[Bug 825931] Re: FFe: multiarch support of qt4-x11 dependencies (libmng, lcms)

[Scott Kitterman]

Ack. Please go ahead.

** Changed in: lcms (Ubuntu)
       Status: New => Confirmed

** Changed in: libmng (Ubuntu)
       Status: New => Confirmed

** Also affects: bluedevil (Ubuntu)
   Importance: Undecided
       Status: New

** Changed in: bluedevil (Ubuntu)
       Status: New => Won't Fix

** Changed in: bluedevil (Ubuntu)
       Status: Won't Fix => Triaged

** Changed in: bluedevil (Ubuntu)
       Status: Triaged => Confirmed

** Changed in: bluedevil (Ubuntu)
   Importance: Undecided => Low

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to lcms in Ubuntu.
https://bugs.launchpad.net/bugs/825931

Title:
  FFe: multiarch support of qt4-x11 dependencies (libmng, lcms)

Status in “bluedevil” package in Ubuntu:
  Confirmed
Status in “lcms” package in Ubuntu:
  Confirmed
Status in “libmng” package in Ubuntu:
  Confirmed

Bug description:
  Two dependency library packages need to be converted for multiarch in
  order for libqtgui4 to be multiarch co-installable: libmng, and lcms.

  Risk: both of these are very simple libraries with no plugins, so the
  risk of runtime regression is small.  (Only from something trying to
  dlopen() a path it shouldn't.)  lcms has a .la file with one reference
  in the archive, in the 'libdevil-dev' package; a no-change rebuild of
  this package is trivial to do to fix this.  Otherwise, the only risks
  are from reverse-build-dependencies of the libraries having build
  systems that make wrong assumptions about how to find the libraries.
  lcms has a number of reverse-build-dependenices, but the vast majority
  of these have traditional build systems that are multiarch-proofed so
  the risk is low.  The highest risk may be from openjdk-6, but
  openjdk-6 has also built successfully without incident after multiarch
  conversion of a number of its other build dependencies so the risk
  here should also be low - and manageable in the unlikely case of any
  breakage.  If necessary, I can do test rebuilds here of all the
  reverse-dependencies before uploading, although the risk is low enough
  and we're soon enough after feature freeze that I believe it would be
  better to upload first and identify any regressions in parallel.

  Benefit: together with a fix for bug #825689, this is sufficient to
  allow all the qt dependencies of skype to be installed in multiarch
  configuration, eliminating a major reason for users to install
  ia32-libs on their systems.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/bluedevil/+bug/825931/+subscriptions