Bug#913120: cups-filters: please favor graphicsmagick-imagemagick-compat over imagemagick

2018-11-11 Thread Didier 'OdyX' Raboud
Control: tags -1 +pending

Le mercredi, 7 novembre 2018, 23.40:23 h CET Samuel Thibault a écrit :
> Mmm, actually it's even a very specific case of braille embossing, which
> is really not common, and most probably really never used in networked
> situations. Perhaps we could just demote to "Suggests" or even drop
> it. If the user tries to emboss an image, she'll get an error message
> hinting to install imagemagick anyway.

That seems like a workable way forward, I'll go with it.

Cheers, 
OdyX

signature.asc
Description: This is a digitally signed message part.


Bug#913120: cups-filters: please favor graphicsmagick-imagemagick-compat over imagemagick

2018-11-07 Thread Samuel Thibault
Hello,

Moritz Mühlenhoff, le mer. 07 nov. 2018 23:04:55 +0100, a ecrit:
> In this specific case other, more promising hardening options would be:
> - IM is only used for the braille support, so this could be split into
>   a separate binary package, reducing the attack footprint for the non-braille
>   installations of cups-filters

Mmm, actually it's even a very specific case of braille embossing, which
is really not common, and most probably really never used in networked
situations. Perhaps we could just demote to "Suggests" or even drop
it. If the user tries to emboss an image, she'll get an error message
hinting to install imagemagick anyway.

Samuel



Bug#913120: cups-filters: please favor graphicsmagick-imagemagick-compat over imagemagick

2018-11-07 Thread Moritz Mühlenhoff
On Wed, Nov 07, 2018 at 08:44:25AM +0100, Jonas Smedegaard wrote:
> Source: cups-filters
> Version: 1.21.3-2
> Severity: important
> Tags: security
> 
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA512
> 
> Graphicsmagick is a drop-in replacement for imagemagick which - apart
> from being faster and lighter - also claims that it "suffers from fewer
> security issues and exploits" - which seems to correspond with the
> amount of issues reported at
> https://security-tracker.debian.org/tracker/source-package/imagemagick
> and
> https://security-tracker.debian.org/tracker/source-package/graphicsmagick

Counting CVE IDs is not a useful metric per se, before we migrate things this
should be researched more in depth (and then whatever magick is preferred
should be migrated distro-wide (and the other variant phased out)).

In this specific case other, more promising hardening options would be:
- IM is only used for the braille support, so this could be split into
  a separate binary package, reducing the attack footprint for the non-braille
  installations of cups-filters
- The respective code shells out to convert; the execution of that binary
  could be contained with firejail (e.g. with a profile running imagemagick
  with the seccomp filter and namespaced), mitigating the effect of an
  exploit in imagemagick.

Cheers,
Moritz
  



Bug#913120: cups-filters: please favor graphicsmagick-imagemagick-compat over imagemagick

2018-11-06 Thread Jonas Smedegaard
Source: cups-filters
Version: 1.21.3-2
Severity: important
Tags: security

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

Graphicsmagick is a drop-in replacement for imagemagick which - apart
from being faster and lighter - also claims that it "suffers from fewer
security issues and exploits" - which seems to correspond with the
amount of issues reported at
https://security-tracker.debian.org/tracker/source-package/imagemagick
and
https://security-tracker.debian.org/tracker/source-package/graphicsmagick

Please list graphicsmagick-imagemagick-compat as favored over
imagemagick to have the former installed by default on new systems.

 - Jonas

-BEGIN PGP SIGNATURE-
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=cM4p
-END PGP SIGNATURE-