Hey,
Related to python-pillow/Pillow#9649, I'd love to renovate this repo so e.g. the blend benchmark from that PR could be incorporated here, and eventually Pillow's performance over time could be measured, either by hand or with CodSpeed or similar solutions.
My plan is, at the moment:
- rewrite (probably with an AI agent, since it's a simple refactor task) the current test suite in pytest-benchmark terms; pytest-benchmark does measurements more scientifically than the current "run it 11 times and see if that works", for one :)
- set up standard Python packaging
- remake the frontend with Astro instead of a homegrown Webpack setup
If that sounds fine, I can make some (probably unfortunately pretty large) PRs.
Some questions though:
Hey,
Related to python-pillow/Pillow#9649, I'd love to renovate this repo so e.g. the blend benchmark from that PR could be incorporated here, and eventually Pillow's performance over time could be measured, either by hand or with CodSpeed or similar solutions.
My plan is, at the moment:
If that sounds fine, I can make some (probably unfortunately pretty large) PRs.
Some questions though: