing and report any issues.
Martin


-----Original Message-----
From: Suffield, David [mailto:***@hp.com]
Sent: Friday, August 31, 2007 12:33 PM
To: Martin Kajdas
Cc: hplip-***@lists.sourceforge.net
Subject: RE: [HPLIP-Devel] Print errors and hangs using HPIJS with HP
DJ4160

Credit pacing should not be enable for LDL. You are not running with
bi-di disabled. You need to look at the hpijs glue code and see how
bi-di is disabled (ie: see RestIOMode) before the printer job
instantiation.

-dave

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

I noticed that when I switched to uni-di mode, the BusyWait() is not
called anymore.
BusyWait function had all my calls to OS to keep it alive.
Now, my keyboard (and screen) are not updated until the printer
finishes.
To fix this I added the OS calls to ToDevice() which fixed the problem
but I am wondering if this is the best place to put it.
Any suggestions?

Also, when the OS detects Escape key, I return JOB_CANCELED from
ToDevice() but it does not stop gracefully.
What is the proper way to cancel the job?
Martin


-----Original Message-----
From: hplip-devel-***@lists.sourceforge.net
[mailto:hplip-devel-***@lists.sourceforge.net] On Behalf Of Martin
Kajdas
Sent: Friday, August 31, 2007 2:10 PM
To: Suffield, David
Cc: hplip-***@lists.sourceforge.net
Subject: Re: [HPLIP-Devel] Print errors and hangs using HPIJS with HP
DJ4160

That was it.
It was my bug; I already did call ResetIOMode but because of my bug it
did not execute properly.
After I fixed the bug, it skips the credit pacing like it suppose to and
it seems to work better and faster.

Amazingly, it worked 90% of the time and I did not suspect something
this critical being broken.

The problem is of course is understanding LDL and as far as I know the
spec is not released.
Is this so and why?

I will do some more testing and report any issues.
Martin


-----Original Message-----
From: Suffield, David [mailto:***@hp.com]
Sent: Friday, August 31, 2007 12:33 PM
To: Martin Kajdas
Cc: hplip-***@lists.sourceforge.net
Subject: RE: [HPLIP-Devel] Print errors and hangs using HPIJS with HP
DJ4160

Credit pacing should not be enable for LDL. You are not running with
bi-di disabled. You need to look at the hpijs glue code and see how
bi-di is disabled (ie: see RestIOMode) before the printer job
instantiation.

-dave

------------------------------------------------------------------------
-
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
HPLIP-Devel mailing list
HPLIP-***@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/hplip-devel

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

I have an embedded system that uses HPIJS only to talk to the printer
(DeskJet D4160) using dj3320.cpp and I it is mostly working.
However, I cannot find a proper way to abort printing.
When a user presses an Escape key while printing, I set DisplayStatus to
DISPLAY_PRINTING_CANCELED and return JOB_CANCELED in ToDevice() and
BusyWait() functions which test the escape flag every time they are
called.

I do not remember all of the details now, but the printing stops but
some debugging errors are generated and/or printed on screen and it
takes a while for the printer to recover or it is not responsive at all
(blinking lights?).

Before I start bothering you with the details, I would like to find out
if what I am doing is correct and what is the proper sequence of
aborting because most likely it is my code's fault.


Anther, issue I have is that the last print line (about 10mm wide) on
the page is only done once instead of twice (?) and it appears much
lighter and the color is not completely filled in.
I tried all kinds of delays and status checking to let the printer
finish the job before ejecting, without any success.
Any ideas?

Martin

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Cauligi, Raghothama S]

Are you looking at the return value from Job::SendRaster?
Looks like hpijs.cpp is ignoring the return value from SendRaster. In
the main loop,
it is calling pSS->ProcessRaster and ignoring the return value. You will
have to
look at the return value and break out of the loop if you get a
JOB_CANCELED error.
Raghu

-----Original Message-----
From: hplip-devel-***@lists.sourceforge.net
[mailto:hplip-devel-***@lists.sourceforge.net] On Behalf Of Martin
Kajdas
Sent: Tuesday, October 23, 2007 8:41 AM
To: Suffield, David
Cc: hplip-***@lists.sourceforge.net
Subject: [HPLIP-Devel] How to abort printing

I have an embedded system that uses HPIJS only to talk to the printer
(DeskJet D4160) using dj3320.cpp and I it is mostly working.
However, I cannot find a proper way to abort printing.
When a user presses an Escape key while printing, I set DisplayStatus to
DISPLAY_PRINTING_CANCELED and return JOB_CANCELED in ToDevice() and
BusyWait() functions which test the escape flag every time they are
called.

I do not remember all of the details now, but the printing stops but
some debugging errors are generated and/or printed on screen and it
takes a while for the printer to recover or it is not responsive at all
(blinking lights?).

Before I start bothering you with the details, I would like to find out
if what I am doing is correct and what is the proper sequence of
aborting because most likely it is my code's fault.


Anther, issue I have is that the last print line (about 10mm wide) on
the page is only done once instead of twice (?) and it appears much
lighter and the color is not completely filled in.
I tried all kinds of delays and status checking to let the printer
finish the job before ejecting, without any success.
Any ideas?

Martin

------------------------------------------------------------------------
-
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
HPLIP-Devel mailing list
HPLIP-***@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/hplip-devel

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

This is a statically linked app using HP's APDK specification so HPIJS
server (hpijs.cpp) is not used.
I checked my loop and I do exit page printing loop when SendRaster()
returns any error > 0.
Then I call NewPage() which returns another status and if this status is
also > 0, then I exit job completely.
This may be my problem and I will confirm it (SendRaster error should
not be overwritten by NewPage error: could be 0 by then).
Martin



-----Original Message-----
From: Cauligi, Raghothama S [mailto:***@hp.com]
Sent: Tuesday, October 23, 2007 9:12 AM
To: Martin Kajdas; Suffield, David
Cc: hplip-***@lists.sourceforge.net
Subject: RE: [HPLIP-Devel] How to abort printing

Are you looking at the return value from Job::SendRaster?
Looks like hpijs.cpp is ignoring the return value from SendRaster. In
the main loop, it is calling pSS->ProcessRaster and ignoring the return
value. You will have to look at the return value and break out of the
loop if you get a JOB_CANCELED error.
Raghu

-----Original Message-----
From: hplip-devel-***@lists.sourceforge.net
[mailto:hplip-devel-***@lists.sourceforge.net] On Behalf Of Martin
Kajdas
Sent: Tuesday, October 23, 2007 8:41 AM
To: Suffield, David
Cc: hplip-***@lists.sourceforge.net
Subject: [HPLIP-Devel] How to abort printing

I have an embedded system that uses HPIJS only to talk to the printer
(DeskJet D4160) using dj3320.cpp and I it is mostly working.
However, I cannot find a proper way to abort printing.
When a user presses an Escape key while printing, I set DisplayStatus to
DISPLAY_PRINTING_CANCELED and return JOB_CANCELED in ToDevice() and
BusyWait() functions which test the escape flag every time they are
called.

I do not remember all of the details now, but the printing stops but
some debugging errors are generated and/or printed on screen and it
takes a while for the printer to recover or it is not responsive at all
(blinking lights?).

Before I start bothering you with the details, I would like to find out
if what I am doing is correct and what is the proper sequence of
aborting because most likely it is my code's fault.


Anther, issue I have is that the last print line (about 10mm wide) on
the page is only done once instead of twice (?) and it appears much
lighter and the color is not completely filled in.
I tried all kinds of delays and status checking to let the printer
finish the job before ejecting, without any success.
Any ideas?

Martin

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

I am sorry, but previous email was incorrect.
I exit both loops when SendRaster() returns any error > 0, so no NewPage
is called().
My guess is that something happens that breaks the printer communication
sequence and the printer or USB driver gets confused.
The latest HPIJS fixed many problems but I believe it is not completely
debugged.
Martin


-----Original Message-----
From: hplip-devel-***@lists.sourceforge.net
[mailto:hplip-devel-***@lists.sourceforge.net] On Behalf Of Martin
Kajdas
Sent: Tuesday, October 23, 2007 9:35 AM
To: Cauligi, Raghothama S; Suffield, David
Cc: hplip-***@lists.sourceforge.net
Subject: Re: [HPLIP-Devel] How to abort printing

This is a statically linked app using HP's APDK specification so HPIJS
server (hpijs.cpp) is not used.
I checked my loop and I do exit page printing loop when SendRaster()
returns any error > 0.
Then I call NewPage() which returns another status and if this status is
also > 0, then I exit job completely.
This may be my problem and I will confirm it (SendRaster error should
not be overwritten by NewPage error: could be 0 by then).
Martin

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/

[Martin Kajdas]

I tried to duplicate the problem and I do not think the problem is with
aborting the print.
It seems to abort fine (at least for now).
However, the problem is with the printer doing its stuff (cleaning?)
after the abort and after normal print also, when if I send another job
to it, it either keeps blinking or just hangs up.
I believe that the printer returns IDLE status when the new job is sent
because I test for it when I close the (previous) job.
I need to reset the computer and printer to recover.
Martin


-----Original Message-----
From: Martin Kajdas
Sent: Tuesday, October 23, 2007 10:16 AM
To: Martin Kajdas; Cauligi, Raghothama S; Suffield, David
Cc: hplip-***@lists.sourceforge.net
Subject: RE: [HPLIP-Devel] How to abort printing

I am sorry, but previous email was incorrect.
I exit both loops when SendRaster() returns any error > 0, so no NewPage
is called().
My guess is that something happens that breaks the printer communication
sequence and the printer or USB driver gets confused.
The latest HPIJS fixed many problems but I believe it is not completely
debugged.
Martin


-----Original Message-----
From: hplip-devel-***@lists.sourceforge.net
[mailto:hplip-devel-***@lists.sourceforge.net] On Behalf Of Martin
Kajdas
Sent: Tuesday, October 23, 2007 9:35 AM
To: Cauligi, Raghothama S; Suffield, David
Cc: hplip-***@lists.sourceforge.net
Subject: Re: [HPLIP-Devel] How to abort printing

This is a statically linked app using HP's APDK specification so HPIJS
server (hpijs.cpp) is not used.
I checked my loop and I do exit page printing loop when SendRaster()
returns any error > 0.
Then I call NewPage() which returns another status and if this status is
also > 0, then I exit job completely.
This may be my problem and I will confirm it (SendRaster error should
not be overwritten by NewPage error: could be 0 by then).
Martin

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems? Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/