Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Bump dependency versions and update Travis CI #591

Merged
merged 6 commits into from
Sep 28, 2019

Conversation

reneeotten
Copy link
Contributor

Description

This PR bumps the minimum required versions for several of the dependencies. Most importantly, both for SciPy and NumPy we now require the LTS versions that are the latest ones to support PY27 (see #544).

  • clean-up of the codebase now that SciPy 1.2 is required (update tests and documentation were applicable)
  • officially remove support for PY34
  • update Travis CI: remove the use of conda and rely on pip to install the exact minimum version requirements for dependencies, also start testing with the development version of PY38. Of note, the failures with PY38 seem to originate from asteval, perhaps one could try to fix this already ahead of the release?).
Type of Changes
  • Bug fix
  • New feature
  • Refactoring / maintenance
  • Documentation / examples
Tested on

Python: 3.7.4 (default, Sep 15 2019, 17:02:10)
[Clang 10.0.1 (clang-1001.0.46.4)]

lmfit: 0.9.14+2.g36a022f, scipy: 1.3.1, numpy: 1.17.2, asteval: 0.9.15, uncertainties: 3.1.2, six: 1.12.0

Verification

Have you

  • referenced existing Issue and/or provided relevant link to mailing list?
  • verified that existing tests pass locally? Failures with 3.8-dev are currently allowed.
  • squashed/minimized your commits and written descriptive commit messages?
  • added or updated existing tests to cover the changes?
  • updated the documentation?

Require the LTS version of both SciPy and NumPy that works for Python 2.7
and above. Also bump the six requirement to 1.11 (released on Sep 17,
2017) and uncertainties to 3.0.1 (released on Aug 15, 2016), and add
missing coverage development dependency.
Support for Python 3.4 was dropped already in version 0.9.13; now also
remove it from the classifiers and python_requires in setup.py
Use the Travis CI Python versions in combination with pip for testing;
this has a few advantages:
- it is slightly faster than using conda
- with pip we can more easily install the *exact* minimum required
    versions of dependences
- allows for testing of Python development versions (i.e., 3.8)

Only potential drawback is that we do not necessarily the latest (micro)
version of Python, but for our testing that shouldn't matter.
@newville
Copy link
Member

@reneeotten Great!! I'm all in on requiring scipy > 1.2.

If the CI tests pass, I'll plan on merging this in a day or so, unless there are objections or more to do...

@newville newville merged commit 9dcfb71 into lmfit:master Sep 28, 2019
@reneeotten reneeotten deleted the bump_dependencies branch September 28, 2019 19:44
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants