Closed hoxbro closed 1 year ago
It might be worth adding a new test to confirm that a ValueError
is raised as in the example.
Added unit test. The failing tests on Python 3.6 are unrelated.
See https://github.com/holoviz/colorcet/issues/96 for issues with this approach. Can those be addressed?
I'm not sure I follow. Why is there an issue with this approach?
See https://github.com/holoviz/colorcet/issues/96#issuecomment-1276576516 . Skipping registering the map is dangerous unless you verify that the map is indeed the same, not just the same name.
I'd be fine with skipping it if the name is the same and a Python equality test indicates that it's the same object, which should be true if it's just about reloading the same code. ETA: Looking again at the code, I now see a line that does check that it's the same object. In that case, this looks fine! Sorry for the noise!
Python 3.6 tests were failing, which we'll have to consider when we do the next release. I assume that's to do with instantiating an environment rather than an actual problem, and I'm happy for this package to be installable on 3.6 even if there are aspects of the docs or other build steps that won't pass on 3.6.
That package won't be installable on Python 3.6 after the next release https://github.com/holoviz/colorcet/pull/110
Ah, yes. I guess it's fine; "colorcet" as a package should be available for Python 3.6, but the very latest versions need not be.
This comes up sometimes when debugging with
panel serve --autoreload
and changing a file in the holoviews/hvplot repo.This has only come up for me when debugging, not actual use.
Steps to reproduce: 1) Make a file called with
import colorcet
2)panel serve --autoreload
that file 3) Make some changes incolorcet/__init__.py
and reload the page