Michael R. Crusoe

> It's also worth mentioning that the negative check should probably just check the top bit. You can have entirely valid -0, -nan, -inf, etc. Here's the psuedocode for `CheckFPClassDP()`...

> Ah, that it is "negative finite" makes sense. I'm over in ARM spaces dealing with a different set of problems. I didn't realize this was fully mapping onto some...

Pre-release is out: https://github.com/simd-everywhere/simde/releases/tag/v0.8.4-rc1

Hello @giuseppe998e ; would you like to have a meeting to discuss how to get this code merged?

It is nice to see the progress @giuseppe998e ; let me know if you need help with that mypyc error. Here is a hint: https://mypyc.readthedocs.io/en/latest/native_classes.html#inheritance

Hello, any thoughts on how to add big-endian compatibility for rhdf5? If that is not desired, please let us know and we'll stop building rhdf5 for s390x in Debian (and...

> Are the test failures reported by @tillea due to endianess then? That's my guess, but I haven't dug into them Here's a fresh log: https://ci.debian.net/packages/r/r-bioc-rhdf5/unstable/s390x/51476144/#S5 > I believe I...

CWL v1.2.1 was released, though we still welcome feedback. https://www.commonwl.org/v1.2/Workflow.html#Changelog_for_v1.2.1 & https://www.commonwl.org/v1.2/CommandLineTool.html#Changelog_for_v1.2.1 CWL v1.3 is under development at https://github.com/common-workflow-language/cwl-v1.3/issues

> It starts getting into very hardware specific sort of incantations to set the floating point rounding modes, and I'm not sure a lot of x86-isms in the test suite...

I'm guessing the error is due to the workflow level inputs are missing the secondary files definitions that are required by the tools https://github.com/simleo/deephealth-pipelines/blob/3037aa6e2e1795268cf52f4170850bab8c3e4b54/cwl/extract_tissue.cwl#L16 https://github.com/simleo/deephealth-pipelines/blob/3037aa6e2e1795268cf52f4170850bab8c3e4b54/cwl/predictions.cwl#L5 There should have been warnings...