Ben Greiner
Ben Greiner
The entry-points do have the same name and get installed in `$prefix/bin`. Of course they do collide. One overrides the other, depending on install order. I referenced exactly those lines...
> This was an idea to clarify things and any feedback can change that. Be aware that this is not a discussion about your future plans but about an existing...
- You don't have control over the install order. A user could have nbclassic already installed and then decide to upgrade notebook from 6.4 to 6.5. - Packaging systems like...
> notebook 6.5 is beta pre-release to gather this kind of inputs. It is already a problem with notebook 6.4
In any scenario where nbclassic and notebook is installed. The trigger could by jupyterlab or something else. There are people out there installing everything into system (distro packages) or user-sitelib.
notebook 6.5 branch already depends on nbclassic
> Can you replace an entrypoint in a distro when upgrading or installing a conflicting package at a later time? I would have assumed that "last write wins". Depends on...
Any news on this? I see the same. workflow: ```yaml - name: Setup Conda uses: conda-incubator/setup-miniconda@v2 if: matrix.distribution == 'conda' with: channels: conda-forge auto-update-conda: true python-version: ${{ matrix.python }} activate-environment:...
Hi @mcepl, it's great that you work on the tests for this package. Just about now there arose problems because the package skipped all the tests even with nose. And...
- https://github.com/eventlet/eventlet/issues/629 - https://github.com/rthalley/dnspython/issues/559 - dnspython itself does not read resolv.conf in their own tests anymore https://github.com/rthalley/dnspython/pull/519 - HTTPretty tells eventlet not to use its greendns but I guess that...