Bug#923457: test_likelihood_nh (Failed)

2019-03-12 Thread Andreas Tille
Hi Julien On Tue, Mar 12, 2019 at 11:56:49AM +0100, Julien Y. Dutheil wrote: > Ok, I have committed a patch removing the test. But just realized I did not > update the changelog... should that now be version 2.4.1-3? I've updated d/changelog but will *not* upload anything since meanwhile the

Bug#923457: test_likelihood_nh (Failed)

2019-03-12 Thread Julien Y. Dutheil
Dear Andreas, Ok, great. Will try to fix that unit test thing upstream for next time. Best, Julien. On Tue, Mar 12, 2019 at 3:18 PM Andreas Tille wrote: > Hi Julien > > On Tue, Mar 12, 2019 at 11:56:49AM +0100, Julien Y. Dutheil wrote: > > Ok, I have committed a patch removing the test. But

Bug#923457: test_likelihood_nh (Failed)

2019-03-12 Thread Andreas Tille
On Tue, Mar 12, 2019 at 11:56:49AM +0100, Julien Y. Dutheil wrote: > Dear Andreas, > > Ok, I have committed a patch removing the test. But just realized I did not > update the changelog... should that now be version 2.4.1-3? Thanks, I'll fix the changelog. Andreas. > Cheers, > > Julien. > >

Bug#923457: test_likelihood_nh (Failed)

2019-03-12 Thread Julien Y. Dutheil
Dear Andreas, Ok, I have committed a patch removing the test. But just realized I did not update the changelog... should that now be version 2.4.1-3? Cheers, Julien. On Tue, Mar 12, 2019 at 11:03 AM Andreas Tille wrote: > On Tue, Mar 12, 2019 at 09:53:10AM +0100, Julien Y. Dutheil wrote: > >

Bug#923457: test_likelihood_nh (Failed)

2019-03-12 Thread Andreas Tille
On Tue, Mar 12, 2019 at 09:53:10AM +0100, Julien Y. Dutheil wrote: > This all boils down to the fact that this unit test is badly designed: it > basically maximizes a function from initial random values. I some > particular combinations of initial values though, the convergence may fail > or take

Bug#923457: test_likelihood_nh (Failed)

2019-03-12 Thread Julien Y. Dutheil
Dear Andreas, This all boils down to the fact that this unit test is badly designed: it basically maximizes a function from initial random values. I some particular combinations of initial values though, the convergence may fail or take too much time. I will redesign the test, but do not have

Bug#923457: test_likelihood_nh (Failed)

2019-03-08 Thread Julien Y. Dutheil
Dear Andreas, Will give it a look asap. Best, Julien. On Fri, Mar 8, 2019 at 9:34 AM Julien Yann Dutheil < julien.duth...@univ-montp2.fr> wrote: > Dear Andreas, > > Will give it a look asap. > > Best, > > Julien. > > On Fri, Mar 8, 2019 at 6:30 AM Andreas Tille wrote: > >> Hi Julien, >> >>

Bug#923457: test_likelihood_nh (Failed)

2019-03-07 Thread Andreas Tille
Hi Julien, after fixing bug #923457 I realised that the package does not build on all architectures[1]. S390x fails with[2] 4/14 Test #5: test_likelihood_nh ...***Failed2.44 sec / 1 - 2 \ 3 - 4 / 5 - 6 \ 7 - 8 / 9 - 10 \ 11 nodes loaded. Theta0 set to 0.444785 Theta1 set to