Skip to content
This repository has been archived by the owner on Apr 4, 2023. It is now read-only.

Latest commit

 

History

History
61 lines (59 loc) · 2.53 KB

2020-07-27.md

File metadata and controls

61 lines (59 loc) · 2.53 KB

ASWF Python 3 Working Group (2020/07/27)

Action Items

  • Create a markdown page that makes the transition message as clear as possible [Martin]
    • What version(s) will support Python 2/3 concurrently?
    • What will be the last version to support only Python 2?
    • When do vendors expect only Python 3 support?
    • Any vendor specific recommendations?
    • Keep it simple so people don’t need to read the tea leaves
  • Create tables listing the most important plug-ins people need migrated to Python 3 (TBD)

Attendees

  • David Aguilar
  • Ryan Bottriell
  • Martin de Lasa
  • Dhrub Ghovil
  • Ollie Harding
  • Jess Lehrman
  • Joe Lohmar
  • JT Nelson
  • Ibrahim Sani Kache
  • Sean Wallitsch
  • Ashley Whetter
  • Mark Wiebe

Agenda

  • Review action points for last meeting
  • Any comments about Added additional Boost Python info. by csyshing · Pull Request #6 · AcademySoftwareFoundation/wg-python3 · GitHub?
  • -Any blockers?-
  • Any status updates?
    • Ryan B - having issues “motivating” teams; lack of commitment ATM
    • Key issues appears to be lack of dates from vendors
  • Python 3 Transition Value Proposition
    • Pros
      • Security patches
      • A number of dev-centric improvements
        • Async and other cool features
        • Thirdparty libraries - many have dropped python 2 support
        • Compiler support (you may be stuck re-building things)
        • Platform compatibility/support
        • Improved type checking support
      • Ability to on-board and hire talent
      • Standardization of Practice
      • The Python 2 toolchain is (bit) rotting
        • Increasingly we expect to run into conflicts b/w the library and tool chain (e.g., SSL versions)
        • People end up having to use a number of workarounds
        • It’s possible that PyPi extensions will cease to work with Python 2
      • Speed improvements
    • Cons
      • It will take effort / time
      • The path forward is unclear
        • No common / shared / established best practices
      • You may bifurcate your code base
        • Near term increase in complexity for a (possible) long-term payoff
        • There is also a bifurcation of your bugs
  • Improving Python Usage across VFX
    • Are there ways we can work more closely with the Python community directly to promote patterns/usages that would make like better / easier for us?
  • What about 3rdparty plug-ins?
    • Do vendors have a complete list of what’s important?
  • Risks
    • Some vendors may not have released versions in every calendar year (e.g., CY17 -> CY20)
    • Propose that vendors ship a “transition” release