Looks like it's missing dep on scimath.
Created attachment 384216 [details] build.log Sorry, this was supposed to be attached already, not sure exactly what happened.
numpy's keywords are: alpha amd64 arm ~arm64 hppa ia64 ~mips ppc ppc64 ~s390 ~sh sparc x86 ~amd64-fbsd ~x86-fbsd ~x86-freebsd ~x86-interix ~amd64-linux ~arm-linux ~x86-linux ~ppc-macos ~x64-macos ~x86-macos ~sparc-solaris ~x64-solaris ~x86-solaris but scimath's keywords are: ~amd64 ~x86 ~amd64-linux ~x86-linux That leaves a pretty big delta to take care of before we can add this dep: alpha amd64 arm ~arm64 hppa ia64 ~mips ppc ppc64 ~s390 ~sh sparc x86 ~amd64-fbsd ~x86-fbsd ~x86-freebsd ~x86-interix ~arm-linux ~ppc-macos ~x64-macos ~x86-macos ~sparc-solaris ~x64-solaris ~x86-solaris
Arches, I'm sorry...numpy is missing this dep so we need to get this fixed. STABLEREQ, target keywords: alpha amd64 arm hppa ia64 ppc ppc64 x86 KEYWORDREQ, target keywords: ~arm64 ~mips ~s390 ~sh sparc ~amd64-fbsd ~x86-fbsd ~x86-freebsd ~x86-interix ~arm-linux ~ppc-macos ~x64-macos ~x86-macos ~sparc-solaris ~x64-solaris ~x86-solaris
Sorry, target version looks like dev-python/scimath-4.1.2-r1
Which packages are supposed to depend on dev-python/scimath?
(In reply to Jeroen Roovers from comment #5) > Which packages are supposed to depend on dev-python/scimath? dev-python/numpy-*
(In reply to Rick Farina (Zero_Chaos) from comment #6) > (In reply to Jeroen Roovers from comment #5) > > Which packages are supposed to depend on dev-python/scimath? > > dev-python/numpy-* But they don't. Please fix that first.
(In reply to Jeroen Roovers from comment #7) > (In reply to Rick Farina (Zero_Chaos) from comment #6) > > (In reply to Jeroen Roovers from comment #5) > > > Which packages are supposed to depend on dev-python/scimath? > > > > dev-python/numpy-* > > But they don't. Please fix that first. I don't believe that dropping all the keywords, including stable ones, from numpy would be appreciated. Or perhaps I'm to mask a stable and widely used package? Or I'll just add the dep and all the users will get visibility issues? No, how about we just add the keywords that I requested please. Thanks.
I've been informed by Arfrever that I read the code wrong and this is numpy's internal scimath that is failing, not the package scimath. Apologies for the noise to the arch teams. The problem seems to be in numpy...just not sure what yet.
Okay, despite replicating this bug on multiple systems it seems there was a deep linking issue with blas-reference. It looks like somehow blas-reference was installed but not eselected (despite no other blas provider ever being pulled in) and this caused several issues. If I can reliably reproduce what causes this I'll open a new bug for whatever the issue with blas or eselect-blas is.