insighttolkit4: remove everywhere except amd64 and i386

Bug #1812874 reported by Gianfranco Costamagna on 2019-01-22
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
ants (Ubuntu)
Undecided
Unassigned
camitk (Ubuntu)
Undecided
Unassigned
fw4spl (Ubuntu)
Undecided
Unassigned
ginkgocadx (Ubuntu)
Undecided
Unassigned
insighttoolkit4 (Ubuntu)
High
Unassigned
itksnap (Ubuntu)
Undecided
Unassigned
nifti2dicom (Ubuntu)
Undecided
Unassigned
otb (Ubuntu)
Undecided
Unassigned
plastimatch (Ubuntu)
Undecided
Unassigned
simpleitk (Ubuntu)
Undecided
Unassigned
vmtk (Ubuntu)
Undecided
Unassigned

Bug Description

Insighttoolkit4 is a complex software, and upstream can't make it work on anything except amd64 and i386.
For such scientific software, disabling testsuite is a bad thing, because it results in a broken software, giving wrong results to the end user without noticing them.

Please kick arm64 armhf ppc64el and s390x out from the archive, including reverse-deps.

We should just sync the next release

Matthias Klose (doko) wrote :

why should we do this, is there any reference from upstream about this?

Changed in insighttoolkit4 (Ubuntu):
status: New → Incomplete

yes, but I forgot where.

https://insightsoftwareconsortium.atlassian.net/browse/ITK-3412

btw the analysis about what can go wrong if ITK misbehaves is done by an ex-colleague of mine, who used it in his tumor-related research university master thesis.

A wrong analysis can draw the user into the wrong conclusions, and the worst part is that different architecture might give different results.

unless somebody does a real porting, I would recommend to not ship that software at all where not supported.

Changed in insighttoolkit4 (Ubuntu):
status: Incomplete → New
importance: Undecided → High

Another prove is here:
https://launchpad.net/ubuntu/+source/camitk/4.1.2-2build1

camitk (a reverse-dependency) has now a testsuite that fails everywhere except amd64 and i386, and failing tests are related to itk bugs

Matthias Klose (doko) wrote :

rdeps are missing

Changed in insighttoolkit4 (Ubuntu):
status: New → Incomplete

fw4spl should just be kicked out from release

Changed in camitk (Ubuntu):
status: New → Fix Released
Steve Langasek (vorlon) wrote :

I have reuploaded insighttoolkit4 to fail the build when the test suite builds (a precondition for this removal is that the binaries won't grow back on reupload) and it has now failed to build on all non-x86 architectures https://launchpad.net/ubuntu/+source/insighttoolkit4/4.12.2-dfsg1-4ubuntu7.

I have removed the binaries for all of these packages from the eoan release pocket on armhf,arm64,ppc64el,s390x. So this removal is done.

I agree that insighttoolkit4 should become a sync again in the next Debian upload. We should not be wasting (considerable) buildd time in Ubuntu only just to try and fail to build the package on architectures unsupported by Debian or upstream.

Changed in ants (Ubuntu):
status: New → Fix Released
Changed in fw4spl (Ubuntu):
status: New → Fix Released
Changed in ginkgocadx (Ubuntu):
status: New → Fix Released
Changed in insighttoolkit4 (Ubuntu):
status: Incomplete → Fix Released
Changed in itksnap (Ubuntu):
status: New → Fix Released
Changed in nifti2dicom (Ubuntu):
status: New → Fix Released
Changed in otb (Ubuntu):
status: New → Fix Released
Changed in plastimatch (Ubuntu):
status: New → Fix Released
Steve Langasek (vorlon) on 2019-04-23
Changed in simpleitk (Ubuntu):
status: New → Fix Released
Changed in vmtk (Ubuntu):
status: New → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers