LSSTDESC / desc-help

DESC Computing Requests
BSD 3-Clause "New" or "Revised" License
2 stars 0 forks source link

td-env #82

Closed wmwv closed 2 years ago

wmwv commented 2 years ago

Description A clear and concise description of what the issue is.

version mismatch between CFITSIO and linked library.

Choose all applicable topics by placing an 'X' between the [ ]:

To Reproduce Steps to reproduce the behavior:

  1. When running desc-td-env Jupyter environment, as set up with

source /global/cfs/cdirs/lsst/groups/TD/setup_td_dev.sh and then source /global/common/software/lsst/common/miniconda/kernels/setup.sh

  1. Start new Jupyter kernel with desc-td-env3.
  2. Run from lsst.daf.butler import Butler
  3. See error:

"""

WARNING: version mismatch between CFITSIO header (v4.000999999999999) and linked library (v4.01).

WARNING: version mismatch between CFITSIO header (v4.000999999999999) and linked library (v4.01).

WARNING: version mismatch between CFITSIO header (v4.000999999999999) and linked library (v4.01).



**Screenshots**
If applicable, add screenshots to help explain your problem.
wmwv commented 2 years ago

I don't get this error with the current desc-stack-weekly

heather999 commented 2 years ago

@wmwv This is a warning due to how cfitsio checks its version - apparently a rounding error. So an upstream, non-fatal issue. Noted by the DM folks on Slack, and in the healpy repo.
I don't see that there is a definitive fix available - perhaps when healpy patches the cfitsio version checking and rubin-env updates to that version of healpy, this warning will disappear - though reading the Slack thread, it wasn't clear which package(s) were causing those warning messages. In the meantime, I think it's safe to ignore.

You won't see this in desc-stack-weekly, since that's an older version of the stack (and something I hope to upgrade this week) and older versions of the dependencies.

wmwv commented 2 years ago

Thanks, @heather999 The idea that a rounding error is involved is mildly terrifying. Versions are not floats and I'm deeply worried about anything that treats them that way. I appreciate that you're on top of it and tracking the issue.

wmwv commented 2 years ago

@rknop I see what you're saying now that 4.000999999... = 4.001 which is still different than 4.01

heather999 commented 2 years ago

Closing - we understand the issue, I believe we are still waiting for a fix upstream from us.