Re: Ghostscript/GhostPDL 9.25 Release Candidate 1

Johannes Meixner <[email protected]> Fri, 14 Sep 2018 10:41:07 +0200 (CEST)
Newsgroups gmane.comp.printing.ghostscript.devel
Message-ID <[email protected]>
Hello Chris,

thanks for your prompt reply.

FYI:
I was on vacation so that fortunately I missed what led to
the Ghostscript 9.24 security release - but unfortunately
I was hit by all of it at once when I retuned from vacation.

On Sep 13 15:51 Chris Liddell wrote (excerpt):
> On 13/09/18 15:43, Johannes Meixner wrote:
>>
> <SNIP>
>> The call
>> --------------------------------------------------------------------
>> gs -q -r50 -dNOPAUSE -dBATCH -dSAFER \
>>  -sOutputFile=/dev/null -sDEVICE=nullpage \
>>  /usr/share/ghostscript/9.25/examples/transparency_example.ps
>> --------------------------------------------------------------------
>> fails with this gs stdout and stderr output:
>> ------------------------------------------------------------------------------
>>
>> Error: /undefined in .pushpdf14devicefilter
...
>> GPL Ghostscript RELEASE CANDIDATE 1 9.25: Unrecoverable error, exit code 1
>> ------------------------------------------------------------------------------
>>
>> I don't know if '.pushpdf14devicefilter' is also used in general
>> by other PostScript files or tools that output PostScript.
> <SNIP>
>
> It's a custom operator, specific to Ghostscript, and no, it should not
> be used anywhere else. I'm not sure if it only fails in SAFER, but,
> frankly, don't use it.

I fully agree: Don't use it.
;-)

/usr/share/ghostscript/9.25/examples/transparency_example.ps
that uses .pushpdf14devicefilter is the Ghostscript
source file examples/transparency_example.ps

Because it is a Ghostscript upstream example others may
assume .pushpdf14devicefilter could be regularly used.
At least a comment in transparency_example.ps should explain
that .pushpdf14devicefilter cannot be regularly used.

I excluded your transparency_example.ps in my ghostscript-testing
package and now the remaining Ghostscript examples succeed, see
https://build.opensuse.org/package/show/Printing/ghostscript-testing

Perhaps you at Ghostscript upstream could add such kind of test
to your test suite as I run in my ghostscript-testing package
so that you at Ghostscript upstream could notice such issues?

For the source of my simple test (bash script code)
see the '%build' section in
https://build.opensuse.org/package/view_file/Printing/ghostscript-testing/ghostscript-testing.spec?expand=1


Kind Regards
Johannes Meixner
-- 
SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard,
Graham Norton - HRB 21284 (AG Nuernberg)