On 13 Oct 2017, at 18:06, Andreas Tille <ti...@debian.org> wrote: > > OK folks, if nobody insists I'll do the following > > 1. proceed as I suggested here: > https://lists.debian.org/debian-science/2017/10/msg00001.html
The body of the rule should be "dh_auto_test || true", not "override_dh_auto_test || true". > 2. Upload after migrating the package to Debian Science as > we previously agreed upon. > > Is there anybody who is no happy about this? > > Kind regards > > Andreas. > > On Fri, Oct 13, 2017 at 07:00:04PM +0300, Adrian Bunk wrote: >> On Wed, Oct 04, 2017 at 05:23:58PM +0200, Andreas Tille wrote: >>> Package: ftp.debian.org >>> Severity: normal >>> >>> Hi ftpmasters, >>> >>> as discussed in detail on Debian Science list[1] please remove pandas >>> for the said architectures to enable lowering severity of bug #877419 >>> and let the package migrate to testing at least for the working >>> architectures. >>> >>> Thanks a lot for your ftpmaster work >>> >>> Andreas. >>> >>> [1] https://lists.debian.org/debian-science/2017/10/msg00009.html >> >> The dak command used in this email is misleading. >> >> -A, --no-arch-all-rdeps Do not report breaking arch:all packages >> or Build-Depends-Indep >> >> Without -A you get: >> # Broken Depends: >> pandas: python-pandas >> python3-pandas >> >> And pandas is not a leaf package. >> >> The removal you are requesting would make pandas no longer available on >> these architectures. >> >> Due to that that you would also remove all binary-any reverse >> (build) dependencies. >> >> As an example, testing migration of statsmodels is currently blocked by >> missing builds due to non-installable pandas on the architectures where >> you want to remove pandas. >> >> >> What causes these FTBFS to appear is likely related to: >> >> pandas (0.20.3-1) unstable; urgency=medium >> ... >> - enabled back all the tests for now >> ... >> -- Yaroslav Halchenko <deb...@onerussian.com> Mon, 10 Jul 2017 20:00:59 >> -0400 >> >> >> This sounds as if pandas previously needed some fiddling with the tests, >> and doing that again is likely a better option.[1] >> >> >> cu >> Adrian >> >> [1] ideally each test failure should be debugged properly, >> but as short-term hack disabling some tests would be an option >> >> -- >> >> "Is there not promise of rain?" Ling Tan asked suddenly out >> of the darkness. There had been need of rain for many days. >> "Only a promise," Lao Er said. >> Pearl S. Buck - Dragon Seed >> >> > > -- > http://fam-tille.de >
