Well, you are used to the workflow ;) The reasoning is to have a consistent release (version, tag, trigger docs build?, …) that can be done by more people. It does not hurt, it is a dev dependendy. We have started to use it for all our libraries … Ouranos (Montreal) was first ;)
Either way, we should document the release process … no matter the tool.
Here is an example where it came from (python cookiecutter template):
AUTHORS.rstto Markdown (
.md) files. Any thoughts/objections? I can submit a PR on this later today.
also noaa service are not available:
We use it for test-data but our
service_ok check is not intelligent enough to notice.
what do folks think about dropping Python 2 support altogether? Areas:
Thoughts? Do none/one/both?
remove-python2branch is part of the OWSLib repo (not my fork) so we can all push to it (not a small effort)
headersto be able to be passed for all services (not just WMS)? e.g. for auth
username/password(Basic Auth) is too limiting. We need e.g. Bearer Token auth. "We"=geopython/GeoHealthCheck#9. Additional problem: GHC only supports Python2 now, and OWSLib Python3 starting 0.18.0, or is a 0.17.2 branch/release still possible?