search

frictionlessdata / datapackage-py

A Python library for working with Data Packages.
https://frictionlessdata.io
MIT License
191 stars 43 forks source link

Support for Python 3.7 #221

Closed uvchik closed 5 years ago

uvchik commented 6 years ago

We use datapackage in our project and tests on travis-ci with python 3.7 fail, due to a problem with the cChardet package. In your requirements cchardet is restricted to <2.0.

  1. Is there a reason why the most actual version of cchardet is not used? I do not know if it actually solves the problem, but it cannot be solved with this restriction in the future.

  2. Are you planing to support Python 3.7 in the future?

roll commented 6 years ago

Hi @uvchik, There was already an effort to increase cchardet version - https://github.com/frictionlessdata/datapackage-py/pull/219/files (but we need need <=3 I think). But tests fail.

I suppose there are minor breaking changes in cchardet (the reason we have it as <=2).

@akariv WDYT - is it OK to bump the cchardet version (will it be non-breaking actually)? Probably they have changes some encoding names.

thenouch commented 5 years ago

Hi there, I stumbled upon this after having the same issues in regards to installing the package. If there's any work arounds, I'd appreciate the support. Version of Python that I'm currently using is 3.72

roll commented 5 years ago

Would someone like to have a go on this issue?

We need to add python3.7 to the .travis.yml and make a PR pass the tests. It could be easy (OR not). But it makes sense to try anyway. Unfortunately, I can't do it by myself upcoming weeks.

billwanjohi commented 5 years ago

I successfully installed the cchardet>=2.1.3 wheel on Python 3.7. Based on the tags attached to the commit, it doesn't look like they're planning on backporting support to the 1.x series.

OriHoch commented 5 years ago

fixed in #227

roll commented 5 years ago

Thanks!

roll commented 5 years ago

v1.5+

uvchik commented 5 years ago

Great. Thanks :clap:

thenouch commented 5 years ago

Awesome, thank you very much. Works like a charm now.