Dev Meeting on May 20
- TP
Resolvelib 0.4.0 update ready for review https://github.com/pypa/pip/pull/8267/
Tried briefly to upgrade packaging 20.4 and NormalizedName caused no linting errors 🤔
Avoid ResolutionTooDeep https://github.com/pypa/pip/pull/8275
- Paul fighting with Pytest to skip test based on fixture
- Pradyun
Filed PR for visualisation https://github.com/sarugaku/resolvelib/pull/52
- Deep in college stuff and expected to be mostly off the radar for the coming week
Dev Meeting on May 19
- Pradyun got 1000+ GH notifications down to ~350-ish.
- Yay!
325 of the remaining 350, are from pypa/pip.
- Pending PRs
- Wheel cache
- Waiting on Paul's review
Resolvelib 0.4.0 update
- Wheel cache
- What are we working on next? (this week)
- "open pool":
- packaging 20.4 is out!
- normalized name stuff.
- yanked warning
- Depends on the wheel cache PR
test_install_yanked_file_and_print_warning
- New resolver needs to print a warning when installing a yanked release.
resolvelib -- removing ResolutionTooDeep raising logic.
- we have to figure out what to replace it with, but this is a required step regardless of what other metrics we pick.
- markers handling
local-dir/[extra]
extras + not SpecifierRequirement
- naming!
- parent vs template
- ???
- packaging 20.4 is out!
- Pradyun
- (overflow) Visualizations (dump into examples/vis)
- Tzu-ping
ResolveLib 0.4.0
- Finder cache
Always passes a bunch of requirements, to find_matches, need to check how caching would work?
- Provider vs Finder?
For experimenting lru_cache()
Refactor the 150-line find_matches() implementation
- Finder cache
- Paul
- Look at remaining errors in the test suite.
Putting down xfails for test suite errors.
Mark that they're issues, and write xfails.
- Constraints
- URL constraints -- PEP 508 style URLs feels like it'd be worth it.
- Look at remaining errors in the test suite.
- "open pool":
- Later later
- yanked vs prereleases