[Distutils] Re: breakages in travis-ci and ci.appveyor

2018-10-02 Thread Robin Becker
On 01/10/2018 20:17, Matthew Brett wrote: Hi, ... in OSX and the same for manylinux. I think that's related to changes in "multibuild/supported_wheels.py" intended to address wheel 0.32.0 :( Thanks for the report. I think that's fixed now, in the `devel` branch of multibuild.

[Distutils] Re: breakages in travis-ci and ci.appveyor

2018-10-01 Thread Matthew Brett
Hi, On Mon, Oct 1, 2018 at 3:51 PM Robin Becker wrote: > > Thanks Dustin & Matthew, > > Appveyor pinning was relatively easy. I tried to pin in multibuild using the > approach of pillow, but now I am getting > > > macosx_10_6_intel.macosx_10_9_intel.macosx_10_9_x86_64.macosx_10_10_intel.macosx_10

[Distutils] Re: breakages in travis-ci and ci.appveyor

2018-10-01 Thread Robin Becker
Thanks Dustin & Matthew, Appveyor pinning was relatively easy. I tried to pin in multibuild using the approach of pillow, but now I am getting macosx_10_6_intel.macosx_10_9_intel.macosx_10_9_x86_64.macosx_10_10_intel.macosx_10_10_x86_64.whl Traceback (most recent call last): File "multibuild

[Distutils] Re: breakages in travis-ci and ci.appveyor

2018-10-01 Thread Dustin Ingram
Here is the relevant issue: https://github.com/pypa/auditwheel/issues/102 On Mon, Oct 1, 2018 at 7:54 AM Robin Becker wrote: > I am using Matthew Brett's manylinux builder setup ( > https://github.com/matthew-brett/multibuild) for reportlab and just > started to > see this > > > File > "/opt/_in