Often, you must wait a bit for libraries you depend on to support it. First, numpy has to work, then matplotlib a month or two later, then pandas, etc. If you depend on an obscure package or two, you have to wait for basically one guy to learn the new stuff, otherwise they’ll pull your whole dependency tree back.
Been seeing most packages atm supporting py310 - py314. Other abandoned packages just have py3+ which is good enough.
The issue is when you go full async native and drop sync codebases. Then no choice but to step up and help maintain other AWOL author/maintainers projects.
Often, you must wait a bit for libraries you depend on to support it. First, numpy has to work, then matplotlib a month or two later, then pandas, etc. If you depend on an obscure package or two, you have to wait for basically one guy to learn the new stuff, otherwise they’ll pull your whole dependency tree back.
Been seeing most packages atm supporting py310 - py314. Other abandoned packages just have py3+ which is good enough.
The issue is when you go full async native and drop sync codebases. Then no choice but to step up and help maintain other AWOL author/maintainers projects.