Bug#1116026: transition: octave-10
What needs to happen to get this migrated: does something need to change
in the octave and/or octave-statistics packages, or do you need to add a
hint to try migrating octave and octave-statistics together? Do other
Octave packages need to stay unchanged until this is resolved (in
particular, is fixing unrelated non-regression RC bugs such as nlopt
#1113328 good or bad at this point)?
Adrian Bunk wrote:
> It installed octave-statistics/sid binaries, but was still using
octave-statistics/testing sources
The failing tests are because Octave 10 changed the text of an error
message, and the fix in 1.7.5-4 is to skip the affected tests in
debian/checkvars, so it makes sense that 1.7.5-3 source + 1.7.5-4 binary
still fails. (And it looks like autopkgtest-pkg-octave takes the test
code from the source not the installed binary, so 1.7.5-3 source + a
hypothetical 1.7.5-5 binary that fixed or removed the tests themselves
would also fail.)
https://ci.debian.net/packages/o/octave-statistics/testing/amd64/64883259/
https://salsa.debian.org/pkg-octave-team/octave-statistics/-/commit/72e3af465b7aad739582d288081188ff76302911
There are autopkgtest runs that did try octave, octave-statistics, nlopt
and astra-toolbox together (and passed), but I can't find an actual
attempt to migrate this combination in
https://release.debian.org/britney/excuses.yaml or
https://release.debian.org/britney/update_output.txt
https://ci.debian.net/packages/o/octave-statistics/testing/amd64/64906953/
Reply to: