Ghostscript command line: /usr/bin/gs :Unrecoverable error: undefined in .putdeviceprops

[William.Yeung]

PDF file data fail to rendering to raster data.

We were reported a bug from our printer users:
I print file use papersize A5 (Ubuntu 16.04.1(x64),Ghostscript 9.18) is Okay.

But, after I update (apt-get upgrade) to Ubuntu 16.04.5(x64),Ghostscript 9.25, I print the same file with the same driver there is an error reports:
"Filter Error"

I figure out that all the papersize with define "LeadingEdge" can't print success after update.
(with "LeadingEdge":A5, Statement, A4(without "LeadingEdge") is print Okay).

I view the "CUPS Error Log",there is an error reports when print job fails:
"Ghostscript command line: /usr/bin/gs -dQUIET -dPARANOIDSAFER -dNOPAUSE -dBATCH -dNOINTERPOLATE -sDEVICE=cups -sstdout=%stderr -sOutputFile=%stdout -sMediaType=Plain -sOutputType=Pass=1,Direction=1 -r600x600 -dLeadingEdge=1 -dMediaPosition=7 -dDEVICEWIDTHPOINTS=420 -dDEVICEHEIGHTPOINTS=595 -dcupsBitsPerColor=8 -dcupsColorOrder=0 -dcupsColorSpace=1 -scupsPageSizeName=A5 -I/usr/share/cups/fonts -c \'<</.HWMargins[14.400000 14.400000 14.399994 14.400024] /Margins[0 0]>>setpagedevice\' -f -_

Unrecoverable error: undefined in .putdeviceprops"

Ubuntu 18.04(x64), Ghostscript 9.22 also have the same problem.

--------------------------------------------------
ghostscript had confirm that this a bug of ghostscript 9.22(and the later version).

Below information is the correspondence of ghostscript:
https://bugs.ghostscript.com/show_bug.cgi?id=700023
---------------------------------------------------

Our printer driver users(also users of Ubuntu 16.04) can't do their normal print jobs.
So, we are eagerly hopping that you can offer us a patch(Ubuntu 16.04) to fix this bug?

[William.Yeung (will323)]

Created attachment 15946
Research file of ghostscript command test in 9.25.

PDF file data fail to rendering to raster data.

We were reported a bug from our printer users:
I print file use papersize A5 (Ubuntu 16.04.1(x64),Ghostscript 9.18) is Okay.

But, after I update (apt-get upgrade) to Ubuntu 16.04.5(x64),Ghostscript 9.25, I print the same file with the same driver there is an error reports:
"Filter Error"

I figure out that all the papersize with define "LeadingEdge" can't print success after update.
(with "LeadingEdge":A5, Statement, A4(without "LeadingEdge") is print Okay).

I view the "CUPS Error Log",there is an error reports when print job fails:
"Ghostscript command line: /usr/bin/gs -dQUIET -dPARANOIDSAFER -dNOPAUSE -dBATCH -dNOINTERPOLATE -sDEVICE=cups -sstdout=%stderr -sOutputFile=%stdout -sMediaType=Plain -sOutputType=Pass=1,Direction=1 -r600x600 -dLeadingEdge=1 -dMediaPosition=7 -dDEVICEWIDTHPOINTS=420 -dDEVICEHEIGHTPOINTS=595 -dcupsBitsPerColor=8 -dcupsColorOrder=0 -dcupsColorSpace=1 -scupsPageSizeName=A5 -I/usr/share/cups/fonts -c \'<</.HWMargins[14.400000 14.400000 14.399994 14.400024] /Margins[0 0]>>setpagedevice\' -f -_

Unrecoverable error: undefined in .putdeviceprops"

I think that this is a bug of ghostscript 9.25.

Below information are the research of I do:

(I am using Ubuntu 16.04.5(x64),Ghostscript 9.25)
I am running gs command below (with "-dLeadingEdge=1"):
gs -dQUIET -dPARANOIDSAFER -dNOPAUSE -dBATCH -dNOINTERPOLATE -sDEVICE=cups -sMediaType=Plain -sOutputType=Pass=1,Direction=1 -r600x600 -dLeadingEdge=1 -dMediaPosition=7 -dDEVICEWIDTHPOINTS=420 -dDEVICEHEIGHTPOINTS=595 -dcupsBitsPerColor=8 -dcupsColorOrder=0 -dcupsColorSpace=1 -scupsPageSizeName=A5 -o ~/Desktop/out_A5.prn ~/Desktop/d00017-001.pdf

Reports an error:
Unrecoverable error: undefined in .putdeviceprops

I am running gs command below (without "-dLeadingEdge=1"):
gs -dQUIET -dPARANOIDSAFER -dNOPAUSE -dBATCH -dNOINTERPOLATE -sDEVICE=cups -sMediaType=Plain -sOutputType=Pass=1,Direction=1 -r600x600 -dMediaPosition=7 -dDEVICEWIDTHPOINTS=420 -dDEVICEHEIGHTPOINTS=595 -dcupsBitsPerColor=8 -dcupsColorOrder=0 -dcupsColorSpace=1 -scupsPageSizeName=A5 -o ~/Desktop/out_A5.prn ~/Desktop/d00017-001.pdf

Reports no error.

Above all, we think this is a bug of ghostscript that ghostscript can't deal with the flag "-dLeadingEdge" on version 9.25 (version 9.18 is Okay.).

My questions are：
1. Do you have any plan to fix it?
2. Is there any workaround?

[Chris-liddell (chris-liddell)]

Regression caused by:

http://git.ghostscript.com/?p=ghostpdl.git;a=commit;h=478a0a862d25

[Chris-liddell (chris-liddell)]

This fixes the problem, and I think behaves correctly:

http://git.ghostscript.com/?p=user/chrisl/ghostpdl.git;a=commitdiff;h=3b91a3aaeb3b

But I'd like Ray to review the commit before I push it.

[Chris-liddell (chris-liddell)]

Fixed in:

http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=73f7c6eb9259

[William.Yeung (will323)]

(In reply to Chris Liddell (chrisl) from comment #3)
> Fixed in:
>
> http://git.ghostscript.com/?p=ghostpdl.git;a=commitdiff;h=73f7c6eb9259

Dear Chris,
   Thanks for your correspondence so quickly.

Now we knew that this a bug of ghostscript(9.25) and you have a solution for it, will you fit it in the next version(9.26) release?

And can you offer the accurate time that you release the version that fix this bug?

And is there any workaround for our printer driver customers to do the print job normally by using paper size like A5(with "LeadingEdge")?

[Ken Sharp (ken-sharp)]

(In reply to YangGuangJing from comment #4)

> Now we knew that this a bug of ghostscript(9.25)

Incorrect, I'm afraid, this was introduced in 9.22, a year ago.

> and you have a solution for
> it, will you fit it in the next version(9.26) release?

Naturally.

> And can you offer the accurate time that you release the version that fix
> this bug?

No.

> And is there any workaround for our printer driver customers to do the print
> job normally by using paper size like A5(with "LeadingEdge")?

No. You simply can't use LeadingEdge with the CUPS device.

[Chris-liddell (chris-liddell)]

I would add:

I'm assuming that as the problem arose via the cups print workflow that the "customers" you refer to are users of Ricoh printers that use Linux - i.e. they use Ghostscript as packaged by their distribution.

That being the case, in situations like this, package maintainers often pick up patches like this before the feature in an official release of ours.

I'd suggest if that doesn't happen for any of you customers, they contact the Ghostscript package maintainer for their distribution, and prompt them to look at this bug, and pull the patch into their package.

[William.Yeung (will323)]

Dear Chris,

   We were reported this error from a lot of users, and almost all of our raster printer drivers were affected by this bug.

   Since more and more our printer driver users were affected by this bug, can you speed up the process of release the next version (Version 9.26) of ghostscript.

And most importantly, our printer driver users can't do their normal print jobs. So, we are eagerly hopping that you can offer us a patch to fix this bug before your official release (Version 9.26).

>That being the case, in situations like this, package maintainers often pick up >patches like this before the feature in an official release of ours.
>I'd suggest if that doesn't happen for any of you customers, they contact the >Ghostscript package maintainer for their distribution, and prompt them to look at >this bug, and pull the patch into their package.

Our printer drivers are only supported on Linux Ubuntu OS, so can you contact the Ubuntu package maintainer to pick up this patch to their distribution?

Best Regards.

[Chris-liddell (chris-liddell)]

(In reply to YangGuangJing from comment #7)
> Dear Chris,
>
> We were reported this error from a lot of users, and almost all of our
> raster printer drivers were affected by this bug.
>
> Since more and more our printer driver users were affected by this bug,
> can you speed up the process of release the next version (Version 9.26) of
> ghostscript.
>
> And most importantly, our printer driver users can't do their normal print
> jobs. So, we are eagerly hopping that you can offer us a patch to fix this
> bug before your official release (Version 9.26).

We're planning a release within the next 6 weeks, and we really can't do it any quicker (that is already ~4 months earlier than our normal schedule).

> >That being the case, in situations like this, package maintainers often pick up >patches like this before the feature in an official release of ours.
> >I'd suggest if that doesn't happen for any of you customers, they contact the >Ghostscript package maintainer for their distribution, and prompt them to look at >this bug, and pull the patch into their package.
>
> Our printer drivers are only supported on Linux Ubuntu OS, so can you
> contact the Ubuntu package maintainer to pick up this patch to their
> distribution?

The maintainer in question (AFAIK) is Till, who you may have noted was the original assignee for this bug, and also added himself to the CC list for this bug - so he is aware.

[William.Yeung (will323)]

>We're planning a release within the next 6 weeks
>and also added himself to the CC list for this bug - so he is aware.
Thanks very much for your correspondence so quickly.

[Till Kamppeter (till-kamppeter)]

YangGuangJing, please report a bug on

https://bugs.launchpad.net/ubuntu/

for the package "ghostscript". Tell also which Ubuntu versions suffer this bug.

Thanks.

[William.Yeung (will323)]

(In reply to Till Kamppeter from comment #10)
> YangGuangJing, please report a bug on
>
> https://bugs.launchpad.net/ubuntu/
>
> for the package "ghostscript". Tell also which Ubuntu versions suffer this
> bug.
>
> Thanks.

Dear Till,
  We had report this bug on ubuntu:
https://bugs.launchpad.net/ubuntu/+source/linphone/+bug/1800062

New, we are eagerly hopping that you can offer us a patch of Ubuntu 16.04 to fix this bug(and the accurate time that we can pick up the patch).

After that(16.04),we hope that Ubuntu 18.04 patch can also be offered.

[William.Yeung (will323)]

Dear Till,

   We(RICOH printer driver developer) are still working hard to looking for an workaround to make sure our printer user can do their job normally before your(or ghostscript) official release.
We appreciate that you can private it to us if you have one.

   By the way, can you offer us the approximate time that Ubuntu release a patch of fixed this bug? If you can't, just tell us can you release the patch before November 12(11/12)?

[Till Kamppeter (till-kamppeter)]

What I can do is to create a Stable Release Update (SRU) of Ghostscript (a new package with the patch applied) for affected Ubuntu versions (18.10, 18.04, 16.04?) and upload it. Then there will be a 1-week period for users to test and confirm that the SRU solves the problem. After that the Ubuntu release team will promote the new package into the official updates.

Is this OK for you?

What leads to the date November 12?

[Till Kamppeter (till-kamppeter)]

seb128, this regression in 16.04 is caused by the security update of updating Ghostscript to 9.25. Should then perhaps the security team issue this update? And could they do it faster than going the regular SRU way? And can we already do SRUs as the D. D. (19.04) cycle has not started yet?

The update itself would be applying a simple patch, the one from the upstream bug report https://bugs.ghostscript.com/show_bug.cgi?id=700023.

[Sebastien Bacher (seb128)]

@Till, did you talk to the security team about it?

[Marc Deslauriers (mdeslaur)]

I'll prepare a security regression fix for this.

[William.Yeung (will323)]

Dear Till & Marc,
   Thanks very much for your correspondence so quickly.

>Is this OK for you?
   It's great that you can provide a SRU(or security team provide a security regression fix).

>What leads to the date November 12?
   We are asked to offer a workaround to our printer users before the date November 12, but our workaround is not stable and reliable. We are still working hard on it.
So we hope that you can fix this bug before the date.

> marc.deslauriers: I'll prepare a security regression fix for this.
   It's great that you'll prepare a security regression fix for this.
   By the way, can you offer us the approximate time that we can get the security regression fix(or a SRU)?

[Launchpad Janitor (janitor)]

This bug was fixed in the package ghostscript - 9.25~dfsg+1-0ubuntu0.14.04.2

---------------
ghostscript (9.25~dfsg+1-0ubuntu0.14.04.2) trusty-security; urgency=medium

  * SECURITY UPDATE: Multiple security issues
    - debian/patches/0218*.patch: multiple cherry-picked upstream commits
      to fix security issues. Thanks to Jonas Smedegaard for cherry-picking
      these for Debian's 9.25~dfsg-3 package.
    - debian/symbols.common: added new symbol.
    - CVE-2018-17961
    - CVE-2018-18073
    - CVE-2018-18284
  * Fix LeadingEdge regression introduced in 9.22. (LP: #1800062)
    - debian/patches/lp1800062.patch: fix cups get/put_params LeadingEdge
      logic in cups/gdevcups.c.

 -- Marc Deslauriers <email address hidden> Tue, 30 Oct 2018 09:05:40 -0400

[Launchpad Janitor (janitor)]

This bug was fixed in the package ghostscript - 9.25~dfsg+1-0ubuntu0.18.04.2

---------------
ghostscript (9.25~dfsg+1-0ubuntu0.18.04.2) bionic-security; urgency=medium

  * SECURITY UPDATE: Multiple security issues
    - debian/patches/0218*.patch: multiple cherry-picked upstream commits
      to fix security issues. Thanks to Jonas Smedegaard for cherry-picking
      these for Debian's 9.25~dfsg-3 package.
    - debian/libgs9.symbols: added new symbol.
    - CVE-2018-17961
    - CVE-2018-18073
    - CVE-2018-18284
  * Fix LeadingEdge regression introduced in 9.22. (LP: #1800062)
    - debian/patches/lp1800062.patch: fix cups get/put_params LeadingEdge
      logic in cups/gdevcups.c.
  * Fix iccprofiles directory to symlink issue (LP: #1800328)
    - debian/libgs__VER__-common.maintscript.in: make sure directory is
      correctly transitioned to a symlink.

 -- Marc Deslauriers <email address hidden> Tue, 30 Oct 2018 09:00:57 -0400

[Launchpad Janitor (janitor)]

This bug was fixed in the package ghostscript - 9.25~dfsg+1-0ubuntu0.16.04.2

---------------
ghostscript (9.25~dfsg+1-0ubuntu0.16.04.2) xenial-security; urgency=medium

  * SECURITY UPDATE: Multiple security issues
    - debian/patches/0218*.patch: multiple cherry-picked upstream commits
      to fix security issues. Thanks to Jonas Smedegaard for cherry-picking
      these for Debian's 9.25~dfsg-3 package.
    - debian/symbols.common: added new symbol.
    - CVE-2018-17961
    - CVE-2018-18073
    - CVE-2018-18284
  * Fix LeadingEdge regression introduced in 9.22. (LP: #1800062)
    - debian/patches/lp1800062.patch: fix cups get/put_params LeadingEdge
      logic in cups/gdevcups.c.

 -- Marc Deslauriers <email address hidden> Tue, 30 Oct 2018 09:04:39 -0400

[Launchpad Janitor (janitor)]

This bug was fixed in the package ghostscript - 9.25~dfsg+1-0ubuntu1.1

---------------
ghostscript (9.25~dfsg+1-0ubuntu1.1) cosmic-security; urgency=medium

  * SECURITY UPDATE: Multiple security issues
    - debian/patches/0218*.patch: multiple cherry-picked upstream commits
      to fix security issues. Thanks to Jonas Smedegaard for cherry-picking
      these for Debian's 9.25~dfsg-3 package.
    - debian/libgs9.symbols: added new symbol.
    - CVE-2018-17961
    - CVE-2018-18073
    - CVE-2018-18284
  * Fix LeadingEdge regression introduced in 9.22. (LP: #1800062)
    - debian/patches/lp1800062.patch: fix cups get/put_params LeadingEdge
      logic in cups/gdevcups.c.

 -- Marc Deslauriers <email address hidden> Tue, 30 Oct 2018 08:38:06 -0400

[Ray Johnston (ray-johnston)]

Hint: Read the documenation. Look for .htm files that contain "halftone".
Ghostscript supports PostScript Level 3 halftones and includes some
tools to do special dithering/halftoning.