12:54 <sb0> bb-m-labs, force build --revision=f5cd0928c3f7fbdc2a09764b452dc27f7422ee4d artiq

12:54 <bb-m-labs> build forced [ETA 39m58s]

12:54 <bb-m-labs> I'll give a shout when the build finishes

12:55 <bb-m-labs> build #1449 of artiq is complete: Failure [failed get_output_name conda_remove] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1449

12:55 <sb0> sigh

12:55 <sb0> yeah...

12:55 <sb0> I wonder what caused the windows lda test to start failing

13:02 <sb0> seems to be some pesky python/setuptools/conda breakage https://github.com/conan-io/conan/issues/298

13:03 <sb0> the other controller test that doesn't use the installed command doesn't fail with this "failed to create process." error

13:03 <sb0> I wonder if commands like artiq_master are also broken then ...

16:24 <GitHub> [artiq] webstersc commented on issue #670: Installing the most recent version of artiq 2.2 on the dev channel on Windows works for me now, apart from the command line tools failing with the error 'failed to create process' - the master & dashboard have to be run using python -m artiq.frontend.xxx. https://github.com/m-labs/artiq/issues/670#issuecomment-292583119

16:38 <GitHub> [artiq] whitequark commented on issue #670: I think that error is https://github.com/conan-io/conan/issues/298. https://github.com/m-labs/artiq/issues/670#issuecomment-292586924

16:44 <GitHub> [artiq] jordens commented on issue #670: Deeper down, more like pypa/setuptools#398 i.e. buildbot might need updating. https://github.com/m-labs/artiq/issues/670#issuecomment-292588454

17:16 <GitHub> [artiq] jordens opened issue #703: unification issue https://github.com/m-labs/artiq/issues/703

17:19 <GitHub> [artiq] whitequark commented on issue #703: Well, the compiler should never raise an exception instead of a diagnostic, so there's at least one clear bug here. https://github.com/m-labs/artiq/issues/703#issuecomment-292597875

17:20 <GitHub> [artiq] whitequark commented on issue #703: That said, Python arithmetics, which automatically expands widths, and is also polymorphic in terms of operands, is problematic to implement in an unification-based system, so the answer to your question is "not necessarily". https://github.com/m-labs/artiq/issues/703#issuecomment-292598105

17:23 <GitHub> [artiq] jordens commented on issue #703: but isn't at least this specific case of `i32 -> i64` well defined? https://github.com/m-labs/artiq/issues/703#issuecomment-292598792

17:26 <GitHub> [artiq] whitequark commented on issue #703: I think it's just a bug, yes. https://github.com/m-labs/artiq/issues/703#issuecomment-292599550

21:03 <GitHub> [artiq] npisenti opened issue #704: plot_xy applet fails with pyqtgraph < 0.10.0 https://github.com/m-labs/artiq/issues/704

21:43 <GitHub> [artiq] jordens commented on issue #704: We have been using Qt5 for much longer. That itself can't be it.... https://github.com/m-labs/artiq/issues/704#issuecomment-292659140

22:01 <GitHub> [artiq] npisenti commented on issue #704: My thought about Qt5 and pyqtgraph 0.10.0 came from google pulling up this issue: https://github.com/Ulm-IQO/qudi/issues/47. Based on here (https://github.com/Ulm-IQO/qudi/issues/47), it might also originate in numpy 1.11 -> 1.12? https://github.com/m-labs/artiq/issues/704#issuecomment-292662404

22:01 <GitHub> [artiq] npisenti commented on issue #704: My thought about Qt5 and pyqtgraph 0.10.0 came from google pulling up this issue: https://github.com/Ulm-IQO/qudi/issues/47. Based on here (https://github.com/Ulm-IQO/qudi/issues/47), it might also originate in numpy 1.11 -> 1.12? https://github.com/m-labs/artiq/issues/704#issuecomment-292662404